Capability template for project and delivery
Client findings validation document
Before research becomes recommendations, the engagement lead sends the client one document: each material finding beside its evidence, the implication proposed and a clear request to confirm, correct or defer it. Nothing is added that the evidence does not show.
Listed under
Add it to a workflow in the Fully Powered Planner, beside the teams and roles that will run it. Book a fit call
The outcome
The client sees what the team found, what it rests on and what the team proposes to do about it, in one place and in a fixed shape. Each finding ends in a plain request: confirm it, correct it, or defer it. Evidence that points the other way sits beside the finding, and anything the team has assumed and not yet shown is written down under it.
The recommendations are then built on findings the client has actually answered. A finding the client defers is not counted as agreed, and the work that depends on it waits. The engagement lead checks every item of evidence against its source, decides what goes to the client, and decides what to do with each answer.
Who starts the work, and when
The engagement lead, or a consultant working for them, starts it when a workstream has finished its research and interviews and is about to turn them into recommendations. It works best a few days before the findings are presented, while there is still time to chase a missing source or reword a claim the evidence cannot carry.
The client sponsor, or whoever the decision brief names as the person who decides, receives the document once the engagement lead has approved it.
What they supply
- The decision brief: the engagement, the decisions the client has to make, and who decides each.
- The evidence register: one row for each interview, data extract, document or observation, with its date, who it came from, what it says, its limits and whether it may be shown to the client.
- The candidate findings: one row for each, with the evidence behind it, any evidence against it, what is assumed, the implication the team proposes and whether it is material.
- Today's date and the date the client should reply by.
- Optional settings: which words count as strong quantity words, how many interviews must stand behind them, and the wording of the three replies. Defaults are built in.
- For the second run only: the client's written replies, by finding, with who gave them and when, together with the same decision brief, register and findings again.
The capability reads only what is supplied. It does not search, contact anyone or open a system.
What comes out
One document in three parts. The first part is for the client: each material finding with its evidence, its counter-evidence, its open assumptions, the implication proposed and a response request with three boxes. The second part is for the engagement lead only, to remove before anything is sent: findings held back and why, wording the evidence does not carry, withheld items, figures that were checked, and a numbered list of decisions. The third part is written after the client replies, and records each answer in the client's own words.
An invented example
Illustrative example with invented data. The firm, the client, the people, the evidence and every figure below are invented. Firm A is an invented advisory firm. Client B is an invented regional distributor of garden machinery parts with about 140 staff, and Firm A has been reviewing why orders leave its main warehouse late. First come the inputs the engagement lead supplied, in full, then the document the capability returned.
The inputs
1. Decision brief
Engagement: Order-to-dispatch review for Client B
Document goes to: Operations director, Client B. From: Engagement lead, Firm A.
Today: 17 November 2026. Reply by: 27 November 2026.
Decisions the client has to make:
D1 Whether to trial zone-sorted pick lists at the main warehouse. Decides: Operations director.
D2 Whether to trial a 16:00 same-day cut-off. Decides: Operations director.
D3 Whether to add a second returns bench. Decides: Operations director.
2. Evidence register
ref,source,kind,date,from_whom,what_it_says,limits,share
E1,Interview 1,interview,2026-10-12,Warehouse supervisor,"Pick lists come out sorted by order number. On mixed orders pickers walk to aisle C, then aisle F, then back to aisle C.",One person's account.,yes
E2,Interview 2,interview,2026-10-13,Picker (morning shift),Confirmed the order-number sorting and the walk back to the first aisle.,One person's account.,no
E3,Interview 3,interview,2026-10-14,Customer service lead,Said late-order calls come mainly from customers who ordered parts from more than one zone.,One person's account; no call log seen.,yes
E4,Dispatch export,data,2026-10-30,Client warehouse system extract,"September 2026: 2,310 orders. 1,184 had lines in more than one zone, median 31 hours from order to dispatch. 1,126 had lines in one zone, median 19 hours.",One month; timestamps as recorded by the system.,yes
