Name the client and their URL, and get back a brand story built only from what their own public record proves, plus the questions that record cannot answer and a ledger of every line nobody should publish yet.
Name the client and their URL, and Juma reads their own published record first: the story pages, the newsroom, the filings and the founder material. Every date and every claim comes back attached to the source that carries it, and the things the public record genuinely cannot settle are listed as questions rather than filled in with plausible invention. Nothing is uploaded and no account access is needed.
What comes back is a branded PDF story bible plus an editable CSV claims ledger: the narrative spine, a sourced timeline, the origin-story shapes competitors already occupy, one canonical narrative with short forms that each stand on their own, cuts for the About page, a founder post, a pitch deck, a sales intro and press boilerplate, and a ledger that marks every superlative and every green claim before it reaches a release. Across 400+ marketing teams, the failure we see most in brand stories is not bad writing. It is a beautiful paragraph nobody can source when legal asks.
Get a category narrative map before a word of the story gets written: five or more competitors, each with a verbatim sentence quoted from their own about or story page, the URL it came from, and the story shape that sentence belongs to. The map then counts how many of them share each shape, so a claim like "everyone in this category tells the founder-frustration story" arrives as evidence rather than as an opinion. What comes out of it is the territory nobody has taken and a short list of the moves to avoid, which is the difference between a story that sounds right and a story that sounds like three other brands.
Map the origin stories this client's category is already telling. Take at least five competitors, quote one verbatim sentence from each of their own story pages with the URL, label the shape it belongs to, and count how many share each shape. Then name the territory nobody has taken.
Ask for the interview guide and the gaps in the public record turn into a booked conversation. The published material can prove that two people started a company in a given year, and it almost never proves why, what nearly stopped them, or which decision went the other way at the last minute. Those are the parts a reader remembers. The guide comes back grouped in asking order, with a follow-up under each question for when the first answer is a press-release answer, and a total time budget so it fits a real calendar slot. Every question is tied to a specific blank in the timeline rather than being generic founder curiosity.
Turn the gaps in the timeline into a founder interview guide: questions grouped in asking order, each tied to the specific blank it fills, a follow-up under each for when the first answer is a press-release answer, and a total time budget.
Get the claims ledger as an editable CSV with one row per claim: the exact wording, the source that carries it, a status of company-stated, filing-stated or third-party corroborated, the risk type, the jurisdiction it is risky in, a named approver role and a review-by date. The risk types stay separate on purpose, because a thin citation, an unproven category first and an environmental claim are three different problems with three different fixes. Any superlative comes back marked as unpublishable until someone produces the substantiation, and the ledger says what that substantiation would have to be. This is the file the comms lead and counsel work from.
Build the claims ledger as a CSV: one row per claim with the wording, its source, a status of company-stated, filing-stated or third-party corroborated, the risk type, the jurisdiction, a named approver role and a review-by date. Mark every superlative as unpublishable until substantiated and say what would substantiate it.
Run the contradiction check and get a table that puts each new line beside the live copy it replaces or disagrees with, the URL that copy sits on, and a call on whether the difference is a deliberate move or an accident nobody noticed. A brand story rarely lands on an empty page. It lands on top of an About page written three years ago, a boilerplate the PR agency keeps reusing and a careers page that describes a different company. Finding those collisions before publication is what stops a journalist finding them afterwards, and the table doubles as the update list for whoever owns each page.
Check the new story against what this client's live pages say today. Put each new line beside the existing copy it replaces or contradicts, give the URL, and say whether the difference is deliberate or an accident. Turn it into an update list by page owner.
Ask for the channel cuts and each one arrives with the constraint it has to live inside: the word or character limit, where it gets published, and the existing copy it replaces. The pitch-deck version comes back as a numbered narrative arc with one beat per slide and the point each beat has to land, not as a paragraph somebody has to break up later. The short forms are written as separate pieces rather than as the long version with words removed, so the 25-word line can open somewhere completely different from the 500-word one. Every copy block stays paste-clean, with the warnings kept beside the text instead of buried inside it.
Cut the story for each channel we publish on, with the limit, the placement and the copy it replaces attached to every cut. Write the pitch-deck version as one beat per slide. Write the short forms as their own pieces rather than trimming the long one, and keep every block paste-clean.
Juma projects work per client rather than per task. The team builds one project for each client, adds context to it over time, and every Flow run inside it starts from that shared ground. For a brand story the context that changes the output most is the material the open web does not carry: what the founders actually said, what the client is allowed to claim, and what their pages say today. If the project already exists, these are additions to it rather than a fresh setup.
Founder and Origin Notes
Interview transcripts, the real origin, the decision that nearly went the other way, the first customer. With this in the project the narrative spine comes from the room rather than from the press release, and the gap list shrinks to what is genuinely still open.
Claims and Legal Constraints
The claims already approved, the markets the client publishes in, and who signs off on what. This is what turns the ledger from a general risk template into the specific review their counsel asked for.
Live Copy Inventory
The current About page, boilerplate, careers copy and sales deck narrative. The contradiction check compares the new story against these, so collisions read as a concrete update list rather than a guess about what is out there.
Brand Voice Guide
How the client sounds, the words they use and the ones they avoid. Every channel cut comes out in the right register from the first draft, which matters most on the founder post, where a wrong register reads as ghostwritten.
Each file in a project can carry a short description in the project info field. Those descriptions tell Juma what a file holds and when it matters, so the right context gets picked up for the task at hand rather than everything at once.
It builds the whole defensible layer: a sourced timeline from the client's own pages, newsroom and filings, the category narrative map with competitors quoted verbatim, the canonical narrative and every channel cut, and the full claims ledger. What it will not do is invent the human parts the record does not carry.
Those parts come back as an ordered interview guide instead, each question tied to the specific blank it fills. That is a deliberate boundary. The quickest way to get a brand story rejected by the people who lived it is to hand them a paragraph about a motivation they never had, and a plausible invention is harder to spot than an obvious one.
Teams usually run it twice for this reason: once to find out what is missing and book the conversation, then again with the notes in the project.
One row per claim in an editable CSV: the exact wording, the source behind it, a status of company-stated, filing-stated or third-party corroborated, the risk type, the jurisdiction, a named approver role and a review-by date. It opens in any spreadsheet and goes straight to whoever signs the copy off.
The risk types stay separate rather than collapsing into one column, because a weak citation, an unproven category first and an environmental claim need three different fixes. Superlatives come back marked as unpublishable until somebody produces the substantiation, with a note on what that substantiation would have to be.
The review-by date matters more than it sounds. A claim that was true at the last funding round can quietly stop being true, and the ledger is what makes that visible on the next run.
An About page is one destination. This Flow produces the narrative itself and then cuts it for everywhere the story gets told out loud: the About page, a founder post, the pitch-deck arc, a 60-second sales intro, the careers page and the press boilerplate, each written to its own constraint.
The other difference is what sits underneath. A page gets written once and published. A story bible carries the spine, the evidence, the gaps and the ledger, so the next person who needs to tell the story in a new format starts from the same source of truth instead of paraphrasing the website again.
Yes. The method starts from whatever public record exists rather than from a consumer-brand template, so it works the same on a B2B software company, a services firm or an industrial supplier. The sources shift toward filings, funding announcements, conference talks and customer case studies, and the channel cuts shift with them.
For a B2B brand the pitch-deck arc and the sales intro usually carry more weight than the About page, because the story gets told in a room far more often than it gets read on a site. The category narrative map tends to be sharper too, since B2B about pages cluster around a very small number of shapes.
A narrative workshop plus the write-up is usually two to three weeks of elapsed time, most of it spent gathering material and chasing sources. Running this Flow gets the evidence base, the category map, the narrative, the cuts and the ledger back in one working session, with the time that remains going into the interview and the judgment calls.
Those calls are the part that stays with the team. Which tension the brand is willing to tell in public, how much of the founder's version survives contact with the facts, and whether a claim is worth defending are decisions a person makes. Strategy, taste and judgment stay human, and nothing in the bible reaches a release without a named person signing it.
The workshop does not disappear either. It gets shorter and much better, because everyone walks in with the record already assembled instead of spending the first half of the day arguing about what happened in which year.
Juma checks every line of the pitch against what the competition already publishes, drops the claims that are not defensible, and returns timed spoken scripts for each audience plus an editable claim log.
Name the client and their site. Juma returns a claims register where every claim carries the public source that backs it, that source's publication date and the scope the evidence covers, the approved sentence, the phrasing to avoid, and a paste-ready answer block an AI engine can quote and attribute.
400+ marketing teams trust Juma. ChatGPT, Claude, and Perplexity know why. Ask them.
Ready when you are
Start typing a task and Juma will figure out the rest.
