The AI-Native Data & AI Governance Platform

Master your data. Govern your AI. Prove it.

TellParrot turns the systems you already run into one governed record of your customers, suppliers, and assets, with evidence an auditor can verify.

Weeks
to a governed estate, not quarters
Guided
every rollout run with our team, at your pace
Any region
your data where you choose, even your own database
Tamper-evident
audit trail, verifiable by anyone
The problem

Your organization already owns the answers. They're just in eleven places.

The same customer, five records

Every system holds its own version of the truth, different spellings, histories, and consent. Nobody can say which is right, so every report starts with an argument.

AI nobody registered

Models and agents are making decisions on data no one governs, no owner, no risk classification, no record of what they touched. The board is asking; the answer is a shrug.

Compliance you can't prove

Policies live in folders. Retention lives in spreadsheets. When the auditor asks what happened, the answer is screenshots and goodwill.

The Platform

Seven pillars. One journey. In order.

Watch the data move, from scattered systems to governed, provable records.

Decide who owns what. Policies attached to real systems, owners with names, gaps visible first to you.

/0.1 · Data Governance
Finance ERPOWNER · M. KHAN
Customer CRMSTEWARD · J. TURNER
Marketing platformSTEWARD · A. LEE
Legacy billingNO OWNER, FLAGGED
Govern

Every model and agent on the register, with an owner, a risk level, and the governance of its data.

/0.2 · AI Governance
Credit decisioningOWNER · M. KHAN
Churn predictorEU AI ACT · LIMITED
Support agentNO OWNER, FLAGGED
Data compliance76
AI

Every asset discovered, classified, and mapped, including the ones nobody admits to.

/0.3 · Data Catalog
customer_contactsPIICONFIDENTIAL
donations_historyPIIRESTRICTED
supplier_registerINTERNAL
meter_readingsLINEAGE MAPPED
Catalog

Quality measured in six dimensions. Bad data held for review before it reaches a report or a model.

/0.4 · Data Quality
Completeness94
Accuracy91
Timeliness96
Uniqueness82
312 records failed threshold→ QUARANTINED FOR REVIEW
Quality

Five conflicting records become one golden record, merges approved by people, never silent, always reversible.

/0.5 · Master Data Management
Sarah Mitchell · CRM
S. Mitchell · EVENTS
Sarah J Mitchell · GIVING
MITCHELL, S · FINANCE
SARAH MITCHELL
GOLDEN RECORD · QUALITY 94
MERGE APPROVED, J. TURNER ✓
MDM

One governed source feeding analytics, operations, and AI, with answers grounded in your own records.

/0.6 · Single Source of Truth
SSOT
ANALYTICS & BI
MICROSOFT FABRIC
AI REGISTER
REST API · MCP
SSOT

Every action recorded in a tamper-evident chain, evidence you can hand a regulator, verifiable without trusting us.

/0.7 · Lens · Audit & Evidence
⛓ 8f3a ✓ ⛓ 4b7d ✓ ⛓ e1c8 ✓ ⛓ 2d9b ✓
CHAIN INTACT, VERIFIABLE BY ANYONE
Lens
Mining & Resources
One governed record per asset, every site, every system.
Explore industries →
Platform architecture

Data flows in messy. It leaves governed, trusted, and provable.

From your business systems, through the TellParrot engine, into your single source of truth, and out to every consumer that needs it.

Business systems
Azure SQLDB
Snowflake · RedshiftDB
Salesforce · CRMCRM
SAP · Business CentralERP
Entra ID · PurviewMicrosoft
CSV · Files · SharePointDocs
REST API · WebhooksPush
TellParrot · Enterprise Data Intelligence Engine
01Data Governance
  • Roles & policies
  • Assessments & knowledge
02AI Governance
  • Model & agent register
  • Risk & compliance
03Data Catalog
  • Profile & classify
  • Lineage & impact
04Data Quality
  • 6-dimension scoring
  • Quarantine & review
05MDM
  • Match & merge
  • Golden records
06SSOT
  • Advisor & answers
  • Governed feeds
07Lens
  • Tamper-evident chain
  • Evidence exports
REST API MCP Server Webhooks Power Automate
SSOT · Consumers
Analytics & BI
AI / ML pipelines
Compliance & audit
Microsoft FabricDelta
Developer tools
Inside the platform

The life of a record.

Follow one customer through the platform, data arriving, changing, merging, and flowing into reporting, with TellParrot's checks at every step. Click a step, or let it play.

Step
TellParrot checks
The graph

Everything connects. Now you can see it.

Under the platform sits a knowledge graph: records, systems, people, policies, and AI in one connected model. Questions that used to take a data team a week become one click across the graph. Click a node, or run a trace.

Context graph, liveExplore
Select a node
Click any node in the graph to see what it is, and everything connected to it.
01 · Unify

Every record, system, person, policy, and AI model lands in one graph as sources connect. No modelling project required.

02 · See

Relationships nobody documented become visible: households, hierarchies, duplicate networks, dependency chains.

03 · Ask

Questions that cross systems, "who owns the data this AI relies on?", answer in one traversal instead of one meeting.

04 · Act

Fix what the graph surfaces, assign the steward, link the policy, flag the model, with approvals and receipts.

The graph answers Duplicate networks & householdsOwnership & stewardship chainsAI-to-data dependenciesSupplier & asset hierarchiesExposure across systems
Why now

Integrations used to be the project. Not any more.

TellParrot is built the modern, AI-native way, and it shows up in your rollout. We work at enterprise pace, agreements, security review, change control, and once you begin, the platform is never the bottleneck.

The legacy playbook
Scoping workshopsWeeks 1–3
Integration build, per sourceMonths 2–6
Consultants embedded on siteOngoing
First value deliveredMonth 9+
The TellParrot rollout
Agreements, security review, setupAt your pace
First sources connectedDays, once begun
First golden records under reviewThat week
Estate governed, liveWeeks, not quarters
Logistics & Supply Chain
Master data consistent enough to automate.
Explore industries →
What our customers say

Trusted by leaders who've seen every governance tool.

"In today's data-driven enterprises, visibility into systems, ownership, and governance gaps is not optional, it's foundational. Tools that map data roles, track maturity, and provide actionable insights, like TellParrot, align perfectly with Microsoft's vision for secure, compliant, and intelligent data estates."
Microsoft
Industry Expert
Cloud Solution Architect
"Having a unified view of systems, roles, and domains would instantly cut through the complexity we face in multi-source projects. It's the kind of foundation that every client needs to truly get a handle on data ownership and governance."
NCS
Mark Bielz
Project Management Office
"After 15 years of trying countless governance solutions, TellParrot is the first that gave us one place to see every system, each owner, and all related projects. We truly became a data-driven organization."
Healthcare
Jassen Surraj
Head of Data & AI, Medtronic
Microsoft-native

A first-class citizen of your Microsoft estate.

TellParrot runs on Azure, signs in with Entra, feeds Fabric, and strengthens Purview, procured through the agreement you already have.

Runs on Microsoft Azure

Hosted on Azure in the region you choose, we stand up a dedicated environment in any geography, and you can bring your own database.

Microsoft Entra ID

Single sign-on and SCIM provisioning. Access is managed where your identities already live, joiners, movers, and leavers included.

Microsoft Purview

TellParrot strengthens your Purview rollout rather than competing with it, a governed data foundation Purview can trust.

Microsoft Fabric

Governed golden records mirrored into Fabric as open Delta tables, trusted data, ready for your analysts and data scientists.

TellParrot for Fabric →

Power Automate

Trigger flows from governance events with the TellParrot connector, approvals, alerts, and downstream updates in your existing automation.

Azure Marketplace

A transactable Marketplace offer, procure TellParrot through your existing Microsoft agreement and Azure commitment.

Enterprise-grade

Security your CISO signs off. Compliance your auditor can check.

