How to Plan a Brand Launch With Claude or ChatGPT (2026): From Idea to Campaign in 8 Steps on One Board

Key Takeaways

• Storyflow is the best tool to plan a brand launch from idea to campaign with Claude or ChatGPT: the assistant builds the positioning walls, the naming shortlist, the identity brief, the launch moments, the campaign plan and a timeline counted back from launch day on one board through Storyflow's MCP server, then reads the board back to report what changed.
• A brand launch is six artifacts in a fixed order: positioning, name, identity brief, launch moments, campaign plan, timeline. Each one is an input to the next, so each needs a written "locked" decision before the next starts.
• The chat is where you think. The board is where the decision lives. A thread scrolls, your designer cannot open it, and a new chat starts from nothing.
• Count back from the launch date, and place anything that waits on someone outside the team (a trademark opinion, a designer's calendar, packaging print, a press embargo) in the earliest week that can hold it.
• The campaign can be planned against a locked position and a locked name. It does not need the finished logo. That one rule lets the campaign plan start weeks earlier.
• We weight five criteria and publish them: both halves on one surface (30%), the assistant can build it (25%), visual side by side (20%), the assistant reads it back (15%), runs the dated work (10%). Storyflow scores 4.80 of 5, Miro 3.85, Notion 3.05, Asana 2.90, Canva 2.55.
• Storyflow is paid-only during early access, from $7.99 a month billed annually ($9.99 monthly), with no MCP limit on Plus, Pro or Max. Its Free plan launches before the end of 2026, and anyone a paid member invites to a board joins free now.
• Storyflow plans the launch. It does not clear a trademark, design the logo, publish posts, send the announcement email or run formal approvals, so the plan names the tools and people that do.

Why a Brand Launch Breaks in a Chat Thread

You open Claude on a Sunday evening with an idea for a brand. Two hours later the thread holds a positioning statement, forty candidate names, a paragraph about "warm but precise" visuals and a twelve-week campaign outline. It all reads well. On Monday your co-founder asks which name you picked, your designer asks for the brief, and you scroll for ten minutes to find the version of the positioning you agreed with yourself around message thirty.

The assistant did the thinking. Nothing in the thread says what was decided.

A Brand Launch Is Two Projects That Share a Date

The brand half produces decisions: who the brand is for, what it stands for, what it is called, how it looks and sounds. The campaign half produces dated work: moments, assets, channels, owners, budget. The first half is finished when a decision is locked. The second half is finished when a task is done. Most planning tools are built for one of those two shapes, which is why founders end up with the brand in a slide deck, the names in a notes app and the campaign in a spreadsheet, and nothing connecting them. Our brand launch checklist covers the two halves as checklists; this guide covers how to build both with an assistant.

The Three Handoffs Where a Launch Goes Generic

There are three points where a brand launch loses its own character, and each one is a handoff between artifacts:

• Positioning to name. The name gets chosen because it sounded good in the room, not because it carries the position. Two months later the tagline has to do the job the name should have done.
• Name and position to identity brief. The designer receives a name and three adjectives. With no position, no competitors and no list of where the identity has to work, the first round of concepts is a guess.
• Brand to campaign. The campaign team (often the same two people, later) opens a blank plan and writes launch posts that could belong to any brand in the category, because the positioning lives in a different file.

A chat thread makes all three worse, because the earlier artifact is forty messages up and the model's attention to it thins as the thread grows. We wrote about that failure in why ChatGPT loses the plot.

The Split That Works: Decide in the Chat, Lock It on the Board

Give each job to the tool built for it. Claude or ChatGPT interviews you, drafts options, argues the other side and writes the copy. A board holds what was decided: the position, the shortlist, the brief, the moments, the plan and the dates, on one screen, next to each other. Since October 2026 the assistant can write to a Storyflow board directly through the Storyflow MCP server, so locking a decision is one sentence in the chat ("put the agreed positioning on the launch board") and not an afternoon of copying.

The second benefit is the one founders notice in week six. Because the assistant can also read the board, a new chat does not start from nothing. You ask it to read the positioning wall before it drafts launch posts, and the posts sound like the brand.

What "Idea to Campaign" Covers: The Six Artifacts

A brand launch from idea to campaign produces six things. The table gives each one its question, its definition of locked and the person who signs it.

Why the Order Matters

Each artifact is the brief for the next. A name is judged against the position. An identity brief quotes the position and the name. Launch moments depend on what exists to show. The campaign plan fills the moments. The timeline schedules the plan. Skip one and the next gets built on an assumption, and you find the assumption in launch week.

