WorkflowHR & Learning

SOP Change-to-Training Impact Agent

SOP training impact assessment that retrains only the people a revision touches

Every SOP change retrains only the people it touches, all trained before it takes effect.

See one case, screen by screen ↓
demo19.8sfor five agents to draft the impact of a 14-page gowning revision — 9 changes, 8 roles, 255 people
demo130hof retraining avoided on that one revision, against blanket read-and-understand of 213 h
target21minof coordinator time per revision, down from about 2.5 h by hand
target98.1%of people trained before the effective date, against a 98 % target
The problem

Why every SOP revision turns into a site-wide re-read

When change control approves a new revision, the training coordinator has until the effective date to get everyone who does the work trained on it. That means putting the old and new versions side by side, working out which changes alter what someone actually does, finding every role on the training matrix for that SOP, choosing how each role trains, and writing a check that proves the change was learned.

Done by hand, that is hours per revision, so the safe default wins: send the whole SOP to everyone on the matrix as read-and-understand. A spelling fix and a new limit get the same treatment. Operators re-read 14 pages to find one number that moved, the people who needed to be shown a new hands-on step only read about it, and the training hours go up with every revision.

target≈2.5hof coordinator time per revision, by hand
target4,259hof blanket read-and-understand at one site in 90 days
Where a revision’s coordinator hours goestimated
By hand2.5 days
With the solution0.3 days
  • Comparing the old and new revision0.8 → 0.1 d
  • Deciding which changes matter for training0.4 → 0.1 d
  • Finding every role and person on the matrix0.5 → 0.1 d
  • Choosing the method for each role0.3 → 0.1 d
  • Writing the knowledge check0.5 → 0.1 d

Estimated split of the coordinator’s time on one revision, by hand and with the solution; the totals are the solution’s targets of about 2.5 h and 21 min.

How it works

How a revision moves

Five specialist agents read the change, find who it touches, pick the method and write the check; the training coordinator and QA review and sign, and a sixth assigns the training.

What comes in
Revision inApproved revision · old and new versions
Agents at work
Revision Readerevery changed passage
Change Classifierfour classes, cited
Then
Training Matrix Mapperroles + people
Method Advisorsite rules
Then
Knowledge Check Writerfrom the changed text
Then
Learning System Assignerafter QA signs
A person decides
Training coordinator and QAreview, then QA e-signs
What comes out
Targeted assignments
Signed impact record
Trained before effective date
One case, step by step

One gowning revision, from change control to trained

SOP-MFG-0217, aseptic gowning and entry to Grade A/B areas, goes from Rev 06 to Rev 07 under CC-2026-0418 and takes effect on Oct 19. Here is how it reaches the right 255 people at Plant 3 — Leiden, screen by screen, in the working solution.

  1. 01Morning

    Every revision on its runway to the effective date

    Priya Raman · Training coordinator

    Priya opens the solution on Wednesday, Oct 7: 1,240 GMP staff and 8 SOP revisions between approval and effective date. Three wait on her, one is with QA for e-signature, and each bar on the runway fills as people train on the new revision. The gowning SOP sits in coordinator review, 255 people, 12 days to go.

    Needs you: “SOP-QC-0203 · HPLC system suitability — 7 people not trained · reminders sent 06:00.”

  2. 02One click

    Every revision starts the same way

    The agents

    A new revision arrives from change control with both versions and the change-control record. Priya presses Run impact analysis, and the agents compare the two revisions section by section, class every change, find the roles on the training matrix, propose a method for each and draft the knowledge check. This is the warehouse SOP that arrived at 07:40; the gowning SOP went through the same steps on Oct 5.

    “Every passage keeps its page and section; every class cites the text and the change-control reason.”

  3. 03Drafted · Oct 5

    Nine changes, each shown in the redline and cited

    Change Classifier

    The Revision Reader aligned the 14 pages of Rev 06 and Rev 07 section by section and found 9 changed passages; the Change Classifier sorted them into 2 changed limits, 2 new steps, 2 clarifications and 3 editorial changes. The first one matters most. In §6.4, operators now re-gown after a maximum of 3 hours in Grade A/B, not 4. The redline shows both revisions, and the change-control reason sits beside it: the gown-integrity study GIS-2026-07.

    “Right class? Your call goes on the record.”

