Map a marketing campaign with ChatGPT or Claude in seven layers (objective, audience, message, channels, moments, assets, measurement), built by the assistant on a Storyflow board, with copyable prompts, an illustrative worked example and an audit that shows which assets to cut.

Category
Marketing
Author
Storyflow Team
Product & Research Team
Topics
2026-10-10
•
45 min read
•
MarketingThe simple AI tool to use for mapping out a marketing campaign is Storyflow. Connect it to ChatGPT or Claude once, describe the campaign in the chat, and the assistant builds the map on a Storyflow board: a mind map for the overview and seven sections for objective, audience, message, channels, moments, assets and measurement. Your team edits that board and a client opens it by link with no account. Choose Miro when your team already runs workshops there, Whimsical when the map is mostly flowcharts, Notion when the plan has to be a database, and Canva when all you need is a designed one-page graphic.
Full disclosure: Storyflow is our product. It ranks first here because the job is mapping a marketing campaign from a ChatGPT or Claude conversation, and its connector builds the mind map, the sections and the dated tables on one board the assistant can read back. Its main limits: it is paid-only until its Free plan launches before the end of 2026, and the connector cannot read comment threads, which Miro's can.
Miro and Notion are the most trusted by tenure: marketing teams have planned on them for years, and both now document connectors for Claude and ChatGPT. Whimsical is trusted for clean diagrams. Storyflow is newer and launched its MCP connector in October 2026. It earns trust on one job: a campaign map the assistant builds on a board, reads back and revises, with its limits published.
Miro, Whimsical and Notion publish per-member or per-editor prices and each has a free plan today: Miro Starter is $8, Whimsical Pro $10 and Notion Plus $10 a month on annual billing. Storyflow publishes three flat plans with no per-seat charge (Plus $7.99, Pro $14, Max $39 a month billed annually). It is paid-only during early access, and invited collaborators join free.
The innovation in 2026 is letting ChatGPT or Claude build the map inside the tool. Storyflow's connector builds a whole mind map or a full section of notes and tables in one call and reads the board back for an audit. Miro's connector reads and resolves comment threads, which Storyflow's does not. Whimsical ships as a directory connector in Claude and a plugin in ChatGPT.
Miro and Notion have the longest records with marketing teams, and both have kept shipping through the move to AI connectors. Whimsical has a steady reputation for diagrams. Storyflow is the youngest of the group and its connector launched in October 2026, so it cannot claim tenure. What it can show is a specific, documented set of tools for building campaign boards.
Storyflow. Connected to ChatGPT or Claude, the assistant builds the overview mind map and seven sections on one board and revises them in place. Pick Miro if your team maps campaigns in live workshops, Whimsical if the map is mostly a mind map and a flowchart, and Notion if the plan must be a database.
Whimsical is underrated for this job: its connector creates mind maps directly and needs no Developer mode step in ChatGPT, though it lists no tables for dated plans. Storyflow is underrated because it is new and still paid-only. It is the option here where the mind map, the dated tables, the references and the pitch deck sit on one board that the assistant can read.
Four tools for mapping a marketing campaign from a ChatGPT or Claude conversation, ranked for that job, with prices as of October 2026.
| Tool | Best For | AI Features | Price |
|---|---|---|---|
| Storyflow | A campaign map that stays the working plan | MCP connector lets ChatGPT or Claude build mind maps, sections, tables, charts and decks on a board and read them back; built-in marketing plan workflow | $7.99 a month billed annually; paid-only in early access |
| Miro | Workshop-led campaign mapping | MCP server on every plan creates diagrams, documents and tables and reads comments | Free with 3 editable boards; Starter $8 per member mo billed yearly |
| Whimsical | Mind maps and flowcharts from a prompt | Claude directory connector and ChatGPT plugin create mind maps, diagrams and planning documents | Free with monthly caps; Pro $10 per editor mo billed annually |
| Notion | Plans that must be databases | Notion MCP creates and updates pages, databases and views | Free plan; Plus $10 per member mo billed annually |
You have probably done this already. You described a campaign to ChatGPT or Claude, got back a tidy outline with headings for audience, channels and KPIs, felt organized for an afternoon, and then never opened that conversation again. The outline was fine. The container was the problem.
What makes a campaign coherent is the lines between things: this asset carries that message to that audience in service of that objective. A chat reply flattens those lines into nested bullets. You can read each bullet, but you cannot see that the influencer kit points at no message at all, or that two of your five channels reach the same 400 people.