E5,Floor timing,observation,2026-10-15,Firm A consultant,"Timed 12 mixed-zone orders and 12 single-zone orders at the morning pick. Mean walking time 9 minutes for mixed, 4 minutes for single.",One morning.,yes
E6,Interview 4,interview,2026-10-14,Dispatch lead,"Once an order is picked, packing takes under 30 minutes, so an order placed by 16:00 could make the 16:30 courier.",One person's account.,yes
E7,Dispatch export,data,2026-10-30,Client warehouse system extract,"September 2026: 412 orders placed between 14:00 and 16:00. 97 dispatched the same day, 315 the next working day.",One month.,yes
E8,Interview 5,interview,2026-10-15,Picker (afternoon shift),"Said the afternoon pick team stops at 15:30, so later orders wait for the next day whatever the cut-off.",One person's account.,yes
E9,Returns log,document,2026-10-28,Client returns clerk,September 2026 returns log lists 188 returned items.,Counts only items logged at the bench.,yes
3. Candidate findings
finding_id,finding,evidence_refs,counter_refs,assumptions,proposed_implication,material,decision
F1,"Orders with lines in more than one zone take longer to dispatch, and the pick-list sort order is a likely cause.",E1;E2;E4;E5,,September is typical of other months.,Sort pick lists by zone at release and trial it at the main warehouse for four weeks.,yes,D1
F2,Most late-order calls come from customers with mixed-zone orders.,E3,,,Prioritise delay messages for mixed-zone orders.,yes,
F3,"Orders placed between 14:00 and 16:00 could ship the same day if the cut-off moved to 16:00.",E6;E7,E8,The 16:30 courier collection stays as it is.,Trial a 16:00 cut-off for two weeks.,yes,D2
F4,The returns queue at the single bench is longest on Mondays.,E10,,,Add a second returns bench on Mondays.,yes,D3
F5,Pallet wrap use is higher on mixed-zone orders.,E5,,,None.,no,
The output
CLIENT FINDINGS VALIDATION DOCUMENT: Order-to-dispatch review for Client B
PART A. FOR THE CLIENT
For: Operations director, Client B. From: Engagement lead, Firm A. Prepared 17 November 2026. Please reply by 27 November 2026.
What this is: 3 findings from the research on Order-to-dispatch review for Client B, each with the evidence under it, what is assumed and the implication we propose. For each, please say whether you confirm it, want to correct it, or want to defer it. A deferred finding is not agreement, and work that depends on it waits.
Finding F1: Orders with lines in more than one zone take longer to dispatch, and the pick-list sort order is a likely cause.
Evidence behind it: 4 items, 3 kinds; 1 not shown, marked not to be shared.
E1 · interview · 12 October 2026 · Warehouse supervisor: Pick lists come out sorted by order number. On mixed orders pickers walk to aisle C, then aisle F, then back to aisle C. Limits: One person's account.
E4 · data · 30 October 2026 · Client warehouse system extract: September 2026: 2,310 orders. 1,184 had lines in more than one zone, median 31 hours from order to dispatch. 1,126 had lines in one zone, median 19 hours. Limits: One month; timestamps as recorded by the system.
E5 · observation · 15 October 2026 · Firm A consultant: Timed 12 mixed-zone orders and 12 single-zone orders at the morning pick. Mean walking time 9 minutes for mixed, 4 minutes for single. Limits: One morning.
Evidence pointing the other way: none supplied
Not yet shown: September is typical of other months.
Implication we propose: Sort pick lists by zone at release and trial it at the main warehouse for four weeks. This bears on D1, whether to trial zone-sorted pick lists at the main warehouse.
Your response: [ ] Confirm [ ] Correct: what is wrong or missing: ______ [ ] Defer: what you need, and by when: ______
Finding F2: Most late-order calls come from customers with mixed-zone orders.
Evidence behind it: 1 item, 1 kind; single source.
E3 · interview · 14 October 2026 · Customer service lead: Said late-order calls come mainly from customers who ordered parts from more than one zone. Limits: One person's account; no call log seen.
Evidence pointing the other way: none supplied
Not yet shown: none stated
Implication we propose: Prioritise delay messages for mixed-zone orders. No decision named.
Your response: [ ] Confirm [ ] Correct: what is wrong or missing: ______ [ ] Defer: what you need, and by when: ______
Finding F3: Orders placed between 14:00 and 16:00 could ship the same day if the cut-off moved to 16:00.
Evidence behind it: 2 items, 2 kinds.
E6 · interview · 14 October 2026 · Dispatch lead: Once an order is picked, packing takes under 30 minutes, so an order placed by 16:00 could make the 16:30 courier. Limits: One person's account.
E7 · data · 30 October 2026 · Client warehouse system extract: September 2026: 412 orders placed between 14:00 and 16:00. 97 dispatched the same day, 315 the next working day. Limits: One month.
Evidence pointing the other way: E8 · interview · 15 October 2026 · Picker (afternoon shift): Said the afternoon pick team stops at 15:30, so later orders wait for the next day whatever the cut-off. Limits: One person's account.
