Capability template for marketing and content
Ninety-day marketing plan
An owner or first marketing hire gets the next ninety days of marketing on one plan: the stage the business is really in, the moves that fit it, who owns each one, the money and hours they use, and what is deliberately left undone. Every count comes from your own records, and no budget is invented.
The outcome
You stop guessing at marketing and start running it. The plan opens with where the business actually stands, read from your own results by a rule anyone can check: still launching, winning its first customers, selling repeatably, or selling and delivering without the founder. Everything after that fits the stage.
Each move says why it should work for your buyers, what number will show whether it is working, who does it and who owns it. A move whose result is not yet known for your business carries its stop rule, written today and never moved afterwards. The ninety days are laid out fortnight by fortnight, with the stuck items cleared first.
The plan is as honest about what it leaves out as about what it puts in. Tactics you considered and set aside are listed with the reason and what would bring them back. Costs nobody has measured stay marked unknown. The plan spends only the money you gave it, uses only the hours you gave it, and sends nothing: every note, post, call and payment is made by a person after the owner approves.
Who starts the work, and when
The owner, or the first marketing hire working for the owner, runs it at the start of a quarter, when the business changes stage, or when something it rests on moves: a new offer, a lost channel, a budget that grows or shrinks. A refresh run carries the last plan's stop rules forward unchanged and reports each one as written.
The owner, or whoever holds the budget, approves the plan and settles every open decision in it.
What they supply
- Your facts. A few lines: what you sell and to whom, how you differ, what you can and cannot say, the channels you will not use, the relationships you can draw on by role, who sells and who delivers, and the stage you believe you are in.
- Money and hours. Your monthly budget by line, or none; what you hold back for experiments; the hours each person gives to marketing and to selling.
- Your results log. A simple spreadsheet, one row per month per source: enquiries, conversations, proposals, new sales and repeat sales. A business with no sales yet starts without one.
- What you already have. Your website, sales material, lists and drafts, each with its status and anything blocking it.
- Evidence of demand, if you hold it. Dated customer notes, emails, a research summary.
- Ideas you are weighing, if any.
What comes out
One plan in twelve fixed sections: the plan in short; the frame, with three demand questions answered from your evidence; the current state, with nine surfaces scored from your materials; a section each for how buyers hear of you, how a conversation becomes a sale, how customers are kept and expanded, and how they bring others; money and hours; the ninety days by fortnight; what changes at the next stage; what is deliberately not done; and the measures, owners and open decisions.
A good plan can be checked line by line against the inputs, every move appears in a fortnight with its owner, and nobody on the team learns of a commitment for the first time when they read it.
An invented example
Illustrative example with invented data. The business, its customers, the accountants and every word quoted below are invented. Quillmoor Bookkeeping is an invented three-person firm that does monthly bookkeeping and payroll for independent cafés and restaurants. Its owner has just taken on a part-time marketing coordinator, the firm's first marketing hire, who runs the plan for the owner to approve.
The inputs come first, in full, so the template can be run on them. The plan follows.
The inputs
Your facts
Business: Quillmoor Bookkeeping, three people, started April 2026.
What we sell: monthly bookkeeping and payroll for independent cafés and
restaurants. Offer terms are on the service sheet.
Who buys: owner-operators of one or two sites.
How we differ: we reconcile card tips against payroll every pay run, and
every customer has a named bookkeeper.
What we can say: five customers so far. None has agreed to be named.
We will not: run paid ads this quarter, cold-call, or offer a discount
for signing up. Never call ourselves "cheap".
Relationships, by role: four outside accountants who have referred a
client or said they would; the events coordinator of the local
restaurant association, whose members are independent restaurant
owners.
Who sells and who delivers: the owner runs every first conversation and
every sale; the owner and the bookkeeper deliver the work.
Approver: the owner.
Our view of the stage: repeatable. "Referrals keep coming."
The results log is updated by the owner by the 3rd of each month for the
month before.
First day of the plan: Monday 16 November 2026.
Note from the owner: two people asked for something in writing after a
first conversation and I had nothing to send.
Money and hours
Budget: $600 a month.
Tools (website, email, scheduling): $120
Restaurant association membership: $40
Held for experiments: $200
Not committed: $240
Hours a week: owner, 4 on marketing and 6 on selling; coordinator, 10 on
marketing; bookkeeper, none.
