
AI Agency & Freelancer Brief Builder
Writes the brief you send an agency, freelancer, or contractor for one named deliverable, and reviews what comes back against it.
Skills for Marketers
Connect what customers need with what your product can credibly promise. Organize research, positioning, launch decisions, and sales questions around the same evidence.
Collect customer language, buying triggers, alternatives, and proof of product value. Note the date and source of each claim so your team can check it later.
Draft an audience, a problem, a reason to choose the product, and the evidence behind that reason. Review the brief with sales and product before adapting it into launch materials.

Writes the brief you send an agency, freelancer, or contractor for one named deliverable, and reviews what comes back against it.

Chooses which slice of the ideal customer profile to win first, and says why not the others — with the market read that decides whether the window for each one is open, closing or already shut.

Scores a brand on six dimensions against rubric.md, comparing what a customer actually meets with what the company wrote down — then reports at most five findings, each with the evidence under it and what it is costing.

Writes the rules that make a brand usable by people outside marketing — a short list of prohibitions, a pointer to every rule that already exists in a context file, an owner and a trigger on each row, and a brief for whoever owns the visuals.

Tests whether the company's story about what changed in the world is true, and writes it where it is not — an argument with a dated turn, evidence for each beat, and one version short enough for a rep to say out loud.

Decides what the company stands for above any one product, and whether to change it — values that rule something out, a character description separate from how the company writes, and a rebrand recommendation with the cheapest honest version named.

Turns one competitor's entry in the competitor file into the page a salesperson actually uses — the questions to ask, the answer to that rival's usual claim, one customer who switched, and when to walk away — short enough to read in the ninety seconds before a call.

Builds the honest picture of each company the buyer is also considering — what they claim, who they beat you with, where they are genuinely better, where they are weak, what they charge, and what they have changed lately — and keeps it as the file every other skill checks a competitive claim against.

Works out which of the things the company has already published are worth a second life, splits them into pieces that stand up for somebody who never saw the original, and keeps a record of what has been used so nobody meets the same point for the third time.

Decides what the company will publish and what it will not — three to five pillars tested against the argument and against named rivals, a schedule cut to what the team can actually ship and approve, and a written list of the topics being declined.

Takes a draft and gives it back better — five passes against the company's own voice file, an unconditional check for claims the company has said it will not make, and either the rewritten text or a pass-by-pass review depending on what the copy has to survive.

Designs a round of customer research that will actually settle something — who to talk to, what to ask without leading them, how many is enough — and then reads what comes back with a confidence level on every finding.

Writes the announcement for something that shipped — the release note, the in-product line, the customer email and the sentence a salesperson can say out loud — from one pass over what changed.

Writes the company's ideal customer profile, its three to five buyer personas, the segments to disqualify with the signal that identifies each one early, and the buying committee for a typical deal.

Decides whether the company should be posting at all, on which platform, how often, and then writes the posts — with every platform claim marked by where it came from, so nobody acts on folklore without knowing it is folklore.

Scores a marketing function across six areas against a written rubric, then produces quick wins, risks, and three recommendations, with the evidence behind every score.

Writes the one-page brief a campaign runs from — one objective with a number in it, who it is for, the offer, the channels, the dates, what the campaign will deliberately not do, who picks up what it produces, and the point at which it stops whether or not it worked.

Reads everything a buyer can see — the website, the deck, the one-pagers, the emails, the ads — against the wording the company agreed, and reports where two surfaces say different things, which ones matter, and the exact edit for each.

Turns a settled position into the words each audience hears — one claim held constant, a version for every persona and every person on the buying committee, the support each of them accepts, and the thing not to say to each.

Turns a messaging argument into a test plan you can run in two to four weeks with no research budget, then reads the results and says what they do and do not license you to change.

Decides first whether a thing needs a name at all, then generates and screens candidates against criteria you agreed before seeing any of them.

Writes positioning.md — the category, the statement, three differentiators that survive a stress test, why now, and the objections that actually stall deals.

Examines what the company charges for rather than how much — the unit the price counts, how the tiers are separated, what the setup fee does to the package, and how any of it is explained to a buyer.

Sizes a launch against four signals, then builds the plan that size deserves — the story, the channels, the internal rollout, the dates and the review that closes it.

Diagnoses where marketing and sales are actually broken — lead definitions, the handoff, follow-up speed, feedback, shared numbers, and the joint meeting — then writes the stage definitions, the service-level agreement, and the pipeline review that fix it.

Works out which moment in a sale is going badly, audits what the company has already given the sales team, says what to throw away, and then builds at most two things — a deck in the order a buyer needs it, a one-pager of the right kind, or a demo narrative timed by scene.

Builds the list of what buyers actually push back on — in their own words, with what each one really means, the answer that has worked in your own calls, and a mark on anything borrowed from general practice rather than tested here.

Builds 12 to 15 prompts real buyers would type into an AI assistant, gives you a scoring sheet to run them yourself across four answer engines, then reads the results and names the five fixes worth making.

Works out why deals were actually won and lost, starting from the assumption that the reason in the CRM is wrong — with a verdict per deal, a pattern across them, and the evidence cited deal by deal.

Builds brand-voice.md from writing the company has already published — the voice in one line, the traits with do and don't examples, the word lists, and the formatting rules.

Builds and grows customer-language.md — what buyers actually said, word for word, about the problem, the result, the category, and the questions they ask.

Works out what your list is actually big enough to do, checks that mail is arriving at all, and then builds the sequences and writes the emails that the list supports — starting with the one that works at any size.

Builds the kit a launch actually needs: the size it deserves, the claim it rests on, the words each audience hears, the dated plan, what sales is given, the announcement, and the campaign — with every one of those gated on the size rather than produced because the box has a slot for it. Standalone playbook: required skills are sold separately; you must already own them. See the prerequisite list below.

Takes every claim the company makes and says, one at a time, whether anything it can actually show backs it up — with the verdict, the evidence named, and for anything thin a rewrite the evidence does carry.

Separates what the ad platform decides from what you still decide, then does the work on the five things you control — the creative, the inputs, the destination, what counts as a conversion, and who you refuse to bid on.

Re-establishes what a company is and is not, on evidence rather than on argument: interviews with real buyers, a verdict on why deals were actually won and lost, an honest read of the rivals, the rewritten position, a test of whether buyers believe the new claim, and the propagation of all of it into every file that quotes the old one. Standalone playbook: required skills are sold separately; you must already own them. See the prerequisite list below.

Sorts press and communications work into the one thing worth preparing before you need it, the things that wait for news, and the things a company without a comms person should not be doing — then writes whichever applies.

Takes a release that has already been judged newsworthy and gets it out of the building: the approval sequence with real dates, a quote somebody will actually print, the boilerplate, and whether to pay for distribution.

Sorts every SEO fix by who can actually make it, finds the pages competing with each other before anything gets optimised, and builds the plan from the queries your buyers actually type.
The essentials
Build a foundation with 46 skills for recurring marketing work.
One-time purchase
Coming soon · downloads being preparedThe bigger picture
Connect your workflow with all 62 skills and five playbooks.
One-time purchase
Coming soon · downloads being prepared