Tenant isolationDedicated database option
Encrypted at restAnd in transit, always
SSO & SCIMEntra-managed access
Tamper-evident auditThird-party verifiable
DSAR workflowsFound, actioned, evidenced
Data residencyAny region, the client picks

Assessment templates and risk mappings for the frameworks you answer to

GDPRAustralian Privacy ActHIPAASOXPCI DSSAPRA CPS 234EU AI ActNIST AI RMF
Flagship use case

The Single Customer View, and every change that made it.

One customer, resolved across every system, with the full history of how the record came to be, field by field, change by change.

Walk through the use case

Ready to see your own data become one record?

Forty-five minutes, on your data if you bring it. No deck, the product is the pitch.

Platform

One platform. Six jobs. Zero silent changes.

TellParrot doesn't sell features, it takes responsibility for outcomes. Here's what your organization can do after each step that it couldn't do before.

01 · Data Governance

Ownership, decided.

Accountability that holds up in the room, and in the audit.

  • Every system has a named ownerOwners, stewards, and custodians mapped to the systems and domains they're responsible for.
  • Policies attached to realityRetention schedules and compliance obligations linked to the systems they actually cover, with alerts before anything expires.
  • The gaps reach you firstUnowned systems and uncovered domains surface on your desk before they surface in an audit.
  • Maturity, measuredAssessment templates for GDPR, HIPAA, SOX, PCI, and APRA CPS 234 · current versus target state, per system, board-ready.
  • Tribal knowledge, capturedMeeting transcripts become searchable governance issues linked to people and systems, knowledge survives the people who carry it.
Explore Data Governance →
Ownership map14 of 15 covered
MK
Finance ERP
Data Owner: M. Khan · Finance domain
Covered
JT
Customer CRM
Steward: J. Turner · Customer domain
Covered
AL
Marketing platform
Steward: A. Lee · Customer domain
Covered
?
Legacy billing
No owner assigned, flagged for review
Gap
02 · AI Governance

AI, on the record.

Every model and agent your organization runs, owned, risk-rated, and governed by its data.

  • Every model and agent registeredLifecycle from idea to retirement, each with a named owner. Shadow AI ends here.
  • Risk mapped to the rules that applyEU AI Act levels and NIST AI RMF dimensions, per model, with impact assessments and incidents on the same register.
  • Governed by its dataEach model carries a compliance score from the governance state of its inputs. AI on ungoverned personal data is flagged the day it happens.
Explore AI Governance →
AI register12 models · 2 agents
Credit decisioning model
Owner: M. Khan · deployed · inputs governed
Low risk
Churn predictor
EU AI Act, limited · inputs include PII
Monitored
Support agent
No owner assigned, escalated for review
Flagged
Compliance scores derive from the governance state of each model's data inputs.
Data catalog1,284 assets
customer_contacts
Customer CRM · 12 columns · lineage mapped
PIIConfidential
donations_history
Fundraising platform · 15 columns · lineage mapped
PIIRestricted
supplier_register
Finance ERP · 8 columns · lineage mapped
Internal
meter_readings
Operations DB · 6 columns · 3 downstream reports
Internal
03 · Data Catalog

The estate, mapped.

A living catalog, not a spreadsheet compiled once for a project.

  • Every asset, discovered and describedWhat you hold, where it lives, what it contains, and how sensitive it is, kept current.
  • Sensitive data declares itselfPersonal and regulated information is identified and classified as it arrives, nothing hides in an unlabelled column.
  • Change without fearLineage and impact analysis show what depends on what, so you know who's affected before you change anything.
Explore the Data Catalog →
04 · Data Quality

Quality, measured.

"How good is our data?" gets a number, not a feeling.

  • Every record carries a scoreSix dimensions of quality, scored per record and rolled up per system.
  • Bad data stops at the doorRecords that fail your standards are held for review instead of flowing into reports, decisions, and models.
  • Drift gets caught earlyWhen quality degrades, the right steward hears about it while it's still cheap to fix.
Explore Data Quality →
Quality scoring, Customer domain89 overall
Completeness
94
Accuracy
91
Conformity
87
Consistency
89
Timeliness
96
Uniqueness
312 records held for steward review
82
Merge review queue31 awaiting approval
Sarah Mitchell ↔ S. Mitchell
CRM + Events · same email, phone, address history
96% match
OptiCore P/L ↔ Opticore Pty Ltd
ERP + Procurement · same ABN, address
92% match
Demo Widgets 40mm ↔ WIDG-40
Two branch catalogs · same supplier code
88% match
Every merge waits for a person. Every approval is recorded. Every merge is reversible.
05 · Master Data Management

One golden record.

Every customer, supplier, and asset, resolved, with receipts.

  • Duplicates found across every systemThe same entity is recognized however it's spelled, formatted, or fragmented, across all sources at once.
  • Merges are never silentEvery proposed merge comes with its reasoning and waits for a person to approve it. Approved merges can be undone, completely.
  • Every field has a receiptEach value traces to the exact source it came from and why it won. Ask any record to explain itself, at any point in its history.
Explore MDM →
06 · Single Source of Truth

Answers, on tap.

One trusted source feeding everything downstream, including your AI.

  • Downstream runs on trusted dataAnalytics, operations, and AI draw from one governed source instead of eleven conflicting ones, including your Microsoft estate, natively.
  • Ask in plain languageAn advisor grounded in your own records answers with sources cited, and when it proposes a fix, the change waits for your approval.
  • Answers with receiptsCounts and reports carry the evidence behind them, consent-safe and ready to defend.
Explore SSOT →
Where trusted data flows4 live feeds
Analytics workspace
Governed records · refreshed continuously
Live
Microsoft Fabric mirror
Open Delta tables · analyst-ready
Live
AI register
12 models · 3 flagged on ungoverned inputs
Monitored
API consumers
6 downstream applications · governed access
Live
Audit chainChain intact ✓
Merge approved, Sarah Mitchell
09:41 · J. Turner · ⛓ 8f3a…c21e
Verified
Retention policy updated
09:38 · M. Khan · ⛓ 4b7d…9a02
Verified
Evidence export delivered
09:21 · delivery receipt issued · ⛓ e1c8…77f4
Verified
AI model flagged, no owner
09:04 · automatic · assigned for review · ⛓ 2d9b…310a
Verified
07 · Lens · Audit & Evidence

Evidence, not trust.

The audit trail that proves itself, to you, and to third parties.

  • An audit trail that proves itselfEvery action is recorded in a tamper-evident chain. If anything were altered, even a timestamp, verification fails. A property, not a promise.
  • Travel through timeSee your estate exactly as it stood on any date. "What did we know, and when?" becomes a query, not an investigation.
  • Hand the regulator a bundle, not a meetingEvidence exports with delivery receipts, verifiable without taking our word for it.
Explore Lens →
Why seven pillars

The model is the thin layer on top.

On the slide, AI readiness is two boxes: a model and some data. In production, the model stands on a stack, and every layer below the first is governance work. That is why the platform has seven pillars: hover a layer to see which one carries it.

The slide version
AI model
Data
The production version
AI models & agentsAI Governance
Clean, scored training dataData Quality
Agreed definitionsData Catalog
Named owners & stewardsData Governance
Documented lineageData Catalog
Quality rules enforcedData Quality
Consistent mappings across sourcesMDM
One record per entityMDM
Consent, retention & privacyData Governance
Access controlsPlatform
Tamper-evident audit trailLens
A governance operating rhythmData Governance
The things only one person knowsAssessments
Honesty

What TellParrot is not.

Not a data warehouse or BI tool

We master and govern the records, then feed your warehouse and dashboards trusted data. They get better; they don't get replaced.

Not a replacement for your ERP or CRM

Your operational systems keep doing their jobs. TellParrot sits across them as the layer that makes them agree.

Not a black box

No unattended automation touches your data. Everything that changes state is proposed, approved by a person, recorded, and reversible.

Bring one messy source. Leave with golden records.

Use cases

What changes when the data agrees.

Each of these is a before-and-after we deliver, described by outcome, demonstrated live on your own data in a demo.

Master data

One version of everything.

