AIVby TrustBridge ComplianceRequest a demo
Prototype live since July 2026. MVP in build.

AIV tells you how much validation your AI system needs and drafts the evidence to prove it.

AIV is software for pharma, biotech and CDMO quality teams. It places each AI system in a GxP risk tier under the draft EU GMP Annex 22, sizes the validation, prepares the evidence set and records release under e-signature. Your people review and sign every record.

Join the design-partner pilotTry the free classifier
AIV · Vial inspection model, Line 3AIV-REC-0042

4 questions set the GxP risk tier

Does the output feed a GxP decision?Yes
Would a hidden 10% error put product or patients at risk?Yes
Does a person approve every output first?No
Can a person override it in the live workflow?Yes
Tier 3 · Guided workflowThe reason for each answer is kept with the record.

The tier sets the validation package

Risk tier and risk register
Model card
Test cases
Execution record
Release record
Drift monitoring plan

Tier 3 gets full IQ, OQ and PQ. A Tier 1 dashboard would get a light touch.

Acceptance criteria, first draft

Draft · awaiting QA reviewThe drafting layer writes. It never approves.

Release under a named signature

Name
A. Rao
Role
Head of QA
Meaning
Approved for release
Time
04 Nov 2026 · 14:32 IST
SIGNED

The tester who ran the OQ could not have signed this.

Live accuracy against the validated range

91%93%95%97%M1M4M7M10
!Month 10: accuracy fell to 92.7%, below the validated range. AIV opened a review and flagged the system for reclassification.
Illustration with sample data. The drafting layer and e-signatures arrive with the MVP.

This is what AIV shows a quality team

These 3 views use sample systems. The live prototype shows the same information, and we walk you through it on a demo call.

AI portfolio5 systems
SystemTierStatus
Deviation triage assistant2Released
Vial inspection, Line 33In review
CAPA routing model3Draft

Every AI system you run, with its tier, owner and next review.

ClassificationAIV-REC-0042
Vial inspection, Line 3Tier 3
Guided workflow · Ref: draft EU GMP Annex 22
6 records in the package

The tier, the reasons for it and the size of the validation.

Evidence setAIV-REC-0042
Risk registerSigned
Test casesIn review
Release recordDraft

Each record shows who prepared it, who reviewed it and who signed.

Quality teams now run AI systems that nobody has validated yet

Pharma companies use AI to triage deviations, inspect vials, route CAPAs and answer queries. The draft EU GMP Annex 22 sets out how AI in GxP work should be controlled. Most validation teams have a method for ordinary software. Very few have one for a model whose answers can change over time. AIV gives them that method, and it runs the same way every time.

How it works

How AIV works

Classify the system

AIV sets the GxP risk tier from what happens when the output is wrong.

Size the validation

The tier decides how much testing and evidence the system gets. Nothing is padded and nothing is skipped.

Draft the evidence

The drafting layer prepares the intended use, acceptance criteria and test cases for your people to review.

Record and release

Each test result, deviation and release is recorded under a named signature.

Watch for drift

After go-live, AIV compares live performance with the validated range and flags when the system needs another look.

The AI drafts and a person signs

The AI layer drafts and proposes. It never classifies a system, approves a result or lowers a criterion. Those decisions stay with a named person, and the product refuses anything else.

AIV
People
Drafting layerDrafts the intended use and acceptance criteria.
R. Mehta · QA reviewerReads the draft and corrects it.
Drafting layerTries to lower an acceptance criterion.Refused. Only a named person can change a criterion.
S. Iyer · TesterRuns the OQ and records the result.
S. Iyer · TesterTries to approve the same result.Refused. A tester can't approve their own result.
A. Rao · Head of QAApproves the release and signs.
MonitoringWatches live performance against the baseline.

Illustration with sample data. The names are made up.

AIV drafts, checks and watchesPeople decide and signRefused the rule is built into the product
Free classifier

Try the first step on one of your own systems

Answer 4 questions about one AI system and you'll see its risk tier and the validation it calls for. The model is the 4-tier model from the book.

1. Does the output influence a GxP decision at all?

2. If it were wrong on 10% of records for 6 months and nobody noticed, would patient safety, product quality or data integrity be at real risk?

3. Does a qualified person review and approve every output before a GxP action is taken?

4. Can a person override the decision through an exception path that exists in the live workflow?

Your resultAnswer 4 questions

Your tier appears here as you answer.

Get the full record for this system
After go-live

AIV keeps watching the system after you release it

A model can pass validation and drift 6 months later. AIV takes the performance numbers from your own tools each month and compares them with the range you validated. When a number leaves that range, AIV opens a review and sends the system back to be classified again.

Illustration with sample data. AIV doesn't compute model metrics. It turns your numbers into evidence.

91%93%95%97%M1M2M3M4M5M6M7M8M9M10M11M12validated range
!Month 10: accuracy fell to 92.7%, below the validated range. Review opened.

Who AIV is for

AIV is for quality units in pharma, biotech and CDMOs that use AI in GxP work. It's also for AI vendors whose customers keep asking how to validate their product.

How we charge

AIV is sold as an annual subscription. In India it starts at Rs 12 lakh a year per site. Our first customers join as design partners, with a free pilot and the first year at half price.

See pricing

Where AIV is today

Done

Working prototype

Running since July 2026. Its rules engine takes 5 sample systems from intake to release and drift monitoring.

In build

The MVP

The drafting layer, multi-user login, a full audit trail, Part 11-style e-signatures and real client data.

Next

Design-partner pilots

Pharma quality teams each run one live AI system through AIV for 12 weeks.

Later

Seed round

We plan to raise a seed round in 2027, after the pilots.

The people building AIV

Sachin Bhandari

Co-founder and product lead

25 years in pharma quality IT. He has validated 20 AI systems in live GxP use, and 9 of them went through FDA and MHRA inspection. He sits on the GAMP Global Steering Committee, took part in the industry review of the draft EU GMP Annex 22, and wrote Validating AI in GxP Environments: A Practitioner's Handbook.

Abhyuday Sachdeva

AI developer

He builds the AIV engine and its drafting layer.

Anu Bhandari

Director

She leads the company. TrustBridge is a women-led, DPIIT-recognised startup based in Jaipur.

TrustBridge also trains and advises quality teams

Training and advisory are our main business, and they fund AIV. They also show us every week where validation breaks in real companies.

Go to TrustBridge advisory

The book behind the method

Validating AI in GxP Environments: A Practitioner's Handbook sets out the method AIV runs. It covers the regulatory spine, risk tiering, the evidence that holds and monitoring after go-live. Its first readers rate it 4.7 out of 5.

Get the book

Questions people ask about AIV

Who builds AIV?

TrustBridge Compliance (OPC) Private Limited, a DPIIT-recognised startup in Jaipur, India. Sachin Bhandari leads the product, and he wrote the method AIV runs.

Can I buy AIV now?

AIV is a working prototype today, and the MVP is in build. Our first customers join as design partners. They get a free pilot on one live AI system and the first year at half price.

How is AIV priced?

AIV is an annual subscription. In India it starts at Rs 12 lakh a year per site. Outside India it starts at EUR 25,000 a year per organisation.

Does AIV replace my validation platform?

No. AIV sits beside the platform you already run and handles the AI-specific part of the work.

Is my data used to train anything?

No. Nothing you put into AIV trains a model, ours or anyone else's.

Does TrustBridge offer training and advisory?

Yes, and it is our main business. We train quality teams and advise on AI validation, CSV to CSA and QMS work. See TrustBridge advisory.