botminds.aiRequest a demo
SystemManufacturing & Quality

Deviation & CAPA Investigator

GMP deviation investigation and CAPA, with five agents doing the legwork

Every deviation reaches a ranked, evidence-backed root cause in days, closed inside the 30-day target.

Watch the filmBook a demo
See one case, screen by screen ↓
demo2.6sto assemble the evidence — 13 sources read, 3 similar deviations found
demo82%confidence on the most likely root cause, with the evidence for and against
target5daysto a root cause, from the day the deviation is opened
target100%of claims in the report carry a citation to their source
The problem

Why deviation investigations miss the 30-day target

When a batch fails a test, the clock starts. The investigator has 30 days to find the root cause, prove it, and agree a corrective and preventive action with QA — while the batch waits on hold.

Most of those days are not spent thinking. They go on chasing the batch record, exporting equipment and utility logs, searching three years of past deviations, and writing the report by hand. By the time the evidence is together, the clock is half gone — and the shortcut answer, “human error”, is the one that lets the same failure come back on another line.

target19daysto a root cause by hand
typical15–20%of deviations recur — the same failure, again
estimated≈60%of investigations closed inside the 30-day target
Where an investigation’s days goestimated
By hand19 days
With the solution5 days
  • Gathering records and logs7 → 0.5 d
  • Finding similar past cases3 → 0.5 d
  • Working out the root cause5 → 2 d
  • Writing the report and CAPA plan3 → 1 d
  • Review and approval1 → 1 d

Estimated split for a typical major deviation, by hand and with the solution.

How it works

How a deviation moves

Five specialist agents gather the evidence, find past cases, rank root causes and draft the report; QA approves the root cause and CAPA.

What comes in
Deviation opened4 sources · for the event's time window
Agents at work
Evidence gathererrecords and logs
Similarity finderpast cases
Then
Root-cause analyst5-Why · fishbone
Then
Report writerreport · CAPA plan
A person decides
QA approverapproves root cause and CAPA
What comes out
Investigation report
CAPA plan
Closed in 30 days
One case, step by step

One deviation, from opened to signed

A filter integrity test fails on aseptic filling Line 2 at 22:11. Here is what happens next, screen by screen, in the working solution.

  1. 01Morning

    Everything that needs her, on one screen

    Dana Okafor · QA lead

    Dana opens the solution and sees the plant as it stands: 10 open deviations, 3 waiting for her approval, 2 at risk of the 30-day target. Overnight the agents already flagged a likely repeat and drafted an extension request.

    Agents at work: “Similarity finder flagged DV-2026-0416 as a repeat of DV-2026-0177.”

  2. 02Opened

    The deviation, with the clock already running

    Sam Patel · Investigator

    DV-2026-0412: the post-use diffusion test of the final sterilizing filter failed — 18.9 mL/min against a limit of 13.6. It arrives classified Major, flagged as a possible repeat, on day 4 of its 30-day clock.

  3. 03One click

    “Gather evidence” — and the agents start

    Sam Patel · Investigator

    Sam presses one button. The agents read the deviation report, pull the batch history and the equipment and environment logs for the time of the event, search 36 months of past deviations, rank the possible causes and start the report.

    Every step is shown as it runs, with what it found.

  4. 042.6 seconds later

    An evidence board, every fact cited

    The agents

    Thirteen sources on one board — the integrity tester run log, the water loop log, calibration certificates, the fill room’s environmental log — each with the exact passage highlighted. And a warning nobody had to look for: this is a repeat. The 2025 fix for the same failure changed the Line 1 batch record only. Line 2 was never covered.

    “Evidence assembled in 2.6 s by 4 agents — 13 sources read, 3 similar deviations found, 4 candidate root causes ranked.”

  5. 05Next

    Root causes ranked, with the evidence against them too

    Root-cause analyst

    The most likely cause — the filter was tested while still warm from the hot water flush — sits at 82 % confidence, with every supporting fact and every fact that argues against it. Membrane wetting (11 %) and tester drift (5 %) are shown and ruled down, not hidden.

  6. 06Next

    Fishbone and 5-Why, down to a system cause

    Root-cause analyst

    The fishbone maps every candidate across method, machine, material, people, measurement and environment. The 5-Why does not stop at “the operator tested too early” — it goes on until it reaches why the procedure let that happen.

    Human error is never the answer on its own — only with the process reason behind it.

  7. 07Drafted

    The report, drafted in your template

    Report writer

    Summary, event, immediate actions, data reviewed, root cause, product impact and the CAPA plan — in the plant’s own report template, with a citation on every claim and the draft marked as a draft until QA signs.

  8. 08Planned

    A CAPA plan that covers every line

    Sam Patel · Investigator

    Five actions with owners and due dates — a cool-down hold for Line 2, the same change for Line 4 because it has the same equipment, a temperature check on the tester, and a check that CAPA scope covers every line from now on. Batch record changes go through change control.