Single Customer View Flagship

Before
The same customer exists in the CRM, the ERP, billing, and three spreadsheets, each version different, none authoritative.
After
One golden record per customer, every field traceable to its source, and the full history of how the record changed.
Walk through the full use case →

Supplier & vendor master

Before
Vendor lists live in the ERP, regional systems, and spreadsheets. Duplicates mean missed discounts and duplicate payments.
After
A unified vendor master across every region, duplicates merged with approval, cleaner spend analytics.

Product & SKU unification

Before
Every branch and supplier keeps its own catalog. The same item carries six codes and four names.
After
Governed golden SKUs consolidated across branches and suppliers, catalog chaos becomes one sellable range.
Governance

Accountability that holds.

Ownership & accountability

Before
Nobody knows who owns what data. Decisions stall; audit responses take weeks of archaeology.
After
Every system and domain has a named owner, and coverage gaps appear on a dashboard, not in an audit finding.

Policy & retention lifecycle

Before
Policies live in a folder; retention lives in a spreadsheet; neither knows which systems it applies to.
After
Policies linked to the systems they govern, retention enforced with sign-off chains, alerts before anything lapses.

Tribal knowledge, kept

Before
The reasons behind your data decisions live in people's heads. When they leave, it leaves.
After
Decisions, caveats, and context captured against the systems they describe, searchable by the next person, forever.
AI governance

AI you can put your name to.

The AI register

Before
Teams deploy models and agents with no register, no risk classification, no owner. The board asks "where is AI used?", silence.
After
A complete inventory of models and agents, each with an owner, a lifecycle stage, and a risk level mapped to the regulations that apply to you.

AI on governed data

Before
Models train and decide on data of unknown quality and unknown consent, an incident waiting for its date.
After
Every model's data inputs carry a governance score. AI running on ungoverned personal data is flagged the day it happens.
Compliance & privacy

The audit that answers itself.

Data subject requests

Before
A single DSAR means days of searching every system by hand and hoping nothing was missed.
After
One person, every system, one governed workflow, found, actioned, and evidenced with a verifiable trail.

Audit evidence on demand

Before
Audit season is screenshots, spreadsheets, and staff pulled off their jobs for three weeks.
After
Evidence exports on demand, who changed what, when, approved by whom, verifiable by the auditor directly.

Your use case, demonstrated on your data.

Industries

Your industry's version of the same problem.

The systems differ; the fragmentation doesn't. Each use case below is delivered by the same platform, governed, mastered, and provable.

See your industry's use case on your own data.

Flagship use case · Single Customer View

One customer. Every system. Every change.

How five conflicting versions of a customer become one governed record, and how your team can see exactly how that record came to be, field by field, change by change.

The journey of a record

Watch the data flow through the platform.

1 · Five versions arrive
Sarah MitchellCRM
S. MitchellEvents
Sarah J MitchellFundraising
s.mitchell@…Email platform
MITCHELL, SARAHFinance
2 · Matched, scored, approved
Duplicate cluster found96% match
Quality scored, 6 dimensionsDQ
Merge proposed with reasoningMDM
Steward approvesJ. Turner
Recorded in audit chainEvidence
3 · One governed record
Golden record, Quality 94SSOT
Every field carries its sourceProvenance
Consent resolved across channelsPrivacy
Feeds analytics, AI & operationsLive
The result

The record your teams can rely on.

Marketing, service, finance, and compliance finally read from the same page.

  • One view, every systemContact details, membership, giving, consent, consolidated with the best value from each source.
  • Every field has a receiptHover any value and see which system it came from and why it won over the alternatives.
  • Consent-safe by construction"Who can we contact?" is answered from resolved consent across every channel, with the evidence attached.
app.tellparrot.com/acme/mdm/golden-records
SM
Sarah Mitchell Member
Golden record, merged from 5 source systems
Quality 94
FieldValueSource
Emails.mitchell@example.orgCRM
Phone+61 4•• ••• •12Website form
MembershipActive, renews OctCRM
Giving history12 donations since 2019Fundraising
ConsentSubscribedEmail platform
One person, every system, with field-level provenance.
app.tellparrot.com/acme/mdm/golden-records/sarah-mitchell/history
Record history View as at any date
Consent updated, Subscribed
Feb 2026 · Email platform · previous value retained
Verified
Merged with fundraising profile
Jun 2025 · approved by J. Turner · reversible
Verified
Email changed via website form
Aug 2023 · was s.mitchell@oldmail… · source recorded
Verified
Phone corrected by steward
Mar 2021 · reason captured · approver recorded
Verified
Record created from CRM import
Sep 2019 · original values preserved
Verified
The full life of the record, every change, source, and approver.
See how data changes

Not just the record, its whole history.

Most platforms show you the record today. TellParrot shows you how it got there.

  • Every change, on the timelineField updates, merges, and consent changes, each with its source, approver, and reason, verified in the audit chain.
  • View the record as at any date"What did we know on 30 June 2023?", see the record exactly as it stood, for disputes, audits, and reporting periods.
  • Merges you can take backIf a merge was wrong, unwind it completely, both records restored, and the reversal itself on the evidence trail.
What your team asks next

Questions the Single Customer View answers on day one.

"Who can we email this campaign?"

A consent-safe count across every channel, with the evidence behind it, ready if a regulator asks.

"Is this the same person?"

A scored answer with the matching evidence shown, and a steward's approval before anything merges.

"What changed since the audit?"

Every change to every customer record in the period, filtered, exported, and verifiable.

See your own customers become one record.

Bring an export from two systems to the demo, watch them resolve live.

Approach

Built for estates with hundreds of sources.

If you run three systems, any tool works. You run three hundred, across decades, acquisitions, and shadow IT. Our whole approach, the platform, the rollout, the economics, is designed for that reality.

The problem with the old playbook

At enterprise scale, "per-source projects" is a death sentence.

300 sources × 3 months each

The legacy model treats every integration as a consulting project. At estate scale that arithmetic never finishes, so the estate never gets governed. Most programs quietly stop at the first dozen systems.

Serial rollouts stall

One domain at a time, one system at a time, one committee at a time. By the time wave two starts, wave one's sponsors have moved on and the business case has expired.

The long tail stays dark

The top 20 systems get attention. The other 280, the spreadsheets, the departmental apps, the acquired company's CRM, are exactly where the risk lives, and no program ever reaches them.

How an estate rollout actually works

Five phases. Waves, not projects.

This is the operating model we run with every enterprise customer, at your pace, inside your change control, with your team holding the pen.

The estate operating model
01
Map the whole estate first
Every system lands in the catalog before anything connects: name, department, owner-if-known, suspected contents. Anyone in your organisation can propose a system with evidence, so the long tail surfaces itself instead of hiding. Weeks, not a discovery consultancy.
02
Prioritise by risk and value
The assessment scores the mapped estate: where PII lives, which systems feed AI, what your regulator touches, where quality bleeds money. The first wave picks itself, usually 10 to 20 sources carrying 80% of the exposure.
03
Connect in waves, in parallel
Native connectors, APIs, webhooks, and file feeds land sources in parallel, each one in days because AI-native engineering makes integration cheap. Wave two starts while wave one is still being verified. Large loads run in the background with live progress; the jobs monitor shows operations everything.
04
Master domain by domain
Customer first, supplier next, assets after, each domain mastered with steward-approved merges and quality gates while other domains keep connecting. Golden records compound: every new source enriches entities that already exist.
05
Govern continuously
Drift alerts watch every connected source. Assessments re-score quarterly against target. The council works a live agenda of surfaced gaps. Governance stops being a program and becomes an operating rhythm.
Why this scales when others don't

The economics that make 300 sources feasible.

Integration is cheap, so coverage is possible

Because we build AI-natively, a new connection is days of platform work, not a statement of work. That single fact changes the program: you stop rationing integrations and start covering the estate.

Built for volume

Candidate blocking keeps matching fast at millions of records. Loads, matching runs, and syncs execute in the background, gated, monitored, and resumable. Scale is an engineering property, not a services line item.

The long tail registers itself

Proposed-systems intake means the 280 systems you don't know about arrive with evidence attached, from the people who use them, reviewed by admins, never polluting the governed estate until confirmed.