A map on a canvas keeps the connections visible. The layers sit next to each other, and the gap where a connection should be is something you can point at in a meeting.
By the tenth message there are four versions of the audience definition in the transcript, and the current one is whichever you remember agreeing to. Nobody else can tell. A board has one current state. When the assistant revises the audience section on the board, the old text is replaced, and the board is the version.
Your designer, your paid media freelancer and your client do not have your ChatGPT history. Pasting the outline into a document fixes the access and loses the assistant: the next revision means pasting again. The point of a connector is that the assistant and the team work on the same object. Storyflow's connector, described on the MCP page, gives the assistant tools to build on a board and to read what is there, including what your team changed.
Ask for a campaign plan with no budget and you will get one with a budget. It will look reasonable and nobody agreed to it. A map that gets shared carries more authority than a chat reply, so the rule tightens: anything the assistant was not given is written as a visible slot, not as a guess. Every prompt in this guide enforces that.
Our existing guide to what a campaign map is uses five layers. Working inside an assistant, we add two, because they are the two an assistant most reliably skips or invents: moments (when things land) and measurement (how you will know). The test for the whole map is traceability. Every asset traces to a message, every message to an audience, every audience to the objective. Anything that cannot make the trip gets cut.
| Layer | The question it answers | What goes on the board | Shape on Storyflow |
|---|---|---|---|
1. Objective | What changes, by how much, by when? | One sentence with one number and one date | A note at the top of the first section |
2. Audience | Who, specifically, and what do they believe now? | One to three segments, each with a current belief and a barrier | One note per segment |
3. Message | What is the one thing you want them to believe? | One core message and two or three proof points | A note plus a short list |
4. Channels | Where will they meet the message? | Each channel with its job, its audience and its share of budget | A table |
5. Moments | When does each beat land? | Dated beats, from first tease to close | A table with rows kept in date order |
6. Assets | What has to be made? | Every asset with format, channel, message, owner and due date | A table |
7. Measurement | How will you know? | One primary KPI, two to four diagnostics, baseline, target, source | A table, plus a chart when you have numbers |
"Grow awareness of the new line" cannot be mapped, because nothing can be traced to it or cut against it. "Book 180 home trials between April 6 and May 17" can. If the assistant offers three objectives, ask it which one the other two serve and keep that one. The other two usually turn out to be diagnostics that belong in layer 7.
"Urban professionals aged 28 to 45" tells a media buyer where to point the spend and tells a writer nothing. Add what the segment believes today and what stops them acting. "Commuters within five miles of a store who think an e-bike is a toy for people who already cycle" gives the message layer something to change.
Keep it to three segments at most. An assistant will happily produce six personas with names and hobbies. Six audiences on a small budget means six thin campaigns.
Write one core message as a sentence a customer could repeat, then two or three proof points that make it credible. The message is not the tagline. The tagline comes later and belongs to the creative team. If your message contains the word "and", it is two messages, and the map will show you that when half the assets trace to one half.
A channel on the map without a job is a habit. For each channel, write what it is there to do (reach new people, prove the claim, convert the interested, remind the undecided), which segment it reaches, and what share of the budget it gets. Two channels with the same job and the same segment is a prompt to cut one.
A campaign has a shape in time: a first beat, a peak, a reminder, a close. Without that layer, the map is a diagram of a campaign that never starts. Moments are dated beats, each with the message emphasis and the channels that carry it. On Storyflow they live in a table with a date cell per row. Storyflow has no calendar, and its tables do not sort, so tell the assistant to write the rows in date order and keep them that way.
List every asset with five things: format, channel, the message or proof point it carries, owner, and due date. The "message it carries" column is the one that matters. An asset with nothing in that column is either missing its purpose or should not exist. In practice a first asset list from an assistant is long, because assets are cheap to list and expensive to make.
Pick one primary KPI that restates the objective, then two to four diagnostics that explain it (visits, booking rate, cost per booking). For each, record the baseline, the target and where the number comes from. If you do not know a baseline, the cell says "[baseline needed]" and someone owns finding it. An assistant asked for "KPIs" will list eight. A map with eight KPIs has none.
