destill.ai · All sectors · Deutsch · Trust Center

For QA, regulatory affairs and hospital IT

Who shows what the AI agent saw and did?

AI already checks images and documents in GxP-related processes. Later, nobody can show which model decided what, and whether the image was the one that was captured. A POAW receipt per step and a sealed image give your QA something to check instead of something to believe.

Try it yourselfRequest a pilot

See also: Pharma Medical Affairs: a receipt for every answer an AI agent prepares

Quality assurance

An audit trail inside the vendor’s system is only as strong as its administrator. QA wants a record it can check on its own.

Regulatory affairs

Part 11 and ALCOA+ ask for attributable, contemporaneous, original records. For AI steps that is still mostly a promise.

Hospital management

Fictional case from the film: a triage agent marks a patient “not urgent”; later her X-ray in the record looks different. That is a liability question.

The film

The opening scene, as a short clip.

The full explainer

Chapters

Try it yourself: a signed receipt you can check

This is the kind of receipt you would get from every run of your agent after the pilot. Synthetic data, real cryptography: an ML-DSA-65 signature and an RFC 3161 time stamp from an independent authority. Check it in your browser, then check a copy where one value in step 3 was changed.

Visual quality-control step (fictional batch)

StepWhat the agent didTime (UTC)
1Batch received
batchId: B-26-0417 · product: tablet 50 mg (fictional) · line: L3
06:15:00
2Images captured
station: inspection camera 2 · images: 3 × hash
06:15:42
3Defects classified
model: vision-qc 1.7 · defectsFound: 2 · classes: chipped edge, spot · confidence: 0.94
06:15:44
4SOP limit checked
sop: SOP-QC-112 v5 · acceptLimit: 3 · result: within limit
06:15:44
5Disposition proposed
disposition: forward to QA review
06:15:45
6Human approval
role: QA reviewer · approver: hash
07:02:19
7Record archived
system: eQMS · record: hash
07:02:21

Check this receiptCheck a tampered copy

Download the receipt (JSON) · Or drag the downloaded file into the verifier yourself, edit one number in a text editor, and drag it in again.

Your engineers can repeat the check offline with OpenSSL 3.5 or later and the published script: check pack.

What you get

Each item carries its status from our feature register: Built, Pilot or Roadmap. Hover a label for the evidence behind it.

The legal frame

Where a receipt helps you evidence an obligation. It does not fulfil the obligation for you and does not make anyone conform with the law by itself. Not legal advice.

AI Act Art. 50

Marking of AI-generated or manipulated content. Fortress marks can help you evidence that content was marked. Whether and how Art. 50 applies to you is a legal assessment.

Source: eur-lex.europa.eu

AI Act, Art. 6(1) with Art. 12 and 26(6)

AI that is a safety component of a medical device under MDR/IVDR is high-risk; it must log events, and deployers keep logs for at least six months. A receipt helps you evidence that a log existed at a fixed time, unchanged.

Source: eur-lex.europa.eu

21 CFR 11.10(e)

Secure, computer-generated, time-stamped audit trails. The receipt helps you evidence order and time of AI steps with an independent time stamp; validation of your system stays with you.

Source: www.ecfr.gov

ALCOA+ (EMA guideline on computerised systems)

Attributable, contemporaneous, original. Signature and time stamp help you evidence the first two for each AI step.

Source: www.ema.europa.eu

EU GMP Annex 22 (draft) and GDPR Art. 9

Annex 22 on AI in GMP is a consultation draft; what the final text requires is open. Health data are special-category data: the receipt stores hashes, not the data themselves.

Source: health.ec.europa.eu

Planned, not built yet

What comes next, by demand and investment. Nothing here is part of the pilot.

The pilot: scope and duration set together, price after scoping

Scope

One use case, for example visual quality control or an image workflow, one site, one team. Runs in your network or cloud tenant. Plus a mapping document (AI Act Art. 12, 21 CFR 11.10(e), ALCOA+) as a working basis for QA and RA, not a validation. Final report with go/no-go. Your own implementation partner can connect it: we hand over the SDK and the sandbox set-up (one machine, Docker). The time frame is fixed in scoping.

Price frame: Price after scoping, based on your process and environment.

Success criteria, agreed in writing at kick-off

  • At least 99 % of AI steps in the pilot workflow produce a verifiable receipt.
  • Your team verifies a receipt on its own, without our help (OpenSSL).
  • SEAL locates changed image areas in at least 95 % of cases, false alarms at most 1 %, on a test set agreed up front.
  • Extra effort per work step is measured; we set the target together at kick-off.

Source: our written pilot offer; the targets are what we commit to measure, not results from an earlier pilot.

What you provide

  • One concrete workflow and an owner
  • Test data or test images and access to a test environment
  • Contacts from IT and compliance or audit
  • A connection to a time-stamp authority, at signing or later (signing itself works offline)
  • Your own legal assessment
  • Validation (CSV/CSA) and QA release stay with you

Want the written pilot offer? Choose "Talk about a pilot" below and we send it with our reply.