One graph, compounding

Every source lands in the same model: same records, same lineage, same evidence chain. Source 200 is cheaper to govern than source 20, because the estate it joins already knows most of what it contains.

Your team holds the pen

Guided rollout, not embedded dependency. Stewards approve, admins confirm, the council decides, in your organisation, building your capability. Help is available; a forward-deployed army is not the business model.

Change control is native

Every wave ships through approvals that land on the evidence chain. Your change-advisory board doesn't fight the platform; the platform is the paper trail they wish they had.

Enterprise reality, respected

We plan around procurement, security, and change control.

Security review, answered with evidence

Questionnaires answered from the same evidence chain we sell. Residency, isolation, encryption, and identity documented for your assessors, in any cloud you require.

AI proposes. People decide.

We use AI aggressively and govern it completely: every AI-proposed change waits for a named person and lands on the record, including refusals. It's also how we build the product itself.

Honesty as policy

Every capability on this site exists in the product today, and we label maturity exactly. If a number isn't measured, we don't publish it.

Bring your source inventory. We'll plan the first three waves.

A working session, your systems list on the table, and a wave plan you keep either way.

Get started

See it on your data.

A 45-minute demo, or a 14-day guided trial, we stand your environment up with you and walk your first sources in. No self-service maze, no services contract.

Book a demo

Forty-five minutes, on your data if you bring an export. No deck, the product is the pitch.

14-day guided trial

Not a self-serve signup. We provision your environment, connect your first sources with you, and stay in the loop until your first golden records are live and reviewed.

Deployment options

Microsoft Azure first, or talk to us about a dedicated environment on AWS or Google Cloud, in the region you choose, with your own database.

Talk to usOne business day
Submissions go straight to the TellParrot team. Replies within one business day.
Legal

Privacy Policy.

Last updated: 15 January 2025

1. Who we are

TellParrot, Inc. ("TellParrot," "we," "us," or "our") provides a software platform that helps organisations govern their data, improve data quality, and comply with regulatory requirements.

RegionAddress
Australia310 Edward Street, Brisbane QLD 4000
United StatesTellParrot, Inc., 524 Broadway, New York, NY 10012
DenmarkLautrup Park, Lautruphøj 1-3, Ballerup DK-2750
United Kingdom1 St Katharine's Way, London E1W 1UN

Contact: info@tellparrot.com

2. Scope of this policy

This policy explains how we collect, use, disclose, and protect personal information when: you visit tellparrot.com (the "Website") or any sub-site we operate; you interact with our marketing, sales, or support teams; and you or your organisation use the TellParrot platform (the "Service").

It does not cover data you load into your own self-hosted TellParrot instance ("Customer Data"). For that, see our Master Subscription Agreement and Data Processing Addendum.

3. Personal information we collect

CategoryExamplesHow we obtain it
Identity & contact dataname, job title, business emailforms, demo requests, event sign-ups
Account datauser ID, role, organisation, authentication logscreated when your organisation onboards users
Usage datapage views, feature interactions, IP addresscookies, server logs, telemetry
Marketing datanewsletter preferences, webinar attendanceopt-in registrations
Support datatickets, chat transcripts, attachmentsdirect from you or your admin

We do not knowingly collect data from children under 16.

4. Why we process your information

PurposeLegal basis (GDPR)
Provide and secure the ServiceArt 6(1)(b) contract
Respond to enquiries or support ticketsArt 6(1)(b) contract
Send product updates & marketingArt 6(1)(a) consent / Art 6(1)(f) legitimate interest
Improve features, fix bugs, analyse performanceArt 6(1)(f) legitimate interest
Comply with law, protect our rightsArt 6(1)(c) legal obligation

You can withdraw marketing consent at any time via the unsubscribe link in every email.

5. How we share information

  • Service providers (e.g., Microsoft Azure, email delivery, analytics) bound by strict confidentiality and data-processing agreements.
  • Legal authorities if required to comply with applicable law or enforce our rights.
  • Successors in the event of a merger, acquisition, or asset sale (you will be notified).

We never sell personal information.

6. International transfers

We host TellParrot on Microsoft Azure regions selected by each customer. Where data moves outside the UK/EU, we rely on European Commission and UK "adequacy decisions," or Standard Contractual Clauses (SCCs) plus supplementary safeguards (encryption in transit and at rest, least-privilege access).

7. Data retention

We keep personal information only as long as necessary for the purposes above, then securely delete or anonymise it. Typical periods: demo-request data, 24 months after last contact; support records, 7 years (audit & legal); telemetry logs, 90 days, aggregated thereafter. Customer Data retention is governed by your organisation's contract settings.

8. Security

  • Encryption: TLS 1.2+ in transit, AES-256 at rest.
  • Role-based access & MFA for admins.
  • Continuous vulnerability scanning and annual penetration testing.
  • ISO 27001-aligned policies and SOC 2 Type II controls (under audit).

Still, no system can guarantee 100% security. We notify affected parties of any data breach as required by law.

9. Your rights

Depending on your jurisdiction, you may have the right to: access and receive a copy of your personal information; correct inaccurate or incomplete data; delete ("right to be forgotten") or restrict processing; object to certain processing (including direct marketing); data portability in a machine-readable format; and lodge a complaint with a supervisory authority.

To exercise any right, email privacy@tellparrot.com and we'll respond within 30 days (or sooner where required).

10. Cookies & similar tech

  • Essential cookies, site functionality (cannot be disabled).
  • Analytics cookies, aggregate usage (opt-out via banner).
  • Marketing cookies, only with your consent.

11. Third-party links

The Website may link to external sites (e.g., Microsoft Docs). We do not control their privacy practices; review their policies separately.

12. Changes to this policy

We may update this Privacy Policy to reflect legal or operational changes. When we do, we'll post the new version here and, if the change is material, notify registered users by email.

13. Contact us

Email privacy@tellparrot.com or write to any regional office listed above.

Industries

The data challenges

Where it hurts, and how it stops hurting.

In the platform

What gets mastered

One governed record for the entities that matter.

Typical systems we connect

← All industries

See this use case on your own data.

A 45-minute demo, or a 14-day guided trial run with our team.

Security & Trust

Security your assessors can verify.

How TellParrot protects your data, residency, isolation, identity, and an audit trail that proves itself.

Data residency

Hosted on Microsoft Azure in the region each customer selects, we stand up a dedicated environment in any geography. Your data stays in the jurisdiction your obligations require.

Tenant isolation

Your estate is yours alone, with the option of a fully dedicated database, bring your own PostgreSQL and your data never leaves your infrastructure. Nothing is co-mingled, ever.

Encryption

TLS 1.2+ in transit and AES-256 at rest, everywhere, always. Key management follows least-privilege access with role-based controls and MFA for administrators.

Identity & access

Single sign-on and SCIM provisioning through Microsoft Entra ID. Access is managed where your identities already live, joiners, movers, and leavers included.

Audit & evidence

Every action is recorded in a tamper-evident, hash-linked chain. If anything were altered, even a timestamp, verification fails. Evidence exports are verifiable by third parties without trusting us.

Practices

ISO 27001-aligned policies, SOC 2 Type II controls (under audit), continuous vulnerability scanning, and annual penetration testing. Breach notification as required by law.

Full details in our Privacy Policy · security questionnaires answered via Get Started.

Put your security team in the room.

We answer assessments with evidence, not adjectives.

Deployment

Runs where you run.

Microsoft Azure first, with dedicated environments available on AWS and Google Cloud, in the region you choose, with your own database if you want it.

Tell us where you need to run.

Pillar 01 · Data Governance

Ownership, decided. Gaps, visible.

Roles mapped to real systems, policies attached to reality, maturity measured against the frameworks you answer to, and the knowledge in people's heads captured before it walks out the door.

What it solves

Governance that lives in documents isn't governance.

"Whose call is this?"

Decisions stall because nobody can say who owns a system or a domain. Audit responses become archaeology projects across org charts and inboxes.