Results log, exported on Friday 13 November 2026
Month,Source,Enquiries,Conversations,Proposals,Won,Repeat sales
2026-05,Owner's network,3,2,1,1,0
2026-06,Owner's network,2,2,2,1,0
2026-06,Website,1,0,0,0,0
2026-07,Accountant referral,2,2,1,1,0
2026-08,Owner's network,1,1,0,0,0
2026-08,Customer referral,1,1,1,1,0
2026-09,Website,2,1,0,0,0
2026-09,Accountant referral,1,1,1,0,0
2026-10,Accountant referral,2,2,2,1,0
2026-10,Customer referral,0,0,0,0,1
What you already have
Item | Status | Since | Notes
Website | live | 2026-04 | One page: "bookkeeping and payroll for
restaurants". Does not mention tips.
Service sheet | stuck | 2026-08 | Drafted. Blocker: the owner has not
decided whether it shows prices.
Welcome checklist | live | 2026-05 | Used with all five customers.
Newsletter | draft | 2026-10 | First issue drafted by the coordinator.
Blocker: no list of people who have agreed to receive it.
Conversation notes | live | 2026-05 | The owner's notes from first
conversations.
Results log | live | 2026-05 | Updated by the owner each month.
Restaurant association membership | live | 2026-09 | Next members'
mixer: Thursday 3 December 2026.
Outreach list or messages | none | - | -
Published content | none | - | Nothing published beyond the website.
Referral ask | none | - | The owner thanks accountants by email after a
referral; customers are not asked.
Evidence of demand
E1 | 14 July 2026 | Conversation notes, café owner, first conversation
"I do the books on Sunday nights and I'm always a month behind."
Note-taker's summary: her brother runs payroll for her.
E2 | 3 September 2026 | Email from an outside accountant
"Three of my restaurant clients still hand me a box of receipts at
year end."
E3 | 20 October 2026 | The owner's own view
Most owners we meet would switch for someone who sorts out tips and
payroll.
Ideas being weighed
1. A weekly social post (the owner's idea).
2. A monthly newsletter (the coordinator has drafted the first issue).
3. Paid search ads (suggested by a friend).
4. A stand at the regional food trade show.
5. A free tips-and-payroll checklist to download from the website.
The plan
NINETY-DAY MARKETING PLAN: Quillmoor Bookkeeping, Monday 16 November 2026 to Saturday 13 February 2027
Prepared for: the owner. Prepared Friday 13 November 2026 from: your facts; money and hours; results log, May to October 2026, 10 rows; what you already have, 10 items; evidence of demand, 3 items; ideas being weighed, 5 items.
First plan.
Stage: First customers. The log holds 5 new sales from May to October 2026, and no new sale was won in September, so the bar for Repeatable (a new sale in each of the last three full months: August, September, October) is not met. Your view: Repeatable. See open decision 2.
Not supplied: none.
1. IN SHORT
Quillmoor's five customers came through the owner's network and through referrals from accountants and a customer; the website brings enquiries, but none has reached a proposal. For the next ninety days the plan puts the 14 marketing hours a week the owner and the coordinator have into the routes that already produce sales, and finishes the written material buyers have asked for. What is at stake is the bar for Repeatable: a new sale in each of November, December and January.
Priorities for the ninety days:
1. Finish the service sheet, once the owner decides by Friday 20 November (proposed) whether it shows prices (M4)
2. Ask the four accountants and the restaurant association for introductions (M1, M2)
3. Call every current customer and ask each for a referral (M6, M7)
By day 90 we expect to know:
- Whether November, December and January each have a new sale, the bar for Repeatable, read in the log
- What keeping their own books costs buyers in hours a month, read in the conversation notes (M5)
- Whether the restaurant association produces enquiries, and whether the check-in calls and the customer ask bring repeat sales and referrals, read in the log on Thursday 4 February 2027 (M2, M6, M7)
2. FRAME
What we sell: monthly bookkeeping and payroll for independent cafés and restaurants.
Who buys: owner-operators of one or two sites.
How we differ: card tips are reconciled against payroll every pay run, and every customer has a named bookkeeper.
The problem, what it costs a buyer each time and how often it comes round (a belief to test): owners fall a month behind and do the books on Sunday nights (E1); receipts reach the accountant in a box at year end (E2). The books come round every month, and payroll at every pay run. Test: M5.
Demand questions:
Q1. Who, other than us, has shown they want this, and how?