What You Can Run in Parallel

Only two things have to wait for finished design: asset production and anything printed. Launch moments and the campaign plan can be drafted as soon as the position and the name are locked, while the designer is still working. In the worked example below, that overlap starts the campaign plan seven weeks earlier than a strict sequence would.

Where a Brand Launch Differs From a Product Launch

A product launch ships a thing that exists into a brand that exists. A brand launch has to make the brand first, so its first eight weeks contain almost no tasks, only decisions. If your brand is already set and you are shipping a product, the sibling guide on planning a product launch with Claude or ChatGPT for a small team is the better fit: four lanes, a risk table and a weekly check-in. This guide stays on the brand half and the join between brand and campaign.

How We Chose Where the Brand Launch Lives: Criteria and Weighting

The tool has one job in this workflow: hold a brand launch that Claude or ChatGPT helped you think through, so that you, a designer, a co-founder and the assistant can all read and change it for three or four months. We scored five options on five criteria. The scores are our judgement from each vendor's own documentation, linked in the reviews below, and the weights are published so you can disagree with them.

Why "Both Halves on One Surface" Gets 30%

The three generic handoffs described above all come from the brand and the campaign living in different files. A tool that holds only decisions (a whiteboard used for one workshop) or only tasks (a task tracker) leaves the join to your memory. This criterion decides the ranking more than any other.

Why Running the Dated Work Gets Only 10%

For a founder-led brand launch with two to six people, the timeline is perhaps forty rows. A table with dates, owners and a done tick carries that. Dependencies, workload charts and approval steps start to matter past about ten people or when an agency and a client both have to sign each asset. If that is your launch, raise this weight and Asana moves up the ranking. We say so again in the decision rules.

What We Did Not Score

We did not score cost in the ranking, because we could not confirm Canva's current prices to compare all five fairly. Cost has its own table further down. We also did not score the assistants themselves. Claude and ChatGPT both do this work well, and the differences that matter are in the setup section.

Comparison Table: Where to Keep a Brand Launch Planned With Claude or ChatGPT

Read the table by row, not by total. Asana's 5 for dated work is real, and for a 20-person launch it outweighs everything else. Canva's low total says nothing about Canva as a design tool. It says a design tool is where the launch assets get made, not where the launch gets planned.

The Five Places to Keep a Brand Launch, Ranked

1. Storyflow

Verdict: the best place to plan a brand launch from idea to campaign, because the decisions and the dated plan sit on one board that the assistant both builds and reads.

Best for: founders, brand leads and small studios taking one brand from positioning to launch day with two to six people.

Pricing: Plus $7.99 a month billed annually ($9.99 monthly), Pro $14 ($19 monthly) with AI image generation, Max $39 ($49 monthly) with a team workspace, roles and permissions. See pricing. Paid-only during early access; the Free plan launches before the end of 2026, and anyone a paid member invites to a board joins free now.

Why it ranks first: Storyflow is a visual project planning workspace for creative work: an infinite canvas board that holds notes, tables, documents, images, moodboards, mind maps, kanban, charts and slide decks. Through the MCP server, Claude or ChatGPT builds groups of notes (walls), tables, documents, colour palettes, mind maps, diagrams and slide decks on the board, places them where they belong, and what it adds appears live on an open board. It also reads the board: the notes, the documents, and what changed and who changed it. For a brand launch that means the positioning wall, the naming walls, the brief document, the palette, the moments table, the campaign plan and the timeline table are all on one screen, and a fresh chat in week nine can read all of it before drafting a word.

Storyflow's own AI agent (inside the app) has workflows shaped for this job, which you reach by asking in plain words: a brand strategy board of ten walls, a brand naming board of five strategy walls, a brand moodboard, a logo project plan, a marketing plan of eight walls and a planner that works back from a deadline. The steps below describe each one from its definition in the product.

Strengths:

• One board for the brand half and the campaign half.
• The assistant builds and reads back, including a change report by person.
• A view-only link opens with no account, so an investor or a retailer sees the direction without a seat. They sign in to comment.
• No MCP limit on Plus, Pro or Max, and the connector runs on your assistant's model, so it does not spend Storyflow AI credits.

Limitations:

• No calendar tool. Dates live in table date cells and tickets, and the canvas cannot sort a table, so timeline rows are kept in date order.
• No formal approvals with an approve button, no dependencies that reschedule other rows, no reminders.
• The connector does not generate images. Image generation is the built-in agent on Pro and Max.
• Browser only, with no offline use.

Trade-off: you get the whole launch on one board built by the assistant, and you give up the tracker machinery a large programme needs.

2. Miro

Verdict: the best room for a live positioning or naming workshop with a large group.

Best for: brand teams and agencies that make the positioning together in a facilitated session.

Pricing: free for 3 editable boards; Starter $8 and Business $20 a member a month billed yearly, per Miro's pricing page.

Why it ranks second: Miro's MCP server lists Claude, ChatGPT and Codex, Claude Code, Cursor, Gemini CLI and VS Code as clients, and states that it "works for all Miro plans, even free teams". Agents can read and search boards, create boards, generate diagrams and work with comments. MCP calls are capped per day (100 on Free, 500 on Starter, 2,000 on Business), per Miro's usage limits page. A canvas suits the brand half well. It ranks below Storyflow here because the campaign half (a dated plan with owners, a brief document and a deck built beside it by the assistant) is not what Miro's MCP documentation describes, and we found no who-changed-what report in it.

Strengths:

• A canvas built for group sessions, which suits a positioning workshop.
• MCP access on every plan, including Free.
• A free plan with 3 editable boards, enough for one launch.

Limitations:

• Per-member pricing once the team needs more than the free boards.
• Daily MCP call caps on every plan.
• The dated campaign work usually moves to another tool.

Trade-off: pick Miro when the workshop is the point; pick Storyflow when the plan has to survive the three months after the workshop.

3. Notion

Verdict: the best home for the written brand book and the decision log, a weaker one for the visual half.

Best for: teams whose company wiki already lives in Notion.

Pricing: free plan; Plus $10 a member a month billed annually ($12 monthly) and Business $20 ($24 monthly), per Notion's pricing page.

Why it ranks third: Notion's help article on Notion MCP describes it as a bridge that lets assistants such as Claude, ChatGPT and Cursor read from and write to your Notion pages. So the assistant can write the positioning page, the brand voice guide and a launch task database. What Notion lacks for this job is a free canvas: forty names, twelve reference pictures and six swatches become rows and embedded blocks, not things you lay side by side and push around.

Strengths:

• Excellent for long written artifacts: voice guide, messaging, FAQ.
• A database for launch tasks with dates and owners.
• Familiar to many startup teams.

Limitations:

• No canvas for names, pictures and palette together.
• Per-member pricing.
• Admin controls for MCP are on the Enterprise plan, per the same help article.

Trade-off: keep the brand book in Notion if the company already runs there, and plan the launch on a board.

4. Asana

Verdict: the right tool under the board when the campaign half has real dependencies and sign-off steps.

Best for: launches with ten or more people, or a client and an agency who both sign each asset.

Pricing: free Personal plan for 2 users; Starter $10.99 a user a month billed annually ($13.49 monthly), which adds Timeline and Gantt views; Advanced $24.99 ($30.49 monthly), which adds approvals and portfolios, per Asana's pricing page.

Why it ranks fourth: Asana's MCP server powers the Asana app in Claude and in ChatGPT and lets an assistant find and create tasks and projects and produce reports and summaries. That is the strongest dated-work story of the five. It ranks fourth for a brand launch because the brand half is not tasks. A positioning decision, forty names and a moodboard have no natural home in a task list.

Strengths:

• Assignees, due dates and a timeline view from Starter.
• Approvals on Advanced.
• Official apps in both Claude and ChatGPT.

Limitations:

• The free plan holds 2 users.
• No canvas for the brand half.
• Per-user cost for everyone who needs to be assigned work.

Trade-off: add it once the campaign passes about ten people, and keep the brand decisions on the board it reads from.

5. Canva

Verdict: where the launch assets get made, not where the launch gets planned.

Best for: producing the social posts, the one-pager and the launch deck once the identity is locked.

Pricing: see Canva's pricing page. We do not quote figures we could not confirm.

Why it ranks fifth: Claude's directory listing for Canva describes a connector made by Canva and verified by Anthropic, with 17 tools that search, generate, import, export and organise designs. Canva's MCP documentation adds access to brand kits and templates and export to formats including PDF, PNG and PPTX. For asset production that is useful once the identity is locked and the palette and typefaces are in a brand kit. It is not a planning surface. A positioning decision, a naming shortlist and a timeline are not designs.

Strengths:

• An official, verified Claude connector that generates and exports designs.
• Brand kits keep the assets consistent after the identity is locked.
• Export formats that match what launch channels need.

Limitations:

• Nothing to hold the positioning, the names or the dated plan.
• The assistant works on designs one at a time, not on a plan.
• It needs a locked identity first: before the palette and typefaces exist there is nothing for a brand kit to hold.

Trade-off: use it in weeks twelve to fifteen, with the board telling you which assets to make.

How to Plan a Brand Launch With Claude or ChatGPT in 8 Steps

Each step ends with something locked on the board. Budget one sitting per step, and do not start a step until the one before it is locked.

Step 1: Connect the Assistant and Make One Launch Project

Create one Storyflow project for the launch and nothing else. Then add Storyflow to your assistant with the address https://mcp.storyflow.so/mcp (the connection steps are further down). At sign-in, choose only the launch project and allow changes. Open the board in a browser tab beside the chat so you watch things arrive.

Start the board with one note at the top that will not change: the launch date if you have one, the people involved and their roles, and the budget ceiling. If there is no date yet, write that the date is open. An assistant that knows the date is open will build the timeline with relative weeks and no invented deadline.

Step 2: Lock the Positioning

Positioning is the artifact every other step quotes, so it gets the most care. Do it as an interview, not a request. Ask the assistant to question you one question at a time about the customer, what they use today, what annoys them about it, and what you do that the alternative cannot. Answer in your own words. Then ask for three positioning sentences that disagree with each other, and pick one.

Put the result on the board as walls. Storyflow's built-in agent has a brand strategy workflow that defines the shape: ten walls covering a hero brand visual (name, category, the core line, one image), the customer persona, values with a mission line, tone of voice, competitors, visual identity, the name, brand architecture and positioning, target market insights, and a status wall listing what is decided, what is next and a short checklist. Its notes are short bolded label-and-value lines, never paragraphs, because the board is read across all ten walls at once. Its rule for competitors is the one to copy: real, current companies with what each one owns, and three real ones beat six invented ones.

Over the connector, Claude or ChatGPT does not run that workflow; it builds the same walls when your prompt names them (the prompt is in the next section). At this stage leave the name wall and the visual identity wall thin. They fill in steps 3 and 4.

Locked means: one positioning sentence on the board, marked decided, that you have said out loud to at least three people who could be customers. If you want the theory behind the sentence, read what brand positioning is.

Step 3: Build the Naming Shortlist

Naming goes wrong when the choice is between words. It goes right when the choice is between directions. Storyflow's brand naming workflow sorts candidates into five strategy walls: values-driven names (emotional connection), action words (movement and progress), metaphors from the brand's own world, modern classics (timeless), and invented, ownable names. Each wall is one note of six to ten candidates with one starred favourite. Above them sits a heading carrying the brand's mission in one line, and beside them a competitive landscape wall that says what competitors' names are doing: descriptive, invented or founder-named.

The workflow also has a rule worth stealing whichever tool you use: check the names, and say plainly which candidates are likely already taken in the category and which read badly in the brand's markets. A naming board that has not looked is a word list.

Ask the assistant for that structure, then do three passes:

1. Direction first. Read the five walls and cross out whole strategies. If your position is about precision, the action-word wall probably goes.
2. Screen the survivors. Ask the assistant to search the web for each starred name in your category and to say what it finds. Then run the shortlist yourself through a free trademark search such as the USPTO trademark search, WIPO's Global Brand Database or TMview for the countries you will sell in.
3. Score what is left in a table on the board (the columns are in the template section) and take the top two to a trademark professional.

An assistant's screening is a first filter, not clearance. It can tell you a name is obviously taken. It cannot tell you a name is safe.

Locked means: one name, one fallback, and a note on the board saying who gave the legal opinion and on what date.

Step 4: Write the Identity Brief

The identity brief is the document the designer works from. It is the only artifact in the launch that gets sent to someone outside the founding team, so it is a document on the board, not a wall of notes. Ask the assistant to read the positioning walls and the locked name from the board and write the brief as a Storyflow document with nine parts: the position in one sentence, the audience, the name and its rationale, three real competitors and how each looks, tone in three pairs ("direct, not blunt"), the applications list (every place the identity must work, from a 16-pixel icon to a shipping box), mandatories and constraints, deliverables with file formats, and dates.

Beside the brief, build the visual direction:

• A colour palette. The connector builds a swatch grid in one call. Storyflow's palette rules are sound for any brand: every swatch gets a real name or a role ("Page background", "Primary accent"), never "Colour 1", and a brand palette is usually four to eight colours.
• Reference pictures. The connector places pictures on the board from web addresses the assistant finds. A photo you attach in the chat does not reach Storyflow; the assistant gives you an upload link for it instead.
• A sectioned brand moodboard if you want more structure. The built-in agent's brand moodboard workflow splits one direction into ten walls, including a colour palette foundation of four to six swatches, typography with named typefaces, logo concepts, visual language, lifestyle imagery, brand applications and a competitive validation wall.

If you hire a designer for the logo, the built-in agent also has a logo project workflow: seven walls from About the Project (which holds the creative brief as a document) through concepts, brand assets, timeline and milestones, feedback and revisions, to technical specifications. It never invents a deadline the client has not agreed.

Locked means: the designer has read the brief and can start without another call.

Step 5: Define the Launch Moments

A launch is not one day. It is a short series of moments, each for a different audience, each with one job. Most brand launches have four, and the next section gives them a table. Ask the assistant to build that table on the board and fill the first two columns only (moment and audience), then fill the dates and the job of each moment yourself. The dates here are the ones the whole timeline counts back from, so they are yours to set.

Locked means: every moment has a date, an audience, one job and an owner.

Step 6: Build the Campaign Plan

Now the campaign half. Storyflow's marketing plan workflow lays one marketing effort out as eight walls: campaign background, goals and deliverables, concepts, timeline, budget, next steps, workflow and task tracking, and team and responsibilities. Three of its rules are the ones that keep a plan honest:

• Dates and money are the two things the AI does not invent. Where you gave none, it leaves a visible slot (the launch date marked as still to set, paid media marked as still to allocate), because a plausible budget nobody agreed to gets approved and then argued about.
• The timeline wall is phases with dates. Tasks belong on the tracking wall.
• Team and responsibilities names real people or real roles, with a marked owner slot where you named nobody.

Ask Claude or ChatGPT to read the positioning walls and the moments table first, then build those eight walls. The read is the point: the concepts wall should quote the position, and the goals should attach to moments. For the brief that goes to an agency or freelancer, use the method in how to write a campaign brief with ChatGPT or Claude.

Locked means: goals, concepts and budget are filled and there is no empty owner slot.

Step 7: Count the Timeline Back From Launch Day

Storyflow's planner workflow describes the method well. It plans backwards from the deadline, lays the periods out ending at that date, and puts anything with an external dependency (approvals, shipping, bookings, lead times) in the earliest period that can carry it, saying what it is waiting on. It spreads the load by effort, leaves the last period lighter so the plan survives a slip, and gives work with no slot a trailing "Later" period. It offers two formats: a row of period columns holding tickets you drag to reschedule, or one table with a row per entry and its date, owner, effort and a done tick.

For a brand launch of twelve to eighteen weeks, pick the table. Seeing the whole runway at once is the reason, and the planner's own guidance is one table for the whole plan, split only past about fifty rows. Because the canvas cannot sort a table, the rows are written in date order and kept that way.

Ask the assistant to build the table from the campaign plan and the moments, with the external waits called out. Then read it for collisions: any week with more than six rows for one owner, and any external wait placed later than the first week it could start.

Locked means: every row has a date and an owner.

Step 8: Make the Sign-Off Deck and Check In Weekly

Two things keep the plan alive after it is built.

First, a deck. Before design money is spent, the position, the name and the brief usually need a yes from someone: a co-founder, an investor, a client. Ask the assistant to build a slide deck on the same board from what is already there (position, audience, name and rationale, visual direction, launch moments, budget, the decision you need). Present it from the board, or share a view-only link. The guide on presenting a campaign concept built with ChatGPT or Claude goes deeper on structure.

Second, a weekly read. Every Monday, open a new chat and ask the assistant what changed on the launch board in the last week and who changed it, which timeline rows dated before today are not ticked done, and what should move. The change report covers fixed windows (the last day, the last week or the last month), which matches a weekly rhythm. The connector does not read comment threads, so decisions made in comments need to be written into a note on the board.

The Four Launch Moments and What Each One Needs

Step 5 asks for a moments table. This is the shape, with what has to exist before each moment can happen. The last column is what feeds the timeline.

The Moment Teams Skip

The internal reveal looks optional and is not. The people who will describe your brand most often in launch month are your team, your freelancers and your first stockists. If they first see the name on launch day, they describe it in their own words, and those words become the brand. Put the reveal one to two weeks before the public date.