Policy on paper, reality elsewhere

Policies sit in folders with no link to the systems they govern. Retention periods expire in spreadsheets nobody watches.

Knowledge that walks out the door

The reasons behind every data decision live in people's heads and old meeting recordings. When key people leave, the organisation forgets.

Try it

The governance map, live.

Pick a system and see everything attached to it: the department it belongs to, the people holding governance roles, the policies that cover it, and the AI running on it. One system has gaps, close one.

Source systemsSteward coverage 3 of 4
Select a system
Data Governance CouncilMeets monthly · next 12 Sep
Recent council decisions
Try it · The assessment

Map the current state. Then tell us what keeps you awake.

This is how an engagement actually starts: your estate scored against target, a few honest questions, and a readback that ties every worry to the specific issues in your sources.

Current state vs target, maturity by dimensionOverall 1.9 of 5 · target 3.5
Ownership & roles
2.5
Policy coverage
2.0
Classification
1.5
Knowledge capture
1.0
Scored per system and domain against GDPR / HIPAA / SOX / PCI / APRA CPS 234 templates · the marker is your target state.
What keeps you awake at night?Pick any that ring true
What you get

Accountability that holds up in the room.

The value: audit responses in hours instead of weeks, decisions that stop stalling, and institutional knowledge that survives staff turnover.

Roles & the council

Data owners, stewards, custodians, quality analysts, and security officers, scoped to specific systems and domains, with a governance council tracked as a working body, not a slide.

Policies attached to reality

Full policy lifecycle, draft, review, active, archived, linked to the systems, domains, and regulations each policy covers. Retention schedules carry treatment workflows and stakeholder sign-off chains, with alerts before anything lapses.

Maturity assessments

Structured templates for GDPR, HIPAA, SOX, PCI, and APRA CPS 234. Score current versus target state per system and domain, with gap analysis and a board-ready compliance dashboard that stays live instead of expiring in a deck.

Tribal knowledge, captured

Upload meeting transcripts; governance issues are extracted with speakers and severity, linked to the people and systems they concern, and land in a knowledge hub searchable across transcripts, policies, documents, and glossary.

Domains & departments, mapped

Link departments to the domains they produce and consume. Coverage gaps, a domain nobody owns, a system no policy touches, appear on a dashboard, not in an audit finding.

Your estate, your worklist

Every steward gets an ownership-shaped home: the systems they own, the decisions waiting on them, and a live feed of governance events across the estate. Anyone can propose a new system with evidence; admins confirm or reject, and the trail is kept.

The day you can't live without it

The auditor's email arrives on a Tuesday.

They want the owner of every system touching customer data, the policies covering each, and evidence of your retention posture, by Friday. Without TellParrot that's three weeks of archaeology. With it, the ownership map, policy links, and assessment scores already exist, live, you export the answer before lunch and spend the week on your actual job.

Pillar 02 · AI Governance

AI, on the record.

Every model and agent registered with an owner, risk-rated against the EU AI Act and NIST AI RMF, and carrying a compliance score derived from the governance of its data.

What it solves

The board is asking. Shadow AI has no answer.

Nobody knows where AI runs

Teams deploy models and agents with no central register, no owner, and no risk classification. "Where do we use AI?" gets a shrug in the very meeting where it can't.

Models on ungoverned data

Decisions ride on data of unknown quality and consent. A model can be flawless and still be a compliance incident because of what it eats.

Regulation arrived first

The EU AI Act and local frameworks demand risk classification, impact assessments, and accountability, most organisations can't produce any of the three.

AI readiness

The model is the thin layer on top.

On the slide, AI readiness is two boxes: a model and some data. In production, the model stands on a stack, and every layer below the first is governance work. Hover a layer to see which pillar carries it.

The slide version
AI model
Data
The production version
AI models & agentsAI Governance
Clean, scored training dataData Quality
Agreed definitionsData Catalog
Named owners & stewardsData Governance
Documented lineageData Catalog
Quality rules enforcedData Quality
Consistent mappings across sourcesMDM
One record per entityMDM
Consent, retention & privacyData Governance
Access controlsPlatform
Tamper-evident audit trailLens
A governance operating rhythmData Governance
The things only one person knowsAssessments
Try it

Put a model on the register.

Register a model, then give it an owner, and watch the portfolio's compliance posture respond.

AI register2 registered
Credit decisioning model
Owner: M. Khan · inputs governed
Low risk
Churn predictor
Owner: A. Lee · inputs include PII
Limited risk
Portfolio compliance
Derived from ownership and the governance of inputs
100
What you get

Answers before anyone asks.

The value: a defensible AI posture, inventory, risk, and accountability, produced from a living register instead of a quarterly scramble.

The complete register

AI initiatives, models, and agents across the whole lifecycle, ideation, build, validate, deploy, monitor, retire, each with a named owner and AI-specific roles like model owner and risk owner. Shadow AI ends here.

Risk mapped to the rules

Risk classification aligned to EU AI Act levels and NIST AI RMF dimensions, with impact assessments captured per model, so classification is a property of the portfolio, not a one-off project.

Incidents & disclosures

An incident register, public disclosures, and contestability workflows, the accountability surface regulators and customers increasingly expect to see.

Governed by its data

Each model carries a data-compliance score derived from the governance state of its inputs. AI running on ungoverned personal data is flagged the day it happens, not the day it's discovered.

Agents under approval

AI agents don't act unilaterally: their proposed changes queue for human approval, with the evidence attached, and every decision lands on the record, including refusals.

Portfolio-level posture

ROI indicators, governance-readiness scores, and risk exposure rolled up across every initiative, the view a board pack needs, generated instead of assembled.

The day you can't live without it

A customer disputes an automated decision.

Under the rules now in force, you must show which model decided, who owns it, its risk classification, what data it used, and how the customer can contest. With TellParrot that's one register entry, model, owner, risk level, impact assessment, data-compliance score, contestability path, exported with evidence. Without it, it's a legal team and a war room.

Pillar 03 · Data Catalog

Every asset, found.

A living catalog of what you hold, discovered, profiled, classified, and mapped with lineage. Including the assets nobody admits to.

What it solves

You can't govern what you can't find.

The estate is a rumour

What data exists, where it lives, and what it contains is tribal knowledge at best. Every project starts by rediscovering the landscape.

PII hiding in plain sight

Personal data sits in unlabelled columns and forgotten extracts. You find out where it was when a subject request, or a breach, makes it urgent.

Change is a gamble

Nobody knows what depends on what, so every schema change and system migration risks silently breaking reports three teams away.

Try it

Search the estate.

Type to filter a sample catalog, then open an asset to see its classification and where it flows.

Data catalog
What you get

Change without fear.

The value: migrations and changes planned against facts, privacy questions answered from a map instead of a search party, and a catalog that stays current by itself.

Discovery & profiling

Assets are profiled as they arrive, column statistics, null rates, cardinality, distribution patterns, and kept current continuously, not compiled once for a project and left to rot.

Classification that finds PII

Personal and regulated information is detected and classified deterministically, the same input always classifies the same way, which is what a defensible privacy posture requires.

Column-level lineage

See how data flows from source columns through transformations to golden records and reports, and run impact analysis before you change anything, so you know who's affected first.

Trust scores

Every asset carries a trust score built from its quality, freshness, and governance state, so consumers can tell at a glance what's safe to build on.

A glossary that does work

Glossary terms aren't descriptions, they're live definitions attached to real records, each showing a live count of what currently qualifies. When the business asks "how many active members?", the term itself answers.

Documents included

Document stores like SharePoint sync into the catalog with incremental updates and the same PII scanning, because your estate is more than databases.

The day you can't live without it

The CRM migration is scheduled for the weekend.

Impact analysis shows the field being retired feeds four downstream assets, including a board report and a regulatory extract nobody in the project had heard of. Two hours of remapping instead of a Monday-morning incident, a broken submission, and a month of rebuilding trust. That's the catalog paying for itself in a single query.

Pillar 04 · Data Quality

Quality is a number, not a feeling.