Five owners bought between May and October (log). A café owner described the problem in her own words (E1), and an accountant sees it among their restaurant clients (E2).
Evidence: buyers who bought: Observed (log, May to October 2026). The problem: Observed (E1, 14 July 2026; E2, 3 September 2026).
Q2. How do buyers manage without us now, and at what cost?
They do the books themselves on Sunday nights while a relative runs payroll (E1), or hand their receipts to the accountant at year end (E2). No buyer has said what this costs in hours or money.
Evidence: how they manage: Observed (E1, 14 July 2026; E2, 3 September 2026). What it costs: None. Closed by M5.
Q3. Which buyers feel the problem hardest, described by role and kind of business?
The log shows that owners referred by an outside accountant reached more proposals than any other source: 5 conversations and 4 proposals, against 12 conversations and 8 proposals from all sources (log). No input shows which buyers feel the problem hardest. That tips and payroll are the reason they switch is the owner's view (E3).
Evidence: accountant-referred owners reaching a proposal: Observed (log, July to October 2026). Which buyers feel it hardest: None. Closed by M5. Tips and payroll as the reason: Reported (the owner, 20 October 2026). Tested by M5.
3. CURRENT STATE
What the log shows, May 2026 to October 2026: 15 enquiries, 12 conversations, 8 proposals, 5 new sales won, 1 repeat sale.
By source: Owner's network: 6 enquiries, 5 conversations, 3 proposals, 2 won. Accountant referral: 5 enquiries, 5 conversations, 4 proposals, 2 won. Website: 3 enquiries, 1 conversation, 0 proposals, 0 won. Customer referral: 1 enquiry, 1 conversation, 1 proposal, 1 won, 1 repeat sale.
New sales won by month: May 1, June 1, July 1, August 1, September 0, October 1.
Reported but not logged: none.
Already done: the website, since April 2026; the welcome checklist, used with all five customers since May 2026; the conversation notes, since May 2026; the results log, updated every month since May 2026; the restaurant association membership, since September 2026.
In flight: the newsletter, first issue drafted: blocker, no list of people who have agreed to receive it.
Stuck: the service sheet: blocker, the owner has not decided whether it shows prices; cost of leaving it, two people asked for something in writing after a first conversation and had nothing sent (owner's note); cleared by M4.
Surface scores (from the materials, not from results):
Positioning: 1. Written only in your facts; the difference (tips reconciled against payroll) is on no item a buyer sees.
Evidence of demand: 1. The owner's conversation notes (E1, E3) and one accountant's email (E2).
Website: 5. Live, and "Website" appears as a source in the log.
Sales material: 1. The service sheet, drafted and stuck.
Outreach: 0. Not yet started; no list or message exists.
Content: 1. The newsletter's first issue, drafted; nothing published.
Referrals: 0. Not yet started; no written ask exists, though referrals appear in the log.
Keeping customers: 4. The welcome checklist, used with all five customers.
Measurement: 3. The results log, updated every month since May 2026.
Shape: the work and the website are in place and referrals arrive without being asked for, but the material that asks for business (the service sheet, a referral ask, outreach) is not yet written.
4. REACH
Claim: buyers hear of Quillmoor from people they already trust, so the plan works through those people before building any audience of its own.
Now: 15 enquiries in six months: 6 from the owner's network, 5 from accountants, 3 from the website, 1 from a customer.
M1. A short personal note from the owner to each of the four accountants, asking whether a client of theirs keeps their own books and offering a first conversation. Why it would work: accountants produced 5 of 15 enquiries and 4 of 8 proposals (log), and one accountant sees the year-end problem among their own restaurant clients (E2). Indicator: enquiries with source "Accountant referral", read monthly in the log. Done by: the coordinator drafts; the owner approves and sends each note. Owner: the owner (4 hours a week on marketing, 6 on selling). Kill criterion: not an experiment.
M2. Attend the restaurant association's mixer on Thursday 3 December 2026 with the finished service sheet, and ask the events coordinator which members keep their own books. Why it would work: the association's members are independent restaurant owners (your facts), and the membership is already paid for; that they will enquire is a belief to test, since no enquiry has yet come from it. Indicator: enquiries with source "Association", a new source value, read monthly in the log. Done by: the coordinator prepares the sheets; the owner attends and asks. Owner: the owner (4 hours a week on marketing, 6 on selling). Kill criterion: fewer than 2 enquiries with source "Association" in the log for December 2026 and January 2027, read on Thursday 4 February 2027, after the log is updated for January on the 3rd; if so, stop attending and put the membership line to the owner.