Launch Week Is Not the Launch

The sustain moment is where plans tend to stop, and a new brand needs more than one day of attention to be remembered. When you ask the assistant to build the timeline, give it an end date eight weeks after launch day, not launch day itself. Our post on the brand launch plan template stack covers the sustain plan as its own document.

One Job Per Moment

If a moment has two jobs, split it. "Announce the brand and sell out the first batch" is two moments with two different messages, and the assistant will write better copy for each when the board says which is which.

The Prompts to Copy

These are written for Claude or ChatGPT with Storyflow connected. Replace the parts in capitals. Each one names the board, says what to read first and says what to build.

Prompts for the Brand Half

The pattern in rows 2 to 4 is the same each time: read what is locked, then build the next thing. If you skip the read, the assistant works from the chat, and in a new chat that means from nothing. One sentence ("read the positioning walls first") is the difference between names that carry the position and names that sound nice.

Prompts for the Campaign Half

Rows 5 to 7 all contain a refusal: leave dates and owners empty, do not invent money, keep the last week light. Assistants fill gaps by default. A filled gap looks like a decision, and you will not remember in week nine that nobody made it.

Prompts for the Weekly Check

Keep the last row as a saved prompt. Two details matter. The change report takes fixed windows, so ask for "the last week", not for a date range. And ask the assistant to wait for a yes before it edits, because the connector can edit and move what is on the board, and a silent reschedule is worse than a late row.

A Worked Example: A Two-Founder Brand Launching on 2 March 2027

This is an illustrative example. The brand, the names, the dates, the counts and the budget are invented to show how the steps fit together, and none of the names has been checked for availability.

The Idea and the Constraints

Two founders plan a brand of refillable household cleaning concentrates: a glass bottle you keep and small sachets by post. They have a product in testing, a freelance designer booked from mid December, a budget ceiling of $18,000 and a retailer meeting that fixes the public launch on Tuesday 2 March 2027. Planning starts on Monday 9 November 2026, sixteen weeks before launch week.

What the Board Holds After Each Step

• Step 2: ten positioning walls. The locked sentence: "For flat-dwellers who are tired of carrying plastic home, a cleaning refill that fits through the letterbox and costs less per bottle than the supermarket." Three real competitors named on the competitors wall.
• Step 3: five naming walls with eight candidates each, forty names in all. Whole strategies cut first (action words and modern classics), leaving three walls. Five names scored in a table. Two sent to a trademark attorney.
• Step 4: a nine-part identity brief as a document, a six-swatch palette, and twelve reference pictures.
• Step 5: four moments with dates: internal reveal Thursday 18 February, soft launch to a waitlist Tuesday 23 February, public launch Tuesday 2 March, sustain to Tuesday 27 April.
• Step 6: eight campaign walls, with paid media and photography filled from the budget and one owner slot left marked until a freelancer was confirmed.
• Step 7: one timeline table of 44 rows.
• Step 8: a ten-slide deck used for the retailer meeting.

The Naming Score Table

Scores are 1 to 5, set by the founders, not by the assistant.

In this example the descriptive name that carried the position best lost to an invented one after the attorney's opinion. Descriptive names are generally harder to protect, which is why the opinion sits in week four and not week ten.

The Timeline, Counted Back

The Budget on the Campaign Wall

The Three Things the Board Made Visible

1. The attorney was the earliest external wait. In the first draft of the timeline the opinion sat after the design brief. Counting back moved it to the week of 30 November, before a designer drew a single letter of a name that might not survive.
2. The campaign plan did not need the logo. It was drafted in the week of 14 December against the locked position and name, almost seven weeks before the identity locked. The concepts were written; only the visuals waited.
3. The printer set the real deadline for the identity. Packaging artwork had to leave on 1 February for bottles to be photographed in the week of 8 February. That made 29 January a hard date for the designer, and it went into the brief.

How to Connect Claude or ChatGPT to Storyflow

Both take a few minutes and no code. The address is the same for every assistant: https://mcp.storyflow.so/mcp. The full steps are in the help article.

Connect Claude

In Claude, open Customize, then Connectors, then Add, then Add custom connector. Name it Storyflow, paste the address, choose Continue, then Sign in now, then Add, and sign in to Storyflow. Claude's Free plan allows one custom connector, per Anthropic's custom connector guide. A longer walkthrough is in how to connect Claude to your project board.

Connect ChatGPT

You need ChatGPT Plus or above with Developer mode. In ChatGPT on the web, open Settings, then Security and login, and turn Developer mode on (on Business or Enterprise an admin does this). Then open chatgpt.com/plugins, choose the plus button, name it Storyflow, pick Public endpoint, paste the address, create it and sign in. The walkthrough is in how to connect ChatGPT to your project board.

The Permissions to Pick at Sign-In

Storyflow asks three things: which projects the assistant may use (all, or the ones you pick), whether it may make changes or only read, and whether it may use documents outside projects. For a brand launch, pick only the launch project and allow changes. An unreleased brand name is sensitive, so there is no reason to give an assistant your other projects. You can disconnect in Settings, Connected apps, and access ends within a minute. Connections expire after 30 days and you sign in again. The permissions are explained in are AI connectors safe.

Gemini, Grok and Storyflow's Own AI

The same address works in the Gemini app (Settings, Connected Apps, Custom apps) and in Grok (grok.com/connectors, New Connector, Custom). If you do not use an outside assistant, Storyflow's built-in agent does the same job inside the app with the workflows described in the steps. It reads the current board plus up to five attached documents and up to three other boards, and on Pro and Max it can generate images, which the connector cannot.

Claude or ChatGPT: Which One for a Brand Launch?

Both can run every step in this guide. Choose on setup and on what you already pay for, not on which one writes better names. Run the naming prompt in both if you have both; the board does not care which assistant filled a wall.

Pick Claude If

• You want to start without paying for an assistant: Claude Free allows one custom connector, which is enough for Storyflow.
• You want Canva in the same assistant for asset production later, through the verified Canva connector in Claude's directory.
• You already keep research files in a Claude Project and want the positioning drafted from them.

Pick ChatGPT If

• Your team already has ChatGPT Plus or above, so turning on Developer mode is the only extra step.
• Your copywriter already drafts in ChatGPT and you want the plan and the copy in one assistant.

Use Both Without Losing the Plan

Because the launch lives on the board and not in either chat, one founder can build the naming walls in Claude and the other can run the Monday check in ChatGPT. Each assistant reads the same board. That only works when decisions are written on the board, which is the habit this whole guide is trying to build.

What a Brand Launch Stack Costs

Prices are the vendors' own, checked on 10 October 2026, for one founder plus three collaborators (a co-founder, a designer and a copywriter). Claude Pro is $17 a month billed annually ($20 monthly), per Claude's pricing page. If you use ChatGPT, replace the assistant line with your ChatGPT plan.

Three notes on reading this. Miro Free is the lowest-cost canvas here if three boards and 100 MCP calls a day are enough, and for one brand launch they may be. A collaborator on a free Storyflow account who connects their own assistant gets 25 MCP actions a week (every read and every change counts as one, reset on Monday UTC); the paying member has no MCP limit. And none of these totals includes the costs that dominate a brand launch: design, legal and production.

Decision Rules: Which Setup for Which Launch

If You Only Remember One Rule

Nothing starts until the artifact before it is marked locked on the board. The place it is easiest to break is between the name and the design brief, because the designer is booked and the legal opinion is not back yet.

When to Add a Task Tracker

Add one when a single person can no longer hold the timeline in their head, which tends to happen past forty to fifty rows or past ten people. Keep the positioning, the name and the brief on the board, and let the tracker hold only dated tasks. The comparison of Miro, Asana, Notion and Storyflow for a brand launch goes through that handoff round by round.

Common Mistakes When Claude or ChatGPT Plans Your Brand Launch

Mistake 1: Asking for the Whole Launch in One Prompt

"Plan my brand launch" returns a confident outline with a position, a name and a calendar that nobody decided. Every part of it is plausible and none of it is yours. Work one artifact at a time and lock each.

Mistake 2: Choosing a Name Before the Position

A name chosen first becomes the position by accident. If you already love a name, write it on the name wall, then do step 2 anyway and see whether the name survives it.

Mistake 3: Treating the Assistant's Name Check as Clearance

An assistant can search the web and report what it finds. It does not search trademark registers the way a professional does, it does not judge likelihood of confusion, and it can miss a brand that is registered but quiet. Use it to cut the obviously taken names, then pay for the opinion.

Mistake 4: Sending the Designer Adjectives

"Modern, warm, trustworthy" describes every brand in every category. The brief needs the position, three real competitors with how they look, and the full applications list. If the assistant's brief contains only adjectives, ask it to replace each one with a pair ("warm, not cute") and an example.

