Decide on a batch knowing every open item against it
An agent that walks the genealogy of a batch, finds every open quality record touching any part of it, and returns a disposition-readiness view with the blocking items named.
The batch looks ready right up until someone remembers something.
The release meeting starts. Testing is complete, the batch record is reviewed, disposition looks straightforward. Then someone remembers there is an open deviation on the chromatography skid that ran this lot, and someone else is not sure whether the CAPA from the March excursion was closed or just extended. A change control on the fill line went effective partway through the campaign and nobody has checked which lots fall either side of it.
So the meeting ends without a decision, and a QA associate spends the next two days reconstructing the picture by hand: pull the genealogy from the ERP, list every material lot and every piece of equipment, then query the QMS four separate ways to find what is open against any of them.
That reconstruction happens for every batch, it is done from scratch every time, and its completeness depends on whether the person doing it thought to check the right things.
One batch number in, a disposition-readiness view out.
The agent is given a batch number. It walks the genealogy to every material lot, every piece of equipment, then finds every open quality record attached to any of them and returns a single view: what is complete, what is outstanding, and what actually blocks disposition.
Genealogy first, then everything hanging off it.
The hard part of release readiness is not reading the quality records. It is knowing which quality records are relevant to this batch, which means the agent has to establish the batch's full footprint before it goes looking.
- ERP and MES to build the genealogy: every raw material and intermediate lot consumed, every piece of equipment used, every room occupied, and the date ranges for each.
- LIMS for release and in-process test status against the batch and against every material lot in the genealogy, including any OOS or OOT result and its investigation status.
- QMS queried against that whole footprint rather than against the batch number alone: open deviations, CAPAs past due, change controls effective mid-campaign, complaints and recalls touching any lot in the tree.
- Effective-date reasoning, which is where manual checks usually fail. A change control that went effective partway through a campaign is evaluated against the batch's actual manufacturing dates, not against the campaign as a whole.
- Blocking versus informational. Items are separated into what genuinely prevents disposition and what a reviewer should simply be aware of, with the rule behind each classification shown rather than assumed.
The agent produces a readiness view, never a disposition. It holds no credentials in the QMS, applies no electronic signature, and cannot release anything. A qualified person makes the call and signs for it inside your validated system, exactly as they do now.
Same rigor, without the two-day reconstruction.
- Per batch, spread across a QA associate, an email thread and a rescheduled meeting
- Genealogy pulled by hand, then four separate QMS queries against it
- Completeness depends on whether the reviewer thought to check the right things
- Repeated from scratch for the next batch, with nothing carried forward
- Full genealogy walked and every attached quality record found in one pass
- Blocking items separated from informational, with the reasoning shown
- The same checks run the same way on every batch, so completeness stops being a function of who is on shift
- A qualified person still reviews the evidence and signs the disposition
These figures come from my demo environment running on synthetic CMC data. They are not client results. The "today" column is what I watched teams actually do across my career; the "with an agent" column is measured on the demo, against a data model built to be representative rather than against your systems.
This one lives or dies on genealogy.
The workflow depends on being able to walk from a batch to every lot, room and asset it touched, and then join that footprint to the quality system. Where that breaks, it usually breaks in the same places: equipment identifiers that differ between the MES and the QMS, material lots tracked in the ERP but referenced by a different convention in the batch record, and change controls whose effective dates are recorded as free text.
The Data Diagnostic establishes whether those joins hold before anyone builds against them. Fixed fee, two to three weeks, and the output names the specific gaps and what closing each one costs.
The Data Diagnostic, including the price
Your IT and quality systems group will want the architecture, the read and write paths, the Part 11 position and the validation approach. That is all written down on the governance and validation page, in a form you can forward or print.
Send me a workflow.
Tell me the process that eats your team's week. I'll record an agent running it on your data model and send it back. No call required.
kyle@kylelangham.comWould rather talk it through? 30 minutes on your workflow, no pitch.