5. CONVERT
Claim: first conversations already turn into proposals (accountant referrals: 5 conversations, 4 proposals; all sources: 12 conversations, 8 proposals, in the log); what is missing is something in writing to take away, and a reason on the website to ask for a conversation.
Now: 12 conversations, 8 proposals, 5 new sales won (log). Two people asked for something in writing and had nothing sent (owner's note). The website brought 3 enquiries and 1 conversation, and no proposal (log).
M3. Add to the website what makes Quillmoor different (tips reconciled against payroll every pay run; a named bookkeeper) and a way to ask for a first conversation. Why it would work: the website brings enquiries that have not reached a proposal (log), and the difference is on no item a buyer sees (current state). Indicator: conversations with source "Website", read monthly in the log. Done by: the coordinator drafts; the owner approves before it is published. Owner: the coordinator (10 hours a week on marketing). Kill criterion: not an experiment.
M4. Finish the service sheet once the owner decides whether it shows prices (open decision 1, by Friday 20 November 2026, proposed), and send it after every first conversation. Why it would work: two people asked for something in writing and had nothing to take away (owner's note). Indicator: proposals, read monthly in the log. Done by: the coordinator finishes it; the owner approves it and sends it. Owner: the owner (4 hours a week on marketing, 6 on selling). Kill criterion: not an experiment.
M5. In every first conversation, ask how many hours a month the books take today and whether tips and payroll are part of why they are looking, and write both answers, with the kind of business and the number of sites, in the conversation notes. Why it would work: it produces the evidence Q2 lacks, shows which buyers feel the problem hardest (Q3), and tests the owner's view in Q3. Indicator: first conversations with both answers recorded, counted at each monthly review in the conversation notes. Done by: the owner. Owner: the owner (4 hours a week on marketing, 6 on selling). Kill criterion: not an experiment.
6. KEEP AND EXPAND
Claim: a check-in after the first months keeps customers and finds the next piece of work for them, a belief to test; the welcome checklist starts each customer well, and nothing follows it yet.
Now: 1 repeat sale, in October (log). The welcome checklist is used with all five customers.
M6. A check-in call with each of the five current customers this quarter, and with each new customer in its third month: what is working, and what else they need. Why it would work: the one repeat sale so far came from an existing customer (log); that a call brings more is a belief to test. Indicator: repeat sales, read monthly in the log. Done by: the owner. Owner: the owner (4 hours a week on marketing, 6 on selling). Kill criterion: no repeat sale in the log for December 2026 and January 2027, read on Thursday 4 February 2027; if so, keep the calls only as the place for the M7 ask and review what customers said in them.
Offer terms: kept in the service sheet; not restated here.
7. REFER
Claim: referrals already bring work without being asked for; asking for them, and thanking the people who give them, should bring more, a belief to test.
Now: 6 of 15 enquiries came from referrals (5 from accountants, 1 from a customer), and 3 of the 5 new sales (log). No written ask exists.
M7. At each check-in call (M6), the owner asks the customer whether they know another owner who keeps their own books, and every referrer is thanked in writing. A customer is named as a reference only with their written permission. Why it would work: one customer referral led to a new sale in August (log). Indicator: enquiries with source "Customer referral", read monthly in the log. Done by: the coordinator drafts the thank-you note; the owner asks, approves and sends. Owner: the owner (4 hours a week on marketing, 6 on selling). Kill criterion: no enquiry with source "Customer referral" in the log for December 2026 and January 2027, read on Thursday 4 February 2027; if so, stop asking at the calls and review what customers said.
8. MONEY AND HOURS
Budget, as you gave it: $600 a month.
Tools (website, email, scheduling): $120 a month. Pays for M3.
Restaurant association membership: $40 a month. Pays for M2.
Held for experiments: $200 a month. Not spent: of the three experiments, M2 runs on the membership line, and M6 and M7 need no money.
Not committed: $240 a month. Spent on nothing until a move earns it.
Hours, as you gave them: the owner, 4 a week on marketing and 6 on selling; the coordinator, 10 a week on marketing; the bookkeeper, none. Every move is owned by someone with hours. The plan does not estimate how long a move takes.
Unknown until recorded: the cost of an enquiry and the cost of a sale: the log records counts by source but no spend; a monthly spend figure for each source would make both known (open decision 3).
9. THE NINETY DAYS
Fortnight 1, Monday 16 November to Sunday 29 November 2026 (clear what is stuck):
M4: the owner decides by Friday 20 November (proposed) whether the sheet shows prices; the coordinator finishes it; the owner approves it. Owner: the owner. Indicator: the sheet approved.
M1: the coordinator drafts the four notes; the owner approves and sends them. Owner: the owner. Indicator: notes sent, 4.
M5: the owner starts asking the two questions in every first conversation. Owner: the owner. Indicator: answers recorded.
M2: the coordinator adds "Association" as a source value in the log. Owner: the owner. Indicator: the value in use.
Fortnight 2, Monday 30 November to Sunday 13 December 2026:
M2: the owner attends the mixer on Thursday 3 December with the service sheet. Owner: the owner. Indicator: enquiries with source "Association".
M3: the coordinator drafts the website change; the owner approves it before it is published. Owner: the coordinator. Indicator: the change published.
M4: the owner sends the sheet after every first conversation. Owner: the owner. Indicator: proposals.
Fortnight 3, Monday 14 December to Sunday 27 December 2026:
M6: the owner books and holds check-in calls with the five current customers. Owner: the owner. Indicator: calls held.
M7: the owner asks for a referral at each call; the coordinator drafts the thank-you note. Owner: the owner. Indicator: referrals asked for.
Fortnight 4, Monday 28 December 2026 to Sunday 10 January 2027:
M6: the owner finishes the check-in calls. Owner: the owner. Indicator: calls held, 5.
M1: the owner follows up with any accountant who has not replied. Owner: the owner. Indicator: enquiries with source "Accountant referral".
Fortnight 5, Monday 11 January to Sunday 24 January 2027:
M5: the owner counts the first conversations with both answers recorded, at the monthly review. Owner: the owner. Indicator: conversations with both answers.
M3: the coordinator checks conversations with source "Website" in December's log. Owner: the coordinator. Indicator: conversations from the website.
Fortnight 6, Monday 25 January to Sunday 7 February 2027:
M2: the owner reads the kill criterion on Thursday 4 February: enquiries with source "Association" in December and January. Owner: the owner. Indicator: the count, against 2.
M6: the owner reads the kill criterion on Thursday 4 February: repeat sales in December and January. Owner: the owner. Indicator: the count, against 1.
M7: the owner thanks in writing every referrer from the quarter, and reads the kill criterion on Thursday 4 February: enquiries with source "Customer referral" in December and January. Owner: the owner. Indicator: notes sent; the count, against 1.
Days 85 to 90, Monday 8 February to Saturday 13 February 2027: read the log for November, December and January against the bar for Repeatable; refresh the plan.
10. THE NEXT STAGE
M1: at Repeatable, grows: a note to the accountants every quarter, because they are a source that has produced sales. Beyond the founder, unchanged, sent by whoever runs first conversations.
M2: at Repeatable, continues only if its kill criterion was not met: the count decides. Beyond the founder, unchanged, attended by whoever runs first conversations.
M3: at Repeatable, unchanged: the website still carries the difference. Beyond the founder, unchanged.
M4: at Repeatable, unchanged: the sheet goes after every first conversation. Beyond the founder, unchanged, sent by whoever runs first conversations.
M5: at Repeatable, stops once the answers settle Q2 and Q3. Beyond the founder, stopped.
M6: at Repeatable, grows if its kill criterion was not met: a third-month check-in for every customer as a routine. Beyond the founder, unchanged, held by the customer's named bookkeeper.
M7: at Repeatable, grows if its kill criterion was not met: the ask becomes part of every check-in. Beyond the founder, unchanged, asked by the customer's named bookkeeper.
11. DELIBERATELY NOT DONE
Paid search ads: your facts rule out paid ads this quarter. Reopen when: you lift that rule and the log records spend by source.
Cold calling: your facts rule it out. Reopen when: not reopened by this plan.
A discount for signing up: your facts rule it out. Reopen when: not reopened by this plan.
A weekly social post: nothing in the log shows buyers finding Quillmoor this way, and the coordinator's hours go to M1, M2, M3, M4 and M7. Reopen when: those moves are done and the coordinator has hours left.
A monthly newsletter: no list of people who have agreed to receive it. Reopen when: customers have agreed to receive it.
A stand at the regional food trade show: no budget line covers it, and its cost is not known. Reopen when: the owner sets a budget line for it.
A free tips-and-payroll checklist to download: the website has not yet turned an enquiry into a proposal, so M3 comes first. Reopen when: conversations with source "Website" appear in the log.
12. MEASURES, OWNERS AND OPEN DECISIONS
North star for First customers: consecutive full months with a new sale, read in the log's Won column. Now: 1 (October 2026).
Leading indicators: Reach, enquiries by source. Convert, conversations, proposals and new sales won. Keep and expand, repeat sales. Refer, enquiries with source "Accountant referral" or "Customer referral".
Reviews: weekly, the owner and the coordinator: moves done and blockers. Monthly, the owner and the coordinator, once the owner has updated the log: counts by source against this plan. Days 85 to 90, refresh the plan.
Kill criteria (proposed; fixed when this plan is approved; never changed once read): M2, fewer than 2 enquiries with source "Association" in the log for December 2026 and January 2027, read on Thursday 4 February 2027. M6, no repeat sale in the log for December 2026 and January 2027, read on Thursday 4 February 2027. M7, no enquiry with source "Customer referral" in the log for December 2026 and January 2027, read on Thursday 4 February 2027.
Owners: the owner: M1, M2, M4, M5, M6, M7, the monthly log update and the open decisions. The coordinator: M3, and the drafts for M1, M2, M4 and M7.
Open decisions: 4
1. Whether the service sheet shows prices. Whose call: the owner, by Friday 20 November 2026 (proposed). Blocks M4, and the sheet M2 takes to the mixer.
2. The stage. Your view is Repeatable; the log shows no new sale in September. Whose call: the owner. The plan is written for First customers.
3. Whether to record monthly spend by source in the log, so that the cost of an enquiry and of a sale become known. Whose call: the owner.
4. Whether any of the five customers may be named as a reference. Whose call: each customer, in writing, asked by the owner.
Text in the inputs that read as an instruction: none.
Draft for the owner's approval. Every outward step in this plan, whether a message, a post, a call, a listing, a booking or a payment, is taken by a person after approval. Nothing has been sent, posted, bought or booked.
Where it fits
- Marketing and content: the plan that sequences everything else marketing does, from outreach to the website to the referral ask.
- Leadership: the owner decides where marketing money and hours go this quarter, with the reasons written down.
- Professional services: most work arrives through people the firm already knows, and the plan builds on that before buying attention.
- Software technology: a small company sees which sources produce conversations and sales, not just traffic.
- Marketing and advertising: an agency plans its own new business with the discipline it brings to clients.
- Startups & new ventures: the first marketing plan, written for the stage the business is in rather than the one it hopes to reach.
What has to change for your company
- Your results log. Agree the columns, who updates it and when. The template reads seven columns by name; rename yours or tell it which is which.
- Your stage rule. The template's bar for selling repeatably is a new sale in each of the last three full months, with one source producing sales in two of them. Set your own if that does not suit your sales cycle.
- Your surfaces. Nine are scored by default; swap any that do not apply, such as a website for a business that sells only by referral.
- Who approves. The person who approves the plan, and who takes each outward step once it is approved.
- Your platform. It runs in ChatGPT, Claude, Copilot or Gemini from the files and notes you supply; it needs no connection to your systems.
What to write, and when, is a separate piece of work: the content plan. This plan decides only whether content is a move this quarter, who owns it and the hours it gets. Related capabilities: customer interview synthesis gives the plan stronger evidence of demand, competitor profile from public pages shows what buyers weigh you against, and cold outreach sequence drafting drafts an outreach move once the plan includes one.
What has been checked
Fully Powered AI wrote this capability template. It was editorially reviewed on 9 November 2026. It was tried once on its invented example in Claude on 9 November 2026; the record is published with it, lists what the run found, and lists the small edits made afterward that have not been run again. 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 a past quarter your team remembers, plus hard cases written in advance: a log whose columns do not match, a stage you believe in that the log does not show, an idea that needs money the budget does not hold. The owner checks every count against the log before anyone relies on the plan. Pilot acceptance-test preparation sets those cases up, and what we test before we call an AI workflow ready describes the record we keep.
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
Bring a real piece of this work to a call. We’ll walk through what it takes to fit this capability to your rules, prove it on your work and put it in the hands of the roles that own it.