Not yet shown: The 16:30 courier collection stays as it is.
Implication we propose: Trial a 16:00 cut-off for two weeks. This bears on D2, whether to trial a 16:00 same-day cut-off.
Your response: [ ] Confirm [ ] Correct: what is wrong or missing: ______ [ ] Defer: what you need, and by when: ______
This document is a draft for the engagement lead to verify. Nothing has been sent.
PART B. FOR THE ENGAGEMENT LEAD ONLY. REMOVE BEFORE SENDING.
Findings supplied: 5. Put to the client: 3. Held back: 1. Not material: 1. Flagged: 5.
Register: 9 items. Prepared 17 November 2026 from decision brief, register, findings.
Flags:
F1: evidence withheld (E2).
F2: single source; wording stronger than the evidence (the finding says "Most"; one counted interview, no counted data item).
F3: counter-evidence present (E8).
F4: reference not in register (E10); held back, no evidence.
F5: item does not address the finding (E5 gives walking times and says nothing about pallet wrap).
Withheld from the client copy: E2, used by F1.
Items that do not address the finding they are cited for: E5, cited by F5.
Figures checked: E4: 1,184 + 1,126 = 2,310, matches. E7: 97 + 315 = 412, matches. No other item gives a total.
Evidence no finding uses: E9.
Not put to the client: F4 (held back, reference not in register); F5 (marked not material).
Decisions for the engagement lead before sending: 5
1. F1, evidence withheld. E2 is marked not to be shared and is not shown, so F1 goes to the client with 3 of its 4 items visible. Does F1 go as it stands?
2. F2, single source and wording. The finding says "Most" and rests on one interview, E3, with no call log seen. Do you reword F2, add evidence, or leave it out of the client list?
3. F3, counter-evidence present. E8 is shown beside E6 and E7, and one assumption about the courier is listed as not yet shown. Does F3 go with both sides shown?
4. F4, reference not in register. It cites E10, which is not in the register. E9, the returns log, is the only item on returns, says nothing about Mondays, and no finding cites it. Which reference is meant, if any?
5. F5, not material, and no cited item addresses it. Does F5 stay out, or is there evidence to add?
PART C. CLIENT REPLIES
Not run: no replies supplied.
May be carried into the recommendations: none
Lead to revisit before use: none
Held: none
Not put to the client: F4, F5
The assistant has changed no finding and no record. Nothing has been sent.
Where it fits
- Professional services, where the lead has to be sure the client agrees with the findings before the team builds on them.
- Engineering firms, where a study's findings are put to the owner before design or scope follows from them.
- Marketing and advertising, where audience and market research is checked with the client before a plan is built on it.
- Project and delivery, which owns the engagement plan and the gate before the team proceeds.
- Customer success and account management, which keeps the client's sponsor confident that nothing reaches them as a surprise.
- Established businesses, where the client has a view of its own operations that the team needs to hear.
- Growth and scaling, when more engagements run at once and the checking cannot rest on one lead's memory.
It is separate from customer interview synthesis, which draws themes from a stack of customer notes, and from research summary brief, which summarizes documents for a reader. This one takes findings that already exist and puts them to the client.
What has to change for your company
Decide who counts as the client's decider for each kind of finding. Set what makes a finding material in your engagements, and whether findings from workshops count as evidence. Choose how many interviews you want behind a claim that uses a word such as most or usually. Fix the wording of the three responses to match how your clients reply, and the number of days they have. Agree what you will not show a client, such as an interviewee who asked not to be passed on, and how that is marked in the register.
What has been checked
Fully Powered AI wrote this capability template. It was editorially reviewed on 10 November 2026. It was tried twice on its invented example on 10 November 2026 in Claude (the second run was not independent), and the record of the second is published with it. 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 running it on an engagement you have already finished, with cases written in advance: a finding with one source behind it, a claim worded more strongly than its evidence, evidence that contradicts the finding, an interview marked not to be shown. The engagement lead checks every item against its source before anything is sent. Pilot acceptance-test preparation sets those cases up.
The full capability template
Everything above is the outline. The complete capability 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 Fully Powered’s public repository.
Free with an account. Sign in by email: no password and no card.
Next step
Put this capability to work for your team: on your plan, or built with you in a sprint.
Build this in a sprint
In a sprint, Fully Powered adapts this capability template to your rules, proves it on your own work and puts it in the hands of the roles that own it.
Two ways to begin: