Workflow template for operations and process improvement

AI adoption measurement sheet

Once a month, the owner of an AI workflow in daily use by one team gets one sheet with four lines: is it installed, is it used, is it useful in the quality of its drafts, and is it useful to the business. It reads the team's work, never a person's, and the owner decides what changes.

View this template on GitHub

Put this on your planBuild this in a sprint

Add it to a workflow in the FullyPowered Planner, beside the teams and roles that will run it. Book a fit call

The outcome

The owner stops wondering how the rollout is going and reads it. Four lines, the same every month: who in the role can open the workflow; how many of the month's cases went through it, week by week; how the reviewer's sample of its drafts sorted out, with the reasons grouped; and where the one number the workflow exists to move stands against the month before it started.

Then a short paragraph reads the lines side by side and says what they point at together. Wide access with thin use is a question about how the team was brought on. Steady use with drafts that keep being rewritten is a question about the workflow itself. Good use, good drafts and a business number that has not moved is a question about whether this was the right work to hand over. That is how a whole team comes onto a workflow, and how you 10X the ROI of your existing AI one workflow at a time.

The sheet measures the team's work as a whole. Nobody is counted, ranked or named, and nobody is singled out as a low user. A change in the business number is reported beside its baseline, never explained as caused by the workflow, and a single month is called one reading, never a trend. The owner reads the four lines and decides.

Who starts the work, and when

The workflow's owner: the person in the team who keeps the workflow current, such as a credit manager or an intake lead. They prepare the first sheet at the end of the first full month after rollout, then every month, on the same inputs counted the same way. They bring it to the team's regular meeting.

What they supply

  • The access confirmation from IT: a short table by role saying how many people are in the role, how many can open the workflow, and whether it can read the records it needs. It is a list of access, not a record of anyone's use.
  • The team's work log for the month, exported from the system the work is done in: one row per case handled, with its completion date, a yes or no for whether the workflow prepared it, and a note. The default column headers are listed in the workflow template; when an export's headers differ, the sheet stops and asks which column is which.
  • The reviewer's sample record: each sampled draft marked sent as prepared or lightly edited, substantially rewritten, or stopped, with a one-line reason.
  • The one business number the workflow exists to move, for the month and for the baseline month before it started, counted the same way.
  • The notes from the team's regular meeting on what the workflow missed and when the team chose not to use it.
  • What else changed that month: an absence, a new rate or price, a changed form, a seasonal rush.

If anyone supplies a usage record by person, or a log with a column naming who handled each case, the sheet stops and asks for team-level counts instead.

What comes out

One sheet, in a fixed order:

  1. Installed: availability. How many in each role can open the workflow, whether its records are confirmed, and the date of the team's first real case.
  2. Used: repeat use. Cases handled, runs, and the share prepared with the workflow, for the month and for each week, with the direction from the first week to the last and the team's own notes on the cases done without it.
  3. Useful: quality. The reviewer's sample sorted three ways, with the reasons for rewritten and stopped drafts grouped in the reviewer's words.
  4. Useful: business effect. The month against the baseline, the change in points and the target, with the plain statement that one month is a reading and the sheet does not show what caused the change.
  5. Read the lines together: a short paragraph using only the facts supplied, ending with what the lines point at: the rollout, the workflow, the choice of work, or nothing.
  6. What else changed, with the line each change touches.
  7. Questions for the owner, at most six, and a blank line for the owner's decisions.

Every count on the sheet can be recounted from the work log and the sample record.

An example

Illustrative example with sample data. The company is an illustrative 70-person freight forwarder; its people, carriers and cases are illustrative too, and nothing here happened. Its quotes desk has four quoting coordinators and a desk lead. Since 14 September 2026 the desk has used a quote request preparation workflow: for each incoming request it prepares a draft quote from the customer's request and the carriers' rate sheets, and a coordinator checks and sends every quote. The desk lead owns the workflow. On Monday 2 November 2026 she prepares the sheet for October, the first full month of use. What she supplied comes first, in full. Then the sheet that came back.

What the desk lead supplied

1. Access confirmation from IT (from the IT manager, 29 October 2026)

RoleIn roleCan open the workflowReads the records it needsConfirmed on
Quoting coordinator44yes29 October 2026
Quotes desk lead11yes29 October 2026

2. The desk's work log for October (exported from the quoting system, 2 November 2026; 48 rows)

The work log has these headers: Case, Completed on, Prepared with workflow, Workflow runs, Note.

CaseCompleted onPrepared with workflowWorkflow runsNote
Q-21411 October 2026yes1
Q-21421 October 2026yes1
Q-21432 October 2026yes2dangerous goods
Q-21442 October 2026yes1
Q-21455 October 2026no0quoted on the phone
Q-21465 October 2026yes1
Q-21476 October 2026yes1
Q-21486 October 2026yes1dangerous goods
Q-21496 October 2026no0dangerous goods
Q-21507 October 2026yes1
Q-21517 October 2026yes2
Q-21527 October 2026yes1
Q-21538 October 2026yes1
Q-21548 October 2026yes1dangerous goods
Q-21559 October 2026no0dangerous goods
Q-21569 October 2026yes1
Q-215712 October 2026yes2
Q-215812 October 2026yes1
Q-215912 October 2026yes1
Q-216013 October 2026no0dangerous goods
Q-216113 October 2026yes1
Q-216213 October 2026yes1
Q-216314 October 2026yes1
Q-216414 October 2026yes2
Q-216514 October 2026yes1
Q-216615 October 2026yes1
Q-216715 October 2026no0dangerous goods
Q-216816 October 2026yes1
Q-216916 October 2026no0same as last week's quote, reissued
Q-217019 October 2026yes1
Q-217119 October 2026no0dangerous goods
Q-217220 October 2026yes2
Q-217320 October 2026no0dangerous goods
Q-217421 October 2026yes1
Q-217521 October 2026yes1
Q-217622 October 2026yes1
Q-217722 October 2026no0dangerous goods
Q-217823 October 2026yes1
Q-217923 October 2026no0dangerous goods
Q-218026 October 2026yes1
Q-218126 October 2026no0inland port, not on the rate sheets
Q-218227 October 2026no0dangerous goods
Q-218327 October 2026yes2
Q-218428 October 2026yes1
Q-218528 October 2026no0dangerous goods
Q-218629 October 2026yes1
Q-218730 October 2026no0dangerous goods
Q-218830 October 2026yes1

3. The reviewer's sample record (from the senior quoting coordinator, 30 October 2026; 16 drafts)

CaseResultReason
Q-2141as prepared or lightly editedsent as prepared
Q-2143substantially rewrittendangerous goods: no declaration surcharge, no packing questions
Q-2146as prepared or lightly editedadded the customer's reference
Q-2148substantially rewrittendangerous goods: declaration surcharge missing
Q-2150as prepared or lightly editedsent as prepared
Q-2154substantially rewrittendangerous goods: packing questions missing, surcharge missing
Q-2156as prepared or lightly editedsent as prepared
Q-2158substantially rewrittenpriced on Carrier B's old rate sheet
Q-2161as prepared or lightly editedchanged the collection date
Q-2164as prepared or lightly editedsent as prepared
Q-2166as prepared or lightly editedsent as prepared
Q-2170stoppedcustomer wants a fixed delivery date; only the desk lead commits to dates
Q-2172as prepared or lightly editedshortened the opening line
Q-2176as prepared or lightly editedsent as prepared
Q-2183as prepared or lightly editedsent as prepared
Q-2186as prepared or lightly editedsent as prepared

4. The business number (the desk lead's note)

Measure: quote requests answered with a complete quote within one working day of receipt. Counted from the received and sent dates in the quoting system, for requests received in the month. Better: up. October 2026: 39 of 46. August 2026, the baseline: 30 of 51. Counted the same way: yes. Target: 90%.

5. Notes from the desk's weekly meeting (October 2026)

9 October. Dangerous-goods requests: the draft leaves out the declaration surcharge and asks none of the packing questions. Three went out only after a full rewrite. For now the desk writes those by hand.

16 October. One coordinator on leave all week. Carrier B's new rate sheet started on 6 October; one draft still used the old rates and was caught at review.

23 October. Still writing dangerous-goods requests by hand: quicker than repairing the draft. Most other drafts go out with a line or two changed.

30 October. What the workflow misses, in the desk's words: dangerous goods first; requests for a fixed delivery date, which only the desk lead commits to; ports that are not on our rate sheets.

6. What else changed (the desk lead's note)

Carrier B's new rate sheet took effect on 6 October. One quoting coordinator was on leave from 15 to 21 October. The quote request form on the website gained a "goods type" field on 20 October.

7. Sheet settings

Workflow: quote request preparation. Team: the quotes desk. Role that runs it: quoting coordinator. Month: October 2026. Baseline: August 2026. Owner: quotes desk lead. Reviewer: senior quoting coordinator. IT: IT manager. First real case worked through as a team: 15 September 2026. Reading lines: the defaults. Earlier sheets: none; this is the first. Today's date: 2 November 2026.