  4. 04Next tab

    Eight roles, a method for each, and the reason

    Training Matrix Mapper · Method Advisor

    The 86 aseptic filling operators get an on-the-job demonstration under rule R4: a new hands-on step and two changed limits in Grade A filling. QC microbiologists only read the changed sections, because the glove-print limit is already tested through the SOP-QC-0088 Rev 12 knowledge check (rule R6). The 41 warehouse operators acknowledge, since every change that touches them is editorial. That comes to 83 h, where a blanket re-read would cost 213 h.

  5. 05Her call

    She changes a method, and writes down why

    Priya Raman · Training coordinator

    The agents proposed a knowledge check for the 31 cleanroom cleaning technicians, who clean Grade A between batches. Priya wants to see the new sanitisation step done, so she switches them to an on-the-job demonstration. The reason is kept on the record next to the proposal it replaced.

    Reason: “Hands-on step — I want to see it done, not only read.” Then: “Method changed — 124 h avoided now.”

  6. 06Drafted

    Questions built from the changed text only

    Knowledge Check Writer

    Five questions, each citing its section — one for each of the four substantive changes and one for the glove-type clarification — with a pass mark of 80 %. The old value stays in as a wrong answer, so a pass proves the change itself was learned: “How long may you stay in the Grade A/B area before you must leave and re-gown?” offers 4 hours, the Rev 06 limit. Sam Patel, the SOP’s owner, can edit any question in place.

    On screen: “‘4 hours’ is the Rev 06 limit — kept as a distractor so a pass shows the change was learned.”

  7. 07Sent to QA

    The training impact record goes to QA

    Priya Raman · Training coordinator

    Everything on the four tabs becomes the training impact record, TIR-2026-0418. Priya sends it to Omar Haddad with her one change listed (cleanroom cleaning technicians, knowledge check to demonstration) and a note: “Checked all 9 changes against the compare.” The record now reads 255 people in 8 roles, 88 h against 213 h blanket.

  8. 08Signed

    QA signs, and only then is anything assigned

    Omar Haddad · QA training approver

    Omar reviews the classes, the roles and methods and the knowledge check, then signs with his user ID and password. The meaning, “Approved — training impact assessment”, is recorded with his name and the date and time. He could also return the record with a comment.

    “Nothing is assigned until you sign.”

  9. 09Seconds later

    255 assignments live, the old ones cancelled

    Learning System Assigner

    Once the signature is verified, the Learning System Assigner creates 255 assignments on Rev 07 in the learning system, due Oct 18, the day before the revision takes effect. It cancels 23 open assignments on Rev 06 so nobody trains on the old version, and tells every person and their supervisor what changed and how they train.

    The assigner runs only after a valid QA e-signature.

  10. 10Across the site

    Targeted retraining, measured against blanket

    Dana Okafor · Head of production

    The site view: revisions assessed, each signed by QA; training hours avoided against blanket read-and-understand, now including the gowning SOP; people trained before the effective date; and coordinator time per revision. Of 336 changes in 90 days, 44 % were editorial and needed only an acknowledgement, and 91 % of the agents’ proposals were signed as drafted.

Who it’s for

Built for everyone between an approved SOP and a trained site.

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

PR
Priya RamanPrepares impact assessments
Training coordinator
Before
Compares revisions page by page, then sends the whole SOP to everyone on the matrix to be safe.
Now
Starts from cited changes and a method per role; checks each one, changes what she disagrees with, and writes down why.
OH
Omar HaddadSigns training impact records
QA training approver
Before
Approves training decisions without seeing which change drove each one.
Now
Signs one record that shows every class, every method with its rule, every override with its reason, and the questions.
SP
Sam PatelAseptic manufacturing
Document owner
Before
Writes quiz questions for every revision of the SOPs he owns.
Now
Gets questions drafted from the changed text, each citing its section, and edits any of them in place.
DO
Dana OkaforHead of production
Site leader
Before
Loses operator hours off the line to re-reading SOPs that barely changed.
Now
Sees hours avoided and on-time training by area, and co-signs the revisions that touch more than 500 people.
Built on the engine

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

Five specialist agents read the change, find who it touches, pick the method and write the check; the training coordinator and QA review and sign, and a sixth assigns the training.

Revision Reader

Aligns the old and new revision section by section and lists every changed passage with its page.

  • Reads approved revisions only
  • Every passage keeps its page and section
  • Aligns by section number and heading, not by page
Change Classifier

Classes each change as editorial, clarification, new step or changed limit, with its reason and confidence.