"Map" makes people reach for a mind map, and a mind map is half the answer. The shapes hold different amounts of text, and choosing the wrong one is the commonest reason an AI-built campaign map looks impressive and is useless.
| Shape | What it holds | Use it for | Do not use it for |
|---|---|---|---|
Mind map | A centre with branches and sub-branches, each a short label | The one-glance overview of all seven layers | Budgets, dates, owners, anything with columns |
Orbit | One ring of cards around a small centre | One idea with its reference material: the message with its proof, images and links around it | A hierarchy with more than one level |
Sections | A wall per layer, holding notes, tables and documents | The working detail of every layer | A first look for someone who has two minutes |
Table | Rows and typed columns, including date cells | Channels, moments, assets, measurement | Narrative, such as the audience belief |
Storyflow's mind map tool builds a whole map in one call: a central topic, branches and sub-branches, laid out and coloured by the tool with the connecting lines drawn. Each node takes a label of three to eight words. That constraint is useful. "Commuters who think e-bikes are toys" fits. A paragraph about the segment does not, and should not be on the overview.
One limit to plan around: the connector builds a mind map whole and has no tool for growing an existing one. It can edit elements already on the board, but a new branch is something you add by hand on the board, or the assistant builds a fresh map and you remove the old one. So settle the tree in the chat first, then build once.
The detail lives in sections. Through the connector, the assistant builds each section in one call: a wall with a heading and every card inside it, top to bottom, in the order you give. A card can be a note, a table, a document, a checklist or a ticket. Seven layers means seven sections, asked for in layer order so the board reads like the argument.
An orbit is one ring of things around a centre. It suits the message layer when you have material: the core message in the middle, with proof points, a customer quote you were given, two reference images and a competitor link around it. If you have no material yet, skip it. An orbit of six notes that repeat the mind map adds nothing.
Inside the Storyflow app there is also a built-in AI agent with a marketing plan workflow. It lays out a fuller execution plan in eight sections (campaign background, goals and deliverables, concepts, timeline, budget, next steps, task tracking, and team and responsibilities), and it is written to leave dates and money as visible slots when you gave none. Use the connector route in this guide when your thinking is already happening in ChatGPT or Claude. Use the built-in agent when you are already on the board. Both produce ordinary board elements you can edit by hand.
The question is narrow: a simple AI tool for mapping out a marketing campaign, used by someone who already plans in ChatGPT or Claude. That is a different job from running a workshop, managing tasks or designing a deck, so the criteria are specific to it.
| Criterion | Weight | What a 10 looks like |
|---|---|---|
Builds the map from the chat | 30% | The assistant creates the whole map in the tool, from ChatGPT and from Claude, with no copy and paste |
Shape fit for a campaign map | 20% | Mind map, free canvas sections and tables with dates on one surface |
Simplicity | 20% | Few steps from a sentence to a map, a price you can state in one line, nothing to configure |
Reads back and revises | 15% | The assistant reads the current map, including your team's edits and comments, and changes it in place |
Sharing | 15% | A client opens the map with no account; the team edits together |
Scores are our judgment out of 10, based on each vendor's own documentation as of October 2026, linked in the reviews. The weights are published so you can re-weight them: if your team will never use ChatGPT, or if a free plan matters more to you than shape, the order can change.
Every tool here can hold a campaign map you draw yourself. The request was for an AI tool, and in 2026 the AI most marketers already use is ChatGPT or Claude. A tool whose own AI button builds a map is useful, but it starts from zero context. The assistant you have been talking to for an hour already knows the product, the audience and the constraint you mentioned in passing.
We did not score execution features: task assignment, approvals, publishing, reporting dashboards. Mapping comes before those, and a tool that is strong at them is often heavy for a map. We also did not score each tool's built-in AI separately. This guide is about the assistant you already use.