What came back

AI ADOPTION MEASUREMENT SHEET: Quote request preparation · Quotes desk · October 2026
For: Quotes desk lead. Baseline: August 2026.
Prepared: 2 November 2026, from IT's access confirmation (29 October 2026),
the quoting system's work log for October (exported 2 November 2026, 48
rows), the reviewer's sample record (30 October 2026, 16 drafts), the desk
lead's note on the business number, the weekly meeting notes (9, 16, 23 and
30 October 2026), the note on what else changed, and the sheet settings.
This sheet reads the team's work as a whole. No person is counted, ranked or named.

1. INSTALLED: AVAILABILITY
Access: 4 of 4 quoting coordinators can open the workflow. 1 of 1 quotes
desk lead can open the workflow.
Records it reads: confirmed by the IT manager on 29 October 2026
First real case worked through as a team: 15 September 2026
Reading: INSTALLED

2. USED: REPEAT USE
Cases handled: 48. Prepared with the workflow: 34 (71%). Runs: 40.
By week: Week 1 (1 to 7 October) 10 of 12, 83% · Week 2 (8 to 14 October)
11 of 13, 85% · Week 3 (15 to 21 October) 6 of 10, 60% · Week 4 (22 to 31
October) 7 of 13, 54%
Direction: fell from week 1 to week 4, 29 points.
Not prepared with it: dangerous goods: 11 · quoted on the phone: 1 · same
as last week's quote, reissued: 1 · inland port, not on the rate sheets: 1
From the team's meetings: "Dangerous-goods requests: the draft leaves out
the declaration surcharge and asks none of the packing questions. Three went
out only after a full rewrite. For now the desk writes those by hand." (9
October) "Carrier B's new rate sheet started on 6 October; one draft still
used the old rates and was caught at review." (16 October) "Still writing
dangerous-goods requests by hand: quicker than repairing the draft." (23
October) "What the workflow misses, in the desk's words: dangerous goods first;
requests for a fixed delivery date, which only the desk lead commits to;
ports that are not on our rate sheets." (30 October)
Reading: BELOW the use line (70%) in week 4, FALLING.

3. USEFUL: QUALITY
Sample: 16 drafts, reviewed by the senior quoting coordinator.
Sent as prepared or lightly edited: 11 (69%) · Substantially rewritten: 4
(25%) · Stopped: 1 (6%).
Reasons for rewritten and stopped drafts:
- dangerous goods: declaration surcharge missing or packing questions
  missing: 3 (3 rewritten, 0 stopped)
- "priced on Carrier B's old rate sheet": 1 (1 rewritten, 0 stopped)
- "customer wants a fixed delivery date; only the desk lead commits to
  dates": 1 (0 rewritten, 1 stopped)
Not counted: none
Reading: BELOW the quality line (80%).

4. USEFUL: BUSINESS EFFECT
Measure: quote requests answered with a complete quote within one working
day of receipt. Source: received and sent dates in the quoting system, for
requests received in the month. Better: up.
October 2026: 39 of 46, 85%. August 2026: 30 of 51, 59%. Change: +26
points. Target: 90%.
This is a single month's reading, not a trend. It does not show what caused the change.
Reading: MOVED the better way against the flat line (5 points); target not
reached.

READ THE LINES TOGETHER
Everyone in both roles can open the workflow and the records are confirmed,
so access is not what the sheet points at. Use fell from 83% of cases in
week 1 to 54% in week 4, below the 70% line, and 11 of the 14 cases done
without the workflow carry the note "dangerous goods". The notes of 9 and
23 October say the desk writes those requests by hand; the 30 October note
lists dangerous goods first among the misses. Quality is below its line at 69%, and 3 of the 5
rewritten or stopped drafts are dangerous-goods requests. The business
number is 85% against 59% in the baseline month, short of the 90% target;
the sheet does not say why it moved.
Points at: the workflow

WHAT ELSE CHANGED
- Carrier B's new rate sheet, 6 October 2026: touches line 3 (one draft
  rewritten as priced on the old sheet)
- One quoting coordinator on leave, 15 to 21 October 2026: touches line 2
  (week 3, 15 to 21 October, handled 10 cases)
- "Goods type" field added to the website quote request form, 20 October
  2026

QUESTIONS FOR THE OWNER
1. The lines point at the workflow, where the notes say the draft misses
   dangerous-goods requests while the desk writes them by hand. What do you make of that?
2. Of the 7 requests received in October that were not answered within one
   working day, how many were dangerous-goods requests? The inputs do not
   say.
3. The meeting notes of 30 October name requests for a fixed delivery date
   and ports not on the rate sheets as misses. The sample holds one stopped
   draft for a fixed date and none for ports. How often do these arise?
   The work log does not say.
4. Carrier B's new rate sheet took effect on 6 October and one draft was
   priced on the old one. Was that the only such case? The sample is 16 of
   the 34 drafts prepared.
5. The "goods type" field arrived on 20 October. Should it be kept in view
   when November's sheet is read?