  • Never classes a change to a number, limit or time as editorial
  • Below 0.85 confidence, flags it for the coordinator
  • Cites the passage and the change-control reason
Training Matrix Mapper

Finds every role on the training matrix for the SOP, its head-count and area, and which changed passages touch each role.

  • Uses the matrix version in force on the approval date
  • Never drops a role that is on the matrix
Method Advisor

Proposes the training method for each role from the site’s rules, with the reason and the hours it costs.

  • Applies only approved rules
  • Avoids testing the same change twice across revisions
  • The coordinator and QA can override with a reason
Knowledge Check Writer

Drafts multiple-choice questions from the changed text only, each citing its section, with the old value as a distractor.

  • Questions come only from changed passages
  • One correct answer, cited
  • Plain language for operators
Learning System Assigner

After QA signs, creates the assignments in the learning system, cancels open tasks on the superseded revision and notifies each person and their supervisor.

  • Runs only after a valid QA e-signature
  • Due date before the effective date
Training coordinator and QA

Review and e-sign. The agents propose; a named person decides.

Ask in plain words

Ask about any revision, in plain words

Ask what changed, who must train or how many hours are saved — or tell it a rule to add. Answers point to the passage they rest on.

What changed in the gowning SOP?

SOP-MFG-0217 Rev 07 has 9 changes; 4 matter for training: re-gown after 3 hours, not 4; any growth on a Grade A glove print is an excursion (it was an action limit above 1 CFU); a new step to sanitise gloves after any touch outside Grade A, a corrective action for DEV-2026-0331; and supervisors check gowning qualification before Grade B entry. The other 5 are 2 clarifications and 3 editorial changes.

Why do editorial changes not need retraining?

Because nothing anyone does changes. Rule R1 sends editorial-only changes to an acknowledgement: the person confirms they have Rev 07 — 2 minutes instead of re-reading 14 pages. The class, its reason and the QA signature sit in the impact record — that is the answer an inspector gets.

Who is at risk of missing an effective date?

SOP-QC-0203, HPLC system suitability, takes effect on Oct 12: 7 people still to train, mostly QC analysts on night shift. Reminders went out at 06:00. If they are not done by Oct 11, the learning system takes them off HPLC work until they pass. Everything else is on track.

Add a rule: changed limits in Grade A need an on-the-job demonstration

Added as rule R7: a changed limit in a Grade A task needs an on-the-job demonstration. It applies to every revision not yet signed; on SOP-MFG-0217 it moves the 31 cleanroom cleaning technicians from a knowledge check to a demonstration. The rule is marked for Omar Haddad’s approval 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 training coordinator today: revisions waiting on her, one with QA, and every revision on its runway to the effective date.
Agents at workA new revision from change control: the passages appear as they are read, then each is classed, mapped to roles and given a method.
Changed passagesEvery change grouped by class, shown in a redline of both revisions, with the change-control reason highlighted beside it.
People and methodEach role on the matrix, the changes that apply to it, the proposed method, the rule behind it and the hours it costs.
Changing a methodThe coordinator overrides a proposal; the reason she gives is kept on the record.
The knowledge checkQuestions drafted from the changed text only, each citing its section, with the old value kept as a wrong answer.
Training impact recordThe revision, its changes and classes, roles and methods, the assessment and the signatures — generated from the tabs, marked draft until signed.
Sent to QAThe record goes to the QA training approver with the coordinator’s changes listed and her note.
Electronic signatureQA signs with user ID, password and the meaning of the signature; nothing is assigned before.
AssignedAssignments created on the new revision, open tasks on the old one cancelled, people and supervisors told.
Training matrixEvery role against every open revision and how it trains; dashed cells are proposals not yet signed by QA.
DashboardRevisions assessed, hours avoided against blanket read-and-understand, on-time training by area and what kinds of change came through.
Your site’s rulesMethod rules, the training window, knowledge-check settings, approvers, people and connected systems.
Governance

Built for GMP training: cited, decided by people, signed by QA.