  9. 09Signed · day 4 of 30

    QA reviews, and signs

    Dana Okafor · QA lead

    Dana reads the root cause, the evidence and the plan, then approves with an electronic signature: her user ID and password, and the meaning of the signature — “I approve the root cause and the CAPA plan” — recorded with it. Signed on day 4 of its 30-day clock.

  10. 10Every month

    The plant, getting better

    Dr. Maya Chen · Head of Quality

    Days to a root cause, investigations closed inside 30 days, root causes by category and repeat deviations by area — with every CAPA checked for effectiveness after it lands, so a fix that did not work comes back to the table.

Who it’s for

Built for everyone who touches a deviation.

The same case, seen by the four people who carry it — what their week looked like, and what it looks like now.

SP
Sam PatelProduction supervisor
Investigator
Before
Spends the first week of every investigation exporting logs and searching old deviations.
Now
Starts from a cited evidence board and ranked causes on day one; spends the time on judgement.
DO
Dana OkaforQA lead
QA approver
Before
Reviews reports she cannot trace, and sends them back for the sources.
Now
Clicks any claim to see the exact passage it rests on, and signs in minutes.
MC
Dr. Maya ChenHead of Quality
Head of Quality
Before
Learns an investigation will miss its date when the extension request lands.
Now
Sees every clock every morning; the extension request is drafted at day 25, with the reason.
OH
Omar HaddadEngineering — equipment and utilities
CAPA owner
Before
Gets actions that fix one line and miss the identical equipment on the next.
Now
Gets actions already checked against every line with the same equipment, with proof of completion tracked.
Built on the engine

5 agents. Each with one job, and hard limits.

Five specialist agents gather the evidence, find past cases, rank root causes and draft the report; QA approves the root cause and CAPA.

Evidence gatherer

Reads the deviation report and pulls the batch record, equipment, utility and environmental logs for the time window of the event.

  • Reads only the plant, batch and equipment named in the deviation
  • Every fact carries its source passage
  • Read-only access to GMP records
Similarity finder

Finds past deviations and CAPAs with the same failure mode, equipment or product, and checks whether earlier CAPAs covered this line.

  • Shows the similarity score and the reason
  • Never merges or closes records
Root-cause analyst

Builds the 5-Why and fishbone, ranks candidate root causes with confidence, and lists the evidence for and against each.

  • Proposes — a named QA person approves
  • Shows evidence against as well as for
  • Human error only with the process reason behind it
Report writer

Drafts the investigation report and CAPA plan in the plant template, with a citation on every claim.

  • No claim without a citation
  • Marks the draft as draft until QA signs
Clock keeper

Watches the 30-day investigation clock, warns owners at day 20 and drafts extension requests at day 25.

  • Cannot extend a clock — only drafts the request
QA approver

Approves root cause and CAPA. The agents propose; a named person decides.

Ask in plain words

Ask it anything, in plain words

Everyone on the team can ask about any deviation, root cause or CAPA — or tell it what to change. Every answer is cited.

Why did the Line 2 filter test fail?

Most likely, the filter was tested while still warm. It was flushed with water from the hot loop at 79.6 °C and tested 6 minutes later at 41.2 °C; the housing cooled 3.4 °C during the measurement, which reads as extra flow. The retest after a 30-minute cool-down passed at 9.8 mL/min. It is a repeat: the 2025 fix changed the Line 1 batch record only.

Which investigations could miss the 30-day target?

Two are at risk: visual inspection rejects on Line 2 is on day 28 of 30, with an extension request drafted for the Head of Quality; buffer pH in Suite 2 is on day 26, with one CAPA due Oct 12 still in progress.

Add the same temperature check to the pre-use test

Good catch — the procedure applies to any pressure-based integrity test, and the pre-use test after steam sterilization also starts on a hot housing. I added a preventive action to the CAPA plan: a housing temperature check before the pre-use test on Lines 1–4.

Make every sterile-filter test failure a Major deviation

Done. New classification rule: any failed integrity test of a sterilizing filter is Major, with the batch on hold. It applies to deviations opened from now on, and you can change it in Settings.

Every screen

The working solution, as it ships.

13 screens from the working solution, on its sample data. Pick one to see it large.

HomeWhat needs the QA lead today: open deviations, approvals waiting, clocks at risk, and what the agents did overnight.
Every deviationEvery investigation on its 30-day clock, filterable by class, area, owner and risk of missing the target.
The deviation workspaceThe event, the batch, the lifecycle and the timeline — with the agents ready to investigate.
Agents at workEach step runs in view: the report read, the batch history pulled, the logs searched, past cases matched.
The evidence boardThirteen sources on one board, each fact highlighted in its source passage — and the repeat flagged.
Ranked root causesCandidate causes with confidence and the evidence for and against each one.
Fishbone and 5-WhyEvery candidate across method, machine, material, people, measurement and environment, then the 5-Why to a system cause.
The investigation reportDrafted in the plant template, a citation on every claim, marked draft until QA signs.
The CAPA planCorrective and preventive actions with owners, due dates, the effectiveness check and change control.
Electronic signatureQA approves with user ID, password and the meaning of the signature, recorded together.
Every CAPAActions from approval to effectiveness check, with overdue and at-risk actions surfaced.
The quality dashboardDays to a root cause, closure inside 30 days, root causes by category and repeat deviations by area.
Your plant’s rulesThe investigation clock, warnings, approvers, look-back window and confidence threshold — all settings.
Governance

Built for regulated work: cited, signed, on the record.