Six dimensions, scored per record and rolled up per system, with bad data held at the door before it reaches a report, a customer, or a model.

What it solves

Bad data is discovered at the worst possible moment.

Quality is anecdotal

"How good is our data?" is answered with feelings. Nobody can say whether it's getting better or worse, or where to spend remediation effort.

Downstream discovers it first

Corrupt records flow silently into reports, campaigns, and models. The finance team, the customer, or the regulator finds the problem before you do.

Feeds rot quietly

A source system changes, a feed degrades, and quality decays for months before anyone notices, because nothing was watching.

Try it

Break the record. Watch the score.

Edit the fields, delete the postcode, mangle the email, and see the six dimensions respond in real time.

Record editorQuality 96
All checks passing. In the platform, records that fall below your thresholds are held for steward review.
What you get

Bad data stops at the door.

The value: remediation effort spent where it matters, downstream teams that trust what they're given, and quality problems caught while they're still cheap.

Six-dimension scoring

Completeness, accuracy, conformity, consistency, timeliness, and uniqueness, scored per record, rolled up per system and domain, and trended over time so improvement is measurable.

Quality gates on the way in

Incoming data is checked before it lands: clean records flow, failing records are held with a diagnosis, so one broken file never quietly poisons your golden records.

Quarantine & resolution

Records below your thresholds queue for human review with the reason attached. Nothing is silently dropped, and nothing bad silently proceeds.

Drift alerts

When a source's quality degrades, the right steward hears about it immediately, with the trend, not just the incident.

Agents that triage

Quality agents continuously queue the worst records for review, so stewards work a prioritised list instead of an ocean.

Repair, not rejection

Structurally broken files, the mangled export, the shifted columns, are diagnosed and repaired where possible, with every repair recorded. Real-world data is messy; the platform expects that.

The day you can't live without it

The campaign is loaded. Send is at 4pm.

A partner's overnight feed had silently corrupted, hundreds of contact records with shifted fields. The quality gate held them at the door that morning, flagged the source, and queued the records for review. The campaign went out clean. The alternative version of that day involves an apology email to your entire customer base.

Pillar 05 · Master Data Management

One golden record, with receipts.

Duplicates found across every system, merges approved by people and never silent, every field tracing to the source that won.

What it solves

Every system tells a different story about the same entity.

Five versions, zero truth

The same customer, supplier, or asset lives in every system under different spellings and identifiers. Reports disagree, and every meeting starts with an argument about whose number is right.

Automated matching you can't trust

Tools that merge silently create a worse problem than duplicates: records combined wrongly, histories contaminated, and no way back.

Migrations that move the mess

Every system migration inherits decades of duplicates because "cleanse later" means never.

Try it

Merge two records, with receipts.

Two systems hold the same person. Edit them if you like, run the match, approve the merge, and undo it, completely.

Match & merge2 source records
Customer CRM
Events platform
What you get

Merges you can defend. Forever.

The value: a single trusted view of every entity that survives audits, migrations, and mistakes, because every decision is human-approved, explained, and reversible.

Ingest from anywhere

File uploads with automatic column mapping, a full API, near-real-time webhooks, and native connectors across CRM, ERP, fundraising, e-commerce, and the Microsoft stack. Large loads run in the background with live progress.

Matching that scales

Configurable per-field rules across exact, fuzzy, phonetic, and token strategies, with candidate blocking so millions of records match in sensible time. Results arrive as scored duplicate clusters, not verdicts.

Humans hold the pen

Survivorship previews show exactly what the merged record will look like before anyone commits. Stewards approve or reject with the evidence in front of them, merges are never silent.

Golden records that explain themselves

Field-level provenance to every source value, identifiers and hierarchies, a relationship graph across entities, and a per-record explain log, ask any record why it looks the way it does.

Rules with receipts

Business rules keep golden records current, set a status from related records, standardise a field across the estate, and every run comes with an inspector showing exactly which records changed, before and after.

Truly reversible, and time-aware

Unmerge is a true inverse: both records restored exactly, with split receipts proving it. View any record as it stood on any date, and push approved corrections back to source systems on the same approval rail.

The day you can't live without it

Two customers were merged that shouldn't have been.

A father and son, same name, same address. In most MDM tools this is a data-loss incident with a restore-from-backup meeting. In TellParrot it's one action: unmerge, both records restored exactly as they were, the correction on the evidence chain, and the matching rule tightened so it can't recur. One of our distribution clients went live on a new ERP with twenty years of history mastered this way, clean masters on day one, full history intact.

Pillar 06 · Single Source of Truth

Ask your data. Get receipts.

One governed source feeding analytics, operations, and AI, and an advisor that answers plain-language questions from your own records, with the evidence attached.

What it solves

Whose number is right?

Every team, a different count

Finance, marketing, and operations each pull from their own system and arrive with different totals. Meetings are spent reconciling instead of deciding.

Data leaves ungoverned

Exports and extracts leave the building with no gate, no lineage, and no record of what went where, the governance you built stops at the door.

Answers take a ticket

Simple business questions, who can we email, how many active members, become analyst requests with a one-week turnaround.

Try it

Ask a question.

This demo answers from illustrative data, in the product, the advisor answers from your live records with sources cited.

Advisor
What you get

One source. Every consumer. All governed.

The value: decisions made on one agreed number, data that leaves the platform as governed as it was inside it, and business questions answered in seconds with evidence.

Governed feeds everywhere

A full read/write API, near-real-time webhooks, an MCP server for AI assistants, and a Power Automate connector, every consumer draws from the same governed source.

Native to your warehouse

Golden records mirror into Microsoft Fabric as open Delta tables and stream to your event infrastructure, analyst-ready without an export ritual.

Exports with receipts

Data leaves through a governed gate: what went out, to whom, under which policy, with lineage and delivery receipts. Outbound is governed like inbound.

An advisor that cites its work

Plain-language questions answered from your live records with sources shown. When it proposes a fix, relink these records, correct these keys, the change waits for your approval.

Segments that stay true

Business definitions, active member, contactable donor, priority supplier, live as governed segments with live counts. Campaign lists come back consent-safe, with the evidence attached.

One operations view

Every background process, loads, matching runs, syncs, exports, visible in one jobs monitor, so operations never wonders what the platform is doing.

The day you can't live without it

Board meeting, 9am. Two decks, two different member counts.

One of our conservation clients lived this monthly: CRM said one number, the fundraising platform another, email a third. Now "how many members can we actually email?" is one question to the advisor, answered from resolved, consent-safe records across every system, with the evidence attached. The meeting argues about strategy now, not arithmetic.

Pillar 07 · Lens · Audit & Evidence

Try to tamper. We'll wait.

Every action lands in a tamper-evident, hash-linked chain. If anything is altered, even a timestamp, verification fails. That's a property, not a promise.

What it solves

Logs record. They don't prove.

Audit trails you have to trust

Ordinary logs can be edited by anyone with enough access, which means their value rests entirely on trusting whoever holds the keys. Auditors know this.

"What did we know, and when?"

Disputes, reviews, and regulators ask about a moment in the past. Reconstructing it from backups and log fragments takes months and still ends in "we believe".

Evidence season eats the team

Every audit pulls people off their jobs for weeks of screenshots and spreadsheet assembly, and the output expires the moment it's produced.

Try it

Break the chain. Get caught.

Verify the chain, then tamper with a record and verify again. This is exactly how an auditor catches an edit.

Evidence chainUnverified
What you get

Evidence, not assurances.

The value: audits that cost days instead of weeks, disputes settled by verification instead of testimony, and a compliance posture you can hand to a third party.

A chain that proves itself

Every action is hash-linked to the one before it. Any alteration, a value, an actor, even a backdated timestamp, breaks verification visibly. Tamper-evidence is a property of the structure, not a policy.

The action ledger

Every executed change is a queryable record: who proposed, who approved, in what role, on whose behalf, and refusals are recorded too. Accountability includes the things that didn't happen.

Two-person control

Sensitive actions require a second approver over the exact content being approved, not a vague "approve" click. Cases collect the full decision trail.

Bundles a third party can verify

Evidence exports ship with everything needed to verify them independently, your auditor or regulator can check the chain without trusting you, or us.

Time travel, built in

View any record, or the whole estate, exactly as it stood on any date, including along business-event time, not just system time. "What did we know on 30 June?" is a query.

Delivery you can prove

Governed exports carry delivery receipts, evidence that the disclosure, the DSAR response, or the regulator's bundle actually went, when, and containing what.

The day you can't live without it

The regulator asks what you knew on 30 June.

A customer dispute has escalated. They want the state of the record on a specific date, every change since, who approved each one, and proof nothing was altered after the fact. With Lens that's an afternoon: an as-at view, the action ledger, and a verification bundle they can check themselves. Without it, it's six months of forensic reconstruction that still ends in "trust us".

Deployment · Microsoft Azure, primary

Native to the Microsoft estate.

Azure is TellParrot's home platform: hosted in the region you choose, signed in through Entra, feeding Fabric, strengthening Purview, and procured through the Microsoft agreement you already have.

What you get on Azure

First-class, end to end.

Your region, your residency

Deployed to the Azure region your obligations require, including Australian data residency, as a dedicated, single-tenant environment. Nothing co-mingled.

Identity where it already lives

Single sign-on and SCIM provisioning through Microsoft Entra ID. Joiners, movers, and leavers are handled by the directory you already run.

Your database, if you want it

Bring your own PostgreSQL, your data stays in infrastructure you control, with complete tenant isolation either way.

Fabric & Purview, connected

Golden records mirror into Microsoft Fabric as open Delta tables, and governance verdicts flow into Purview through an approval rail. See TellParrot for Fabric →

Procurement, simplified

A transactable Azure Marketplace offer, buy through your existing Microsoft agreement and Azure commitment, not a new vendor onboarding.

Encryption throughout

TLS 1.2+ in transit, AES-256 at rest, role-based access with MFA for administrators, and the tamper-evident evidence chain over everything that changes.

How rollout works

Enterprise pace, without the enterprise drag.

Guided rollout, Azure
Agreements & security review, questionnaires answered with evidenceAt your pace
Environment provisioned in your chosen regionWith your team
Entra sign-on, SCIM, and network posture connectedDays
First sources connected, first golden records reviewedDays, once begun
Guided go-live, our team stays in the loopWeeks, not quarters

Run TellParrot where your Microsoft estate lives.

Deployment · Amazon Web Services

A dedicated environment, inside your AWS world.

The full TellParrot platform, nothing rebuilt, nothing removed, provisioned as a single-tenant environment in the AWS region you choose, aligned to your cloud standards, as part of a guided rollout.

What "dedicated on AWS" means

Your region. Your network posture. Your database.

Single-tenant by construction

Your environment runs alone: dedicated application services and a dedicated PostgreSQL database. No shared infrastructure, no co-mingled data, no noisy neighbours.

Your region, your residency

Provisioned in the AWS region your obligations require, Sydney, Singapore, Frankfurt, Virginia, wherever you need to be. Data stays in-jurisdiction.

Bring your own database

Point TellParrot at your own PostgreSQL, including RDS in your VPC. Your records never leave infrastructure you administer, while the platform does the governing.

Your identity provider

Single sign-on through the IdP you run, Entra ID, Okta, or any SAML/OIDC provider, with SCIM provisioning so access follows your directory.

Network aligned to your standards

Private connectivity to your source systems, egress rules your security team defines, and encryption everywhere: TLS 1.2+ in transit, AES-256 at rest, keys under cloud KMS.

The same evidence chain

The tamper-evident audit chain, approval rails, and third-party-verifiable evidence bundles work identically on AWS, proof is a property of the platform, not the cloud under it.

How rollout works

Provisioned with your cloud team, not around it.

Guided rollout, AWS
Agreements & security review, we answer with evidence, not adjectivesAt your pace
Architecture session with your cloud team, region, VPC, database, egressOne workshop
Environment provisioned; identity and monitoring connectedWith your team
First sources connected, first golden records reviewedDays, once begun
Guided go-live, backups and runbooks handed overWeeks, not quarters

Operations & continuity

Automated backups with point-in-time recovery, health monitoring with alerting, and updates applied through change control you can see, every operational action on the evidence chain like everything else.

The honest version

Azure is our primary platform and gets new platform integrations first. On AWS you get the complete TellParrot product with your residency and network requirements met, Microsoft-specific integrations (Fabric mirroring, Purview write-back) connect across clouds via their public endpoints.

Tell us your region. We'll bring the platform.

Start the conversation with your security questionnaire, we like those.

Deployment · Google Cloud

The same platform, on Google Cloud.

A dedicated, single-tenant TellParrot environment in the Google Cloud region you choose, provisioned with your team, through your security review, to your standards.

What "dedicated on Google Cloud" means

Nothing shared. Nothing diluted.

Single-tenant by construction

Dedicated application services and a dedicated PostgreSQL database, including Cloud SQL in your project, if you prefer to bring your own.

Your region, your residency

Provisioned in the Google Cloud region your obligations require. Data stays where your regulator expects it.

Your identity provider

SSO through Entra ID, Google Workspace, Okta, or any SAML/OIDC provider, with SCIM so access follows your directory automatically.

Network to your standards

Private connectivity, your egress rules, TLS 1.2+ in transit, AES-256 at rest, keys under cloud KMS.

The same evidence chain

Tamper-evident audit, approval rails, and verifiable evidence bundles, identical on every cloud, because proof lives in the platform.

The honest version

Azure is our primary platform. On Google Cloud you get the complete product with your residency met; Microsoft-specific integrations connect cross-cloud via public endpoints.

How rollout works

Guided, end to end.

Guided rollout, Google Cloud
Agreements & security review answered with evidenceAt your pace
Architecture session, region, project, database, networkOne workshop
Environment provisioned; identity and monitoring connectedWith your team
First sources connected, first golden records reviewedDays, once begun
Guided go-liveWeeks, not quarters

Running on Google Cloud? So can we.

TellParrot for Fabric

Your data stays in your tenant. Your governance goes further.

Golden records land in your OneLake as open Delta tables. Governance verdicts write back into Purview. One hub governs your whole Microsoft estate, and TellParrot runs with your Fabric estate, through public APIs, without moving your data.

Three commitments

Yours, yours, and yours.

Your OneLake

Mastered, governed records arrive as open Delta tables in your OneLake, no export step, no proprietary format, ready for every Fabric workload you already run.

Your capacity

What the pipeline runs and what it draws is visible and auditable. Capacity consumption is a line item you can see, never an architectural tax you discover on the bill.

Your commitment

Consumption counts toward the Microsoft agreement you already have. One cloud bill, not a new line item, not a new procurement cycle.

The Microsoft Governance Hub

Five moves. One estate.

One place to connect, see, and govern everything Microsoft, with TellParrot's judgment layer working for it.

Connect, a guided wizard using your own Entra app registration, credentials encrypted. Your tenant, your keys.01
Discover, one click scans Entra ID, Purview, Fabric, Power BI, Power Platform, and Teams into the governance graph.02
Govern, the Context Graph surfaces the gaps: unclassified workspaces, systems with no steward, AI running on ungoverned data, each linking to an approvable remediation.03
Write back, verdicts flow into Purview, classifications, glossary, stewardship, through an approval rail. Purview gets richer, never overwritten.04
Prove, a consumption panel shows what the pipeline runs and what it draws, every figure traceable to its model.05
Architecture

The data plane is yours. The judgment layer is ours.

In your tenant

OneLake Delta tables · Purview · Dataverse · Event Hubs. Your data, your capacity, your keys, where it already lives.

In TellParrot

Match and survivorship decisions · the cross-estate governance graph · approval rails · connectors beyond Microsoft. The judgment, the evidence, the reach.

Integration is via public APIs. No part of the product is rebuilt, downgraded, or removed to fit.

Fabric-native, not Fabric-shaped

Built to respect your estate.