Every class cites its sourceEach change is shown in a redline of both revisions with its section and page, beside the change-control reason it came from.
A changed number is never “editorial”A change to a number, limit or time is never classed as editorial, and anything below 0.85 confidence is flagged for the coordinator.
People decide, with a reasonThe agents propose; the coordinator checks every change and method. Any change she makes needs a written reason, kept on the record.
Nothing is assigned until QA signsThe QA training approver signs the impact record with user ID, password and the meaning of the signature. Revisions that touch more than 500 people need a second signer.
Trained before it takes effectAssignments are due before the effective date and name the exact revision. People not trained by then are taken off the role, and tasks on the old revision are cancelled.
Every step on the audit trailEach agent step, proposal, override and signature is recorded with who and when. The method rules are versioned and approved by QA.
Configuration

Your site’s training rules, not ours

The method rules, the training window, the knowledge-check pass mark and who signs are settings. Every change is versioned and on the audit trail.

SettingDefaultChoose from
Only editorial changes apply to the roleAcknowledgeOn-the-job demonstration · Knowledge check · Read & understand · Acknowledge
A new step or a changed limit in the role’s own taskKnowledge checkOn-the-job demonstration · Knowledge check · Read & understand · Acknowledge
A new hands-on step performed in Grade AOn-the-job demonstrationOn-the-job demonstration · Knowledge check · Read & understand · Acknowledge
Assignments are due1 day before the effective date1 · 3 · 5 days before the effective date
Reminders to the person and their supervisor7, 3 and 1 days before due7, 3 and 1 days · 3 and 1 days before due · Daily in the last week
Block work for people not trained by the effective dateOnOn · Off
Knowledge-check pass mark80 %60–100 %, in steps of 5
Who signs site-wide procedures (more than 500 people)Omar Haddad and Dana OkaforOmar Haddad and Dana Okafor · Omar Haddad
Connections

Works with the systems you already run

Document management systemapproved revisions and the current effective versions
Change controlthe reason for change, with linked deviations and CAPAs
Training matrixwhich role trains on which SOP
Learning systemassignments created and cancelled, completions read back
Site and role rosterwho holds which role, with work areas and cleanroom grades
Training impact recordthe signed record, as a PDF
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
21min
of coordinator time per revision
By hand≈ 2.5 h
With agents21 min
target
2,140h
of retraining avoided at one site in 90 days, against blanket re-reading
Blanket4,259 h
Targeted2,119 h
target
98.1%
of people trained before the effective date; target 98%
trained on time

“demo” = seen in the working solution, on its sample site data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · Training before the work: EU GMP Chapter 2 and 21 CFR 211.25. People, companies and products named on this page are fictional — characters and sample data in the working solution.

Questions

What training and quality teams ask us.

What is an SOP training impact assessment?

It is the decision, made each time an SOP is revised, about who must be retrained on the new revision and how — before it takes effect. The SOP Change-to-Training Impact Agent drafts it: it compares the revisions, classes each change, maps the roles from the training matrix, proposes a method per role and drafts the knowledge check, and QA signs the result.

How does it decide who needs retraining?

The Training Matrix Mapper lists every role on the matrix for the SOP and which changed passages touch each role. The Method Advisor then applies your site’s rules: the strongest change that applies to a role decides its method, from acknowledge for editorial-only changes to an on-the-job demonstration for a new hands-on step in Grade A.

Does it replace read-and-understand training?

It keeps read-and-understand where it fits and stops using it everywhere by default. People whose task changes take a knowledge check or a demonstration, people who need to know the change read only the changed sections, and people touched only by editorial changes acknowledge the new revision.

Where do the knowledge-check questions come from?

Only from the changed passages. The Knowledge Check Writer drafts multiple-choice questions from the changed text, cites the section behind each one, and keeps the previous value as a wrong answer so a pass shows the change itself was learned. The document owner can edit any question, and your site can require the owner to check them before the record goes to QA.

Do people stay in control?

Yes. The agents propose; the training coordinator checks each change and method and can change any of them with a written reason. Nothing is assigned until the QA training approver signs the training impact record, and he can return it with a comment instead.

Which systems does it connect to?

Your document management system for the approved and effective revisions, change control for the reason for change, the learning system for the training matrix, assignments and completions, and the site and role roster for who holds which role.

What happens to people who are not trained by the effective date?

Reminders go to the person and their supervisor before the due date. With the blocking setting on, the learning system removes anyone not trained by the effective date from the role until they complete, and open assignments on the superseded revision are cancelled.

How long does it take to go live?

The Agentic Solution Engine builds and deploys it from your requirements — your SOPs, training matrix and method rules — and it goes live once every quality gate has passed. We will walk you through it on a few of your own recent revisions first.

See it on
your SOP revisions.

We’ll run the SOP Change-to-Training Impact Agent on a few of your own recent revisions and your training matrix.