Life sciences
Who compares every partner document with your specification, line by line?
For each batch record, certificate or test report, an AI agent marks the lines that differ from your specification, with page and source. Your QA lead decides on each finding.
- Assay · matchesSpec PS-020, line 4
- Storage temperature · matchesSpec PS-020, line 9
- Expiry date · matchesCertificate, p. 1
- Deviation: test method versionReport p. 3, line 12
Waiting for your QA lead to decide
An ordinary week in your back office
For pharma, biotech and medtech firms, including those that own the product and buy development, manufacturing and testing from partners.
- Monday, 08:30
Three batches in one email
Your contract manufacturer sends the documents for three batches at once. Every page has to be checked against the specification before quality assurance can decide on the batches.
- Tuesday, 11:00
Which version applied?
A test lab reported against an older method. Now someone searches the document system for the version that was valid on that day.
- Wednesday, 14:00
A questionnaire from a customer
A customer wants a long supplier questionnaire back by Friday. Most of the answers are already in your SOPs and certificates, spread over several folders.
- Friday, 16:00
Board meeting on Tuesday
The board wants the state of each partner, open deviations and deadlines on a few pages. The pack gets built at the weekend, from folders only a few people know.
What our agents prepare for a life-sciences firm
The agent works between the systems you already have.
A chat answers one question for one person. An agent does a whole job. It takes data from several of your systems, applies your rules and hands the result to one of your people to check.
- Document management
- Quality system
- ERP
- Mailbox
The specification agent
Works by your rules
skillsplugins
Quality assurance decidesAccept or reject each finding
Set up before the first run
- AccessOnly the folders and systems you allow.
- GuardrailsRules the agent works within.
- LogEvery step is written down.
Batch release and every quality decision stay with your people.
Microsoft 365 is an important place for agents. It shouldn't be the only one. So we keep a second way open: your own server and open models.
See it first on partner documents you have already checked.
After the trial run you have three things in hand.
You start with a trial run on 20 of your partner documents. We run an agent, read only, over documents you have already checked. In a 90-minute session its draft sits beside your original. Then we give you a clear yes or no. Firms with 10 to 250 staff start with the trial run, larger firms with the AI assessment.
Questions from life-sciences firms
Does this work with our document and quality systems?
We clarify that in writing before we set anything up. At first the agent only reads. If your system already does the same job, we tell you.
What does this mean for validation?
We work that out with your QA lead before an agent becomes part of a controlled process. Until then it only suggests, and every decision stays with a person.
How do you deal with our partners?
The agent shows you the gaps before you decide, so you can clear them up with the partner. You keep the relationship with your partners.
Can you also help in development and manufacturing?
We start in the back office. Whether and when AI goes into development or manufacturing is your decision, once the first job runs.
Which partner sends you the most pages to check?
In 30 minutes we work out whether your partner documents suit a trial run.
In the first call you talk to Lukas Hertig.