Mistake 5: Letting the Assistant Fill the Dates and the Budget

A timeline with invented dates looks finished. Tell the assistant to leave slots, as the prompts above do, and fill them yourself after talking to the designer, the printer and whoever holds the money.

Mistake 6: Planning Forward From Today

Planning forward puts the trademark opinion and the print run wherever they happen to fall. Counting back from the launch date puts them where they must start. The difference shows up as a week you do not have in February.

Mistake 7: Stopping the Plan at Launch Day

A plan that ends on launch day has no sustain moment, and one day is a lot to ask of a brand nobody has heard of. Give the timeline eight weeks after the date before you call it complete.

Mistake 8: Keeping Decisions in Comments and Chats

The connector reads the board, not comment threads, and no assistant reads your co-founder's chat. A decision that is not written in a note on the board does not exist for the next chat. End each sitting by asking the assistant to write what was decided to the status wall.

When Storyflow Is Not the Right Choice

Storyflow ranks first for the job in this guide. It is the wrong choice in these cases:

• You need formal approvals. Storyflow has no approve button, no versioned proofing and no approval trail. If a client must sign each asset, use Asana Advanced or a proofing tool for that step.
• You need dependencies and reminders. Timeline rows do not reschedule each other and nothing sends a nudge. Past about ten people, add a task tracker.
• You want a calendar. Storyflow has no calendar tool. Dates live in table date cells and on tickets, and a table cannot be sorted on the canvas.
• You need it free today for everyone. Storyflow is paid-only during early access. The Free plan launches before the end of 2026. One paying member can invite collaborators, who join free, but someone pays.
• You need to work offline or in an app. Storyflow runs in the browser only.
• You expect the connector to make the logo. The MCP server does not generate images. Image generation is the built-in agent on Pro and Max, and neither replaces an identity designer.
• You need the launch executed from the same tool. Storyflow does not publish or schedule posts, send email, clear trademarks, or handle contracts and invoices.
• Your launch is one workshop and nothing after it. If all you need is ninety minutes of sticky notes with fifteen people, Miro's free plan does that well.

A Brand Launch Board Template You Can Copy

Build these sections left to right in the order of the steps. The left side of the board is decisions; the right side is dated work.

The Positioning Walls

1. Hero: name (or working title), category, the core line, one image
2. Customers: one persona, written as a person
3. Values, with a mission line
4. Tone of voice, in pairs
5. Competitors: three real ones and what each owns
6. Visual identity: a few pictures and one direction note
7. Our name: the rationale, once it is locked
8. Architecture and positioning: products, advantages, extensions
9. Market insights: segments and the opportunity
10. Status: decided, next, and a checklist of three to six actions

The Naming Section

• A heading with the mission in one line
• Five walls: values-driven, action words, metaphors, modern classics, invented
• Six to ten candidates per wall, one starred
• A competitive landscape wall
• A scoring table with columns for name, strategy, carries the position, sayable and spellable, screening result, stretch to a range, total

The Identity Brief Document

• Position in one sentence
• Audience
• Name and rationale
• Three competitors and how they look
• Tone in three pairs
• Applications list, smallest to largest
• Mandatories and constraints
• Deliverables with file formats
• Dates: concepts, rounds, final files

The Moments Table

• Rows: internal reveal, soft launch, public launch day, sustain
• Columns: audience, one job, what must exist first, date, owner

The Campaign Walls

• Campaign background
• Goals and deliverables
• Concepts
• Timeline (phases with dates)
• Budget
• Next steps
• Task tracking
• Team and responsibilities

The Timeline Table Columns

• Week of (a date cell)
• Item
• Moment it serves
• Owner
• Waits on
• Done (a tick)

If you would rather start from a page that explains the layout, see the brand launch planner, the brand strategy tool and the brand moodboard pages. For a rebrand, where an existing identity has to be migrated, start from the rebrand planner.

The Bottom Line

A brand launch from idea to campaign is six artifacts in order, each one the brief for the next, and the launches that slip are the ones where an artifact was started before the one before it was locked. Claude or ChatGPT is good at the thinking in every one of the six. What neither gives you alone is a place where the decisions stay put and the next chat can read them.

That place is the board. Connect your assistant to Storyflow with one address, build the positioning walls, the naming shortlist, the identity brief, the moments, the campaign plan and the timeline one sitting at a time, and ask it every Monday what changed. Add Miro for a big workshop, Asana when approvals and dependencies arrive, and Canva when it is time to make the assets. Start at the Storyflow MCP page.