Status: Prepared for the owner, who decides. Nothing has been sent or changed.
Owner's decisions: [blank until decided]   Date: [ ]

The sheet prepares; the desk lead reads the four lines with the desk at its next weekly meeting and writes her own decisions on the last line.

Where it fits

It is the measuring part of step three, deploy, in how organizations plan their use of AI. The four lines are explained in installed, used or useful?.

  • Operations and process improvement, which usually owns the rollout of a team's first workflows and the monthly look at how they are doing.
  • Leadership, for owners who want a straight answer on what the AI they pay for is doing for one team.
  • Logistics and trade, where a quotes desk or a credit team handles a steady stream of similar cases and one turnaround number matters.
  • Professional services, where an intake lead or a billing team runs the same piece of work every day and a reviewer already reads what goes out.
  • Insurance, where submissions and claims arrive in volume and the desk that handles them can count what went through the workflow.
  • Manufacturing, where order entry and customer service teams work from one system that can export the month's cases.
  • Growth & scaling, when a first workflow is in use and the owner needs to know it is working before the next one starts.
  • Established businesses, where a team has done the work one way for years and the sheet shows whether the new way has taken hold.

What has to change for your company

Choose the one business number, write down how it is counted, and count the baseline month before the workflow starts. Decide how a case is marked as prepared with the workflow in the system the work is done in, and write your column names down once so every month reads the same. Name who reviews the sample and how many drafts they take each month. Set your own reading lines if the defaults do not fit: 70% of cases in the last week, a fall of more than 10 points, 80% of drafts sent as prepared or lightly edited, and a move of 5 points on the business number. It runs in ChatGPT, Claude, Copilot or Gemini, from files and notes you supply; it needs no connection to any system.

What has been checked

FullyPowered AI wrote this workflow template. It was editorially reviewed on 19 November 2026. It was tried once on its example on 19 November 2026; the record is published with it, and lists what that run found and what was fixed afterward. It has not yet been tested in a client environment. It has not yet been tested as a packaged version in ChatGPT, Claude, Copilot or Gemini.

Testing in your environment means preparing the sheet for two past months the owner already knows well, with hard cases written in advance: a log with a column naming who handled each case, renamed headers, a sampled case the log marks as not prepared, and a baseline counted from a different report. The owner recounts every figure before relying on the sheet. Pilot acceptance-test preparation sets those cases up, and workflow fit check is where the choice of work is made again when the lines point there.

Who decides

People decide

Your people keep

  • What changes after the owner reads the four lines
  • Which business number the workflow exists to move
  • Which drafts the reviewer samples
By design

It does not

  • Reads the team's work, never a person's use
  • Never explains the business number as caused by the workflow
  • Calls one month a reading, not a trend
  • Needs no connection to any system

Questions people ask

Who is this for?

The owner of one AI workflow in daily use by one team, such as a credit manager or an intake lead. They prepare the sheet after the first full month, then every month.

What do I need to supply?

The access confirmation from IT, the team's work log for the month, the reviewer's sample record, the one business number this month and in the baseline month, the team's meeting notes and what else changed.

Does it track individual people?

No. The sheet measures the team's work as a whole. Nobody is counted, ranked or named, and if a usage record by person is supplied, the sheet stops and asks for team-level counts.

Does it say the workflow caused the change?

No. The business number is reported beside its baseline, one month is called a reading, not a trend, and the sheet does not show what caused the change. The owner decides.

What has been checked?

It was editorially reviewed and tried once on an example. It has not yet been tested in a client environment.

The full workflow template

Everything above is the outline. The complete workflow template is the working version: the roles, the inputs, the procedure, the exceptions and stop conditions, the output format, the reviewer checklist, the acceptance tests, the instructions for an AI assistant and what to customize for your company. It comes with a download and the link to its folder in FullyPowered’s public repository.

View the full template

Free with an account. Sign in by email: no password and no card.

Next step

Put this workflow template to work for your team: on your plan, or built with you in a sprint.

  1. 01

    The full workflow template

    Free with an account. Sign in by email: no password and no card.

  2. 02

    FullyPowered Planner

    Add it to a workflow in the FullyPowered Planner, beside the teams and roles that will run it.

  3. 03

    Build this in a sprint

    In a sprint, FullyPowered adapts this workflow template to your rules, proves it on your own work and puts it in the hands of the roles that own it.

These workflow templates are examples and a starting point, not a complete list. We can build any workflow for your company, including one that runs several steps, in a Workflow Sprint or a scoped implementation.

Rather talk it through first? Book a fit callDiscuss adapting this

FullyPowered Planner. A card to start, everything unlocked, nothing charged for seven days.Security & your informationPlanner termsRefunds