| Tool | Score | Best for | Works with ChatGPT / Claude | What the assistant can build | Entry price | Free plan |
|---|---|---|---|---|---|---|
Storyflow | 8.3 | A campaign map that stays the working plan | Yes / Yes (custom connector) | Mind maps, sections of notes, tables, charts, diagrams, documents, slide decks | $7.99 mo annual, flat | Planned before the end of 2026 (invited collaborators join free now) |
Miro | 7.8 | Teams already running workshops in Miro | Yes / Yes | Diagrams, documents, tables with kanban and timeline views, shapes and cards | $8 per member mo, billed yearly | Yes, 3 editable boards |
Whimsical | 7.4 | Maps that are mostly flowcharts and mind maps | Yes (plugin) / Yes (directory connector) | Mind maps, flowcharts, sequence diagrams, wireframes, planning documents | $10 per editor mo, billed annually | Yes, 50 board objects a month |
Notion | 6.4 | Plans that have to be a database | Yes / Yes | Pages, databases and database views | $10 per member mo, billed annually | Yes |
Canva | 5.8 | A designed one-page campaign graphic | Claude connector verified | Generated designs, exports | See Canva's pricing page | See Canva's pricing page |
How the weighted scores break down:
| Tool | Builds from chat (30%) | Shape fit (20%) | Simplicity (20%) | Reads and revises (15%) | Sharing (15%) | Total |
|---|---|---|---|---|---|---|
Storyflow | 9 | 9 | 7 | 8 | 8 | 8.3 |
Miro | 8 | 7 | 7 | 9 | 8 | 7.8 |
Whimsical | 8 | 7 | 8 | 7 | 6 | 7.4 |
Notion | 7 | 4 | 6 | 8 | 7 | 6.4 |
Canva | 6 | 3 | 7 | 5 | 8 | 5.8 |
Storyflow does not score 10 anywhere, and the gaps are specific. It gets a 7 on simplicity because it has no free plan yet and because ChatGPT needs Developer mode switched on before you can add it, where Whimsical is a plugin in ChatGPT's own directory. It gets an 8 on reading back because its connector cannot read comment threads, which Miro's can. It gets an 8 on sharing because it runs in the browser only, with no desktop or mobile app.

Storyflow board laid out in sections for planning client and campaign work
Connect Storyflow to ChatGPT or Claude, run the mapping prompts, and get a seven-layer campaign map your team edits and your client opens by link.

The ten steps below take a campaign from a paragraph in a chat to a map on a board that someone else can open. Steps 1 to 4 happen in the chat. Steps 5 to 8 build the board. Steps 9 and 10 are what make the map worth having.
In Claude, open Customize, then Connectors, then Add, then Add custom connector. Name it Storyflow, paste the address https://mcp.storyflow.so/mcp, continue, and sign in to Storyflow. Claude's Free plan allows one custom connector, per Claude's help article on custom connectors.
In ChatGPT on the web, you need ChatGPT Plus or above with Developer mode. Turn Developer mode on under Settings, then Security and login. Then go to chatgpt.com/plugins, press the plus button, name the connector Storyflow, paste the same address and sign in.
At sign-in you choose which projects the assistant may reach and whether it may make changes. For a first run, pick one project and allow changes. The full walkthrough is in the help article on connecting AI assistants, with assistant-specific guides for ChatGPT and Claude.
Before any mapping, paste what actually exists: the product, the price, the budget if there is one, the dates if they are fixed, last campaign's numbers if you have them, and anything the client or founder has said out loud about what they want. Then add one instruction: use only these facts, and mark anything missing as a slot in square brackets.
Five true sentences beat a page of background. The assistant maps what it is given, and what it is not given it either marks or invents. You decide which with that one instruction.
Ask the assistant to question you layer by layer, objective first, and to move on only when a layer passes its test (one number and one date for the objective, a current belief for each audience). Expect roughly ten to twenty minutes of typing. It is the difference between your campaign and a generic one. If the campaign starts from a client document, do the campaign brief first and map from that.
Ask for the map as an indented outline in the chat: the campaign name at the centre, seven branches, and no more than five sub-branches each, every label three to eight words. Read it. Fix it there. Building is one call, and a rebuilt map arrives as a second map beside the first, so the cheap place to argue about the tree is the chat.
Now ask the assistant to build that tree as a mind map on your board, and to show you the board while it works. In clients that support it, the live board appears in the chat and updates as elements land. On an open Storyflow tab, the map appears live as well.
Look at it for thirty seconds before going further. If a branch is wrong, you will see it faster here than in an outline.
Ask for one section per layer, to the right of the mind map, in layer order. Be explicit about the shape of each: notes for objective, audience and message; tables for channels, moments, assets and measurement, with the columns named. The template later in this guide has the columns. Asking for the columns by name is what stops the assistant choosing its own.
Give the assistant the fixed dates and have it propose the beats in between, written as rows in date order: date, moment, message emphasis, channels, owner. Where a date is not set, the cell reads "[set date]". Check the gaps. Three weeks with no beat is either a deliberate quiet period or a hole.
This is the prompt that justifies the whole exercise. Ask the assistant to read the board and report three lists: assets that carry no message, messages or proof points that no asset carries, and channels with no asset or no stated job. Then ask it what it would cut if the budget dropped by a third.