Efficiency by design

The engine doesn't reprocess your whole estate on a schedule to stay correct, so capacity draw stays an auditable line item, never an architectural tax.

Permissions with explain

A full permission model where every allow and deny can explain itself, access decisions you can defend, not just configure.

Your own AI keys

Bring your own AI provider keys, model usage under your account, your terms, your visibility.

Whole-estate reach

Microsoft is the centre of your estate, not the edge of it, the same governance covers your non-Microsoft systems in the same graph.

Column-level lineage

From source columns through decisions to Delta tables, lineage that answers impact questions before you change anything.

Actions with receipts

Every write-back and remediation is approval-gated and lands on the tamper-evident evidence chain, including the ones that were refused.

Availability

We label maturity exactly.

The same discipline we apply to your data quality metrics, applied to our own claims.

Available today

OneLake Delta export and mirroring · Purview write-back through approvals · Dataverse export · Event Hubs streaming · the Hub with live Entra ID and Purview discovery.

Rolling out

Live discovery for Fabric, Power BI, Power Platform, and Teams · per-tenant consumption reporting.

Roadmap

Processing push-down onto your own capacity · measured, not modelled, consumption telemetry.

Govern your Microsoft estate without moving it.

TellParrot × Microsoft

Built for the Microsoft estate.

Runs on Azure. Signs in with Entra. Feeds Fabric. Strengthens Purview. Reads SharePoint. Triggers Power Automate. Procured through the agreement you already have, and honest about what ships when.

The estate, end to end

Every Microsoft surface, working for governance.

Azure · Hosting

Runs on Microsoft Azure

A dedicated, single-tenant environment in the region you choose, including Australian data residency, with bring-your-own PostgreSQL if you want your data in infrastructure you administer. Encryption in transit and at rest, always.

The full Azure deployment page →
Entra ID · Identity

One directory, one truth for access

Single sign-on and SCIM provisioning through Microsoft Entra ID. Joiners, movers, and leavers flow from the directory you already operate, and access decisions can explain themselves.

Security & Trust →
Purview · Governance

Purview gets richer, never overwritten

Governance verdicts, classifications, glossary, stewardship, flow into Microsoft Purview through an approval rail. TellParrot strengthens your Purview rollout rather than competing with it: a governed data foundation Purview can trust.

Fabric · Analytics

TellParrot for Fabric

Golden records land in your OneLake as open Delta tables, your data stays in your tenant, your capacity draw stays visible, and consumption counts toward your existing Microsoft commitment.

The full Fabric page →
SharePoint · Documents

Documents in the same estate

SharePoint document stores sync into the catalog with incremental updates and the same deterministic PII scanning, because your governance obligation doesn't stop at databases.

The Data Catalog →
Power Automate · Marketplace

Automation and procurement, native

Trigger flows from governance events with the TellParrot connector, approvals, alerts, downstream updates in your existing automation. And procure through a transactable Azure Marketplace offer against your Microsoft agreement.

Talk procurement →
The Microsoft Governance Hub

One hub. Five moves. Your whole Microsoft estate.

Connect with your own Entra app registration · discover across Entra ID, Purview, Fabric, Power BI, Power Platform, and Teams · govern the gaps the Context Graph surfaces · write verdicts back into Purview through approvals · and prove what the pipeline runs and draws.

Hub, estate discovery
Entra ID
Users, groups, roles, live discovery
Available today
Purview
Classifications, glossary, live discovery + write-back through approvals
Available today
OneLake / Fabric export
Delta export and mirroring · Dataverse export · Event Hubs streaming
Available today
Fabric · Power BI · Power Platform · Teams
Live discovery, in review, not yet tenant-verified
Rolling out
Per-tenant consumption reporting
What the pipeline runs and what it draws
Rolling out
We label maturity exactly, the same discipline we apply to your data quality metrics. Full availability detail →

Govern your Microsoft estate without moving it.

Company

Twenty-five years in the room where it went wrong.

TellParrot was founded by veteran data governance and MDM consultants who spent decades inside the engagements, watching the same complex problem get solved the wrong way, with the same tools, for the same reasons.

Where we come from

We didn't study this problem. We sat in it.

Across utilities, banks, government agencies, and distributors, every engagement began the same way: six months mapping current state, chasing stakeholder buy-in, begging for accountability, with tools designed before the problems they were supposed to solve existed. We watched organisations spend six figures on a master data tool, six figures more on a separate governance platform, then scramble for an AI governance answer when the regulator called.

Three vendors. Three contracts. Three support teams. Zero shared metadata.

The wrong answer, bought again and again

MDM was bought to solve governance. It built silos instead.

Each discipline got its own tool, and the gaps between them became the real problem.

The pattern

Buy MDM to fix DG

Organisations reached for master data tools to solve a governance problem. The records got merged; the accountability, policy, and knowledge questions were never answered, just deferred.

The result

A committee of tools that don't talk

Governance in one platform, quality in another, mastering in a third, the catalog in a fourth. Different data models, different audit trails, different versions of the truth about the truth.

The cost

Silos between the disciplines

DQ couldn't see what DG decided. MDM merged what the catalog hadn't classified. Every integration between the tools was another project, another gap, another thing nobody owned.

1990s, 2000s

The monolith era

Batch jobs, on-premise monoliths, 18-month implementations, per-record pricing. Data lived in silos, and so did the tools that managed it.

2010s

The cloud-catalog era

SaaS catalogs arrived, still separate from mastering, still blind to quality, still nothing watching AI. The stack got newer; the fragmentation got worse.

Then AI arrived

AI on top of legacy tools and ungoverned data isn't intelligence. It's acceleration.

The last few years added a new layer to the mess: AI bolted over legacy stacks, trained on duplicated records, deciding on data of unknown quality and unknown consent. Models multiplied faster than anyone could register them. Agents began acting on records nobody trusted. The organisations with the most fragmented estates adopted AI fastest, and amplified every problem underneath it.

The moment

Two records, merged. Nobody knew why.

On one engagement we watched a legacy tool silently combine two customers. No evidence, no approver, no way back, and no one in the building who could explain it. The data was wrong, and worse: it was unaccountable. That moment is why Lens exists, why in TellParrot every merge carries its evidence, waits for a person, can be undone completely, and lands in a chain that proves itself.

So we started with a blank page

One platform. One data model. One audit trail.

Founded in 2023 · no legacy code, no acquired modules stitched together, no architecture decided before the cloud existed. The seven pillars were built together from day one, sharing the same records, the same workflows, and the same evidence chain.

The journey
Prototype scripts born of consulting frustration, automating the six-month stakeholder-mapping slog2020
First auto-mapping demo cuts traction from months to weeks; decision to build a platform, not scripts2021
Private beta, multi-tenant architecture, bring-your-own-database, security hardening2022
General availability, MDM, governance, catalog, and AI governance unified; first customers on three continents2023
AI governance deepens, agent oversight, MCP server, the evidence chain hardened2024
Sector accelerators, enterprise connectors, Fabric and Purview integration, scaling across energy, finance, health, and government2025 →
What we believe

Beliefs we're happy to be audited on.

Mission

Replace the fragmented, decades-old governance stack with one modern platform that delivers value in weeks, not quarters.

Vision

A world where every organisation, not just those with seven-figure budgets, can govern its data and AI responsibly, and prove it.

Values

Honesty in claims: every capability on this site exists in the product today, and we label maturity exactly. Transparency with customers. Evidence over assurances, including about ourselves.

Where we run

Any cloud, any region. Azure first, with dedicated environments on AWS, Google Cloud, or wherever your obligations demand, on your own database if you choose. If you can run it, we can deploy to it.

How we work

Engineering-led and AI-native. Capability ships in days, every claim on this site sits behind a live screen, and every release passes through the same evidence gates we sell you.

Who we serve

Data-intensive and regulated organisations everywhere, finance, government, utilities, health, and industry. Any enterprise whose decisions ride on data it must be able to trust, and prove.

Book a demo

Forty-five minutes, on your data if you bring it. No deck, the product is the pitch.

Prefer email? info@tellparrot.com reaches the same team.