Every fact cites its sourceClick any value on the evidence board or any sentence in the report to see the exact passage of the batch record, log or certificate it came from.
GMP records are read, never changedThe agents have read-only access to batch records and logs. Changes to a batch record go through your change control.
A named person signsRoot causes and CAPA plans are approved with an electronic signature — user ID, password and the meaning of the signature, recorded together.
No “human error” without the reasonThe root-cause analyst only accepts human error with the process reason that allowed it, so the fix changes the process — not just a person.
CAPA scope, checked across linesEvery action is checked against every line with the same equipment, so a fix on Line 1 does not leave Line 2 exposed.
Every step on the recordEach agent step and each human decision is recorded in the activity trail of the deviation, with who, what and when.
Configuration

Your plant’s rules, not ours

Everything that differs between sites is a setting, not a project.

SettingDefaultChoose from
Target to close an investigation30 days30 · 45 · 60 days
Warn the ownerAt day 20Day 15 · 20 · 25
Draft an extension requestAt day 25Day 25 · 27
Who approves root cause and CAPAQA leadAny named approver
Second approval for critical deviationsHead of QualityAny named approver
How far back to look for similar deviations36 months24 · 36 · 60 months
Call a cause “most likely” only above60 % confidence50 · 60 · 70 %
Classification rulesYour rulesAdd in plain words
Connections

Works with the systems you already run

Your quality systemdeviations and CAPAs, in and out
Batch recordssteps, timings and results for the batch
Equipment and utility logstesters, water loop, HVAC
Environmental monitoringroom and suite readings
SOP libraryprocedures the causes are checked against
Document controlthe report, exported in your template
What it changes

The difference, in numbers.

Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.

target
5days
to a root cause, from the day it is opened
By hand19 days
With agents≈ 5 days
target
94%
of investigations closed within the 30-day target
up from about 60% before agents
typical
40%
shorter deviation closure in published gen-AI programmes

“demo” = seen in the working solution, on its sample plant data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (McKinsey, Gen AI: a game changer for biopharma operations — 15–20% of deviations recur). People, companies and products named on this page are fictional — characters and sample data in the working solution.

Questions

What quality teams ask us.

What is deviation management software?

Software that takes a deviation from the moment it is reported, through investigation and root cause analysis, to an approved corrective and preventive action (CAPA) and an effectiveness check — with every step recorded. The Deviation & CAPA Investigator adds agents that do the evidence gathering, the search for similar past deviations and the first draft of the report.

How does it find the root cause?

The root-cause analyst lists candidate causes across method, machine, material, people, measurement and environment, gives each a confidence with the evidence for and against it, and builds a 5-Why and fishbone for the top cause. It proposes; a named QA person decides.

Does the AI change our GMP records?

No. The agents have read-only access to batch records and logs. Any change to a batch record goes through your change control, and every agent step is recorded.

How are approvals signed?

With an electronic signature: the approver’s user ID and password, and the meaning of the signature, recorded together on the deviation. Critical deviations can require a second approval from the Head of Quality.

Will it catch repeat deviations?

Yes — that is one of its main jobs. The similarity finder searches past deviations and CAPAs for the same failure mode, equipment or product, shows the similarity score and the reason, and checks whether an earlier CAPA actually covered the line where the new deviation happened.

Which systems does it connect to?

Your quality system, batch records, equipment and utility logs, environmental monitoring, your SOP library and document control. It reads what it needs for the time window of the event and writes the report in your template.

Can we set our own investigation clock and approvers?

Yes. The 30-day target, when owners are warned, when an extension request is drafted, who approves, how far back to look for similar deviations and the confidence needed to call a cause most likely are all settings.

How long does it take to go live?

The Agentic Solution Engine builds and deploys it from your requirements — your SOPs, report template and a sample of past deviations — and it goes live once every quality gate has passed. We will walk you through it on your own deviations first.

Manufacturing & Quality

More in this division.

All 11 solutions →
Batch Record Assistant

Any batch-record question answered in seconds, cited to the exact page and step.

1-min filmValidation Documents

Validation protocols and reports drafted in minutes, every value traced to its source page.

1-min filmAsk the Plant

Batch analysis that took three months now takes hours, with the query behind every answer.

1-min film

See it on
your deviations.

We’ll run Deviation & CAPA on a sample of your own past deviations.

Book a demoBrowse solutions