It is reading the real board here, not its memory of what it wrote, so the audit also catches what you changed by hand.
Build the measurement table with five columns: metric, role (primary or diagnostic), baseline, target, source. Fill baselines from numbers you paste. Leave the rest as "[baseline needed]" with an owner. The assistant can also add a chart to the board, and it should only do that from numbers you supplied, such as last campaign's weekly bookings. Storyflow does not pull data from ad platforms or analytics.
Share the board with a view-only link. The client or stakeholder opens it with no account, and signs in if they want to comment. Teammates you invite can edit the board directly.
A week later, ask the assistant what changed on the board in the last week and who changed it. The connector reads changes over fixed windows (the last day, the last week or the last month). Then ask it to re-run the traceability audit against the board as it stands. It cannot read comment threads, so paste any comments you want it to act on into the chat.
Each tool below is placed for one job only: mapping a marketing campaign from a conversation in ChatGPT or Claude. All vendor facts were checked on the vendor's own pages in October 2026 and are linked.

The verdict: the simple route from a campaign described in a chat to a map your team can work from.
Best for: marketers, founders and small agencies who think in ChatGPT or Claude and need the result to become the plan.
Pricing: Plus is $7.99 a month billed annually ($9.99 monthly), Pro $14 ($19 monthly) and Max $39 ($49 monthly), all flat prices with no per-seat charge, listed on the pricing page. Storyflow is paid-only during early access. Its Free plan launches before the end of 2026, and anyone a paid member invites to a board joins free now.
Why it ranks first here: Storyflow is a visual project planning workspace for creative work, and its connector exposes the shapes a campaign map needs as single calls. One call builds a whole mind map. One call builds a section with every card in it. Tables carry date cells for the moments layer. The same board then takes the things a map turns into: a moodboard, a content plan, a slide deck for the pitch. The assistant reads the board back, so the map can be audited and revised in place instead of regenerated.
Strengths:
Limitations:
The verdict: the right answer when the map is going to be built or reviewed in a live workshop, and your team is already in Miro.
Best for: marketing teams of five or more who run planning sessions on Miro boards.
Pricing: Free with 3 editable boards; Starter $8 per member a month and Business $20 per member a month, both billed yearly, per Miro's pricing page.
Why it ranks second: Miro's connector is well documented and generous. Miro's MCP documentation says the server "works for all Miro plans, even free teams" and lists Claude and ChatGPT among supported clients. Its tool list covers diagrams, documents, tables that switch between table, kanban and timeline views, and reading, replying to and resolving comments. That last one is a real advantage over Storyflow for review rounds.
It ranks below Storyflow for this job on two counts. Its documented diagram types are flowcharts, ERDs and UML diagrams, and its tool documentation does not name a mind map tool, so the overview is assembled from shapes. And in our view a Miro board is a workshop surface first: strong for the session, less natural as the place the campaign's references, plan and pitch deck later live.
Strengths:
Limitations:
The verdict: the simplest setup of the five, and a good pick when your campaign map is really a mind map plus a customer flow.
Best for: solo marketers and product marketers who want a clean mind map or flowchart from a prompt.
Pricing: Free with monthly caps (50 board objects and 50 doc blocks a month); Pro $10 and Business $20 per editor a month billed annually, per Whimsical's pricing page.
Why it ranks third: Whimsical's MCP page says your AI can "create mind maps, wireframe outlines, and planning documents directly in Whimsical", and can search and read your files. It is a verified connector in Claude's directory and a plugin in ChatGPT, which means no Developer mode step. That is why it scores highest on simplicity.
It ranks below Storyflow and Miro because the campaign map needs tables for channels, moments, assets and measurement, and Whimsical's MCP page does not list tables or date fields among what the assistant builds. The Free plan's cap of 50 board objects a month may also be tight for this job. A seven-branch mind map with five sub-branches each is 43 nodes, so if each node counts as an object, one map uses most of a month.
Strengths:
Limitations:
The verdict: choose it when the map has to become a database the team filters and assigns from, and nobody needs to see it as a picture.
Best for: teams whose briefs, tasks and wikis already live in Notion.
Pricing: Free plan available; Plus $10 per member a month billed annually ($12 monthly) and Business $20 ($24 monthly), per Notion's pricing page.
Why it ranks fourth: Notion's connector is strong at what Notion is. Its supported tools create pages, databases and database views including board, timeline and calendar, and they read and write comments. Notion's help page names Claude, Cursor and ChatGPT as supported. For the assets and moments layers, a Notion database is arguably better than a table on a canvas: it filters, sorts and has a calendar.
It ranks fourth because none of those tools creates a canvas or a mind map. What you get is a well-structured page, which is an outline again. The connections between layers become relation properties you have to click through, not lines you can see.
Strengths:
Limitations:
The verdict: the pick when the deliverable is a good-looking one-page campaign overview for a slide or a wall, not a working plan.
Best for: marketers who need the map to look designed for a leadership update.
Pricing: see Canva's pricing page for current plans and prices.
Why it ranks fifth: Canva's connector listing for Claude shows a connector made by Canva and verified by Anthropic, with 17 tools that search, generate, import and export designs. For a finished graphic, that is the shortest path of the five. For a campaign map it is the weakest fit, because a design is a picture of a plan. The asset table inside it is not a table the assistant can read row by row and audit, and the next revision is a new design.
Strengths:
Limitations:
What follows is an illustrative example. The shop, the numbers and the counts are invented to show the method, and none of it is a measured result.
Northbank Cycles is a three-store bike shop. It wants to sell commuter e-bikes to people who do not think of themselves as cyclists, using a free 48-hour home trial. It has $12,000 and six weeks, April 6 to May 17. The marketing lead works in Claude and has Storyflow connected.
We pasted six facts: the average e-bike sale is $2,400; the budget is $12,000; the dates are fixed; last spring's in-store test rides converted to sales at about one in four; the booking page converts 6% of visitors; and the email list has 4,100 local subscribers. We told the assistant to use only those facts and to bracket anything else.
The owner's first objective was "get the word out about e-bikes". The interview turned it into a number by working backwards.
| Step | Calculation | Result |
|---|---|---|
Sales the budget has to justify | Owner's target | 45 e-bikes |
Revenue at that target | 45 x $2,400 | $108,000 |
Marketing cost per sale | $12,000 / 45 | About $267 |
Trials needed at one in four | 45 x 4 | 180 trials |
Booking page visits needed at 6% | 180 / 0.06 | 3,000 visits |
The objective on the board became: "Book 180 home trials between April 6 and May 17, to sell 45 commuter e-bikes." The 3,000 visits went to the measurement layer as a diagnostic, not into the objective.
| Layer | What went on the board |
|---|---|
Objective | Book 180 home trials between April 6 and May 17, to sell 45 commuter e-bikes |
Audience | Two segments. Car commuters within five miles of a store, who believe an e-bike is a toy for cyclists. Lapsed cyclists aged 40 plus, who believe their riding days ended with the hills |
Message | "Try your actual commute for 48 hours, free." Proof points: the trial costs nothing, the bike is delivered, most commutes under five miles take the same time as driving [confirm with route data] |
Channels | Paid social (reach car commuters), search (convert the interested), email (lapsed cyclists), local ambassadors (prove it), in-store and print (remind) |
Moments | Five dated beats, below |
Assets | 22 listed at first, 18 after the audit |
Measurement | Primary: trials booked. Diagnostics: booking page visits, booking rate, cost per trial, trial-to-sale rate |
Note the bracket in the message row. The "same time as driving" proof point was something the assistant offered, not something we gave it, so the instruction from step 2 made it a slot. Someone has to test three real routes before that line appears in an ad.
| Channel | Job | Budget | Share |
|---|---|---|---|
Paid social | Reach car commuters near each store | $4,800 | 40% |
Search | Convert people already looking | $2,400 | 20% |
Local ambassadors | Prove the commute claim with real riders | $1,800 | 15% |
Photo and video production | Make the trial look easy | $1,800 | 15% |
In-store and print | Remind and collect walk-ins | $1,200 | 10% |
Total | $12,000 | 100% |
The budget was ours, so the assistant split it and we adjusted. The paid channels total $7,200. To bring 2,000 of the 3,000 visits, they have to average $3.60 a visit, and the remaining 1,000 have to come from the email list, the ambassadors and in-store QR codes. That $3.60 is a requirement the map exposes, not a forecast. If the first week comes in at $6, the map tells you which number breaks.
| Date | Moment | Message emphasis | Channels |
|---|---|---|---|
April 6 | Trial booking opens | Free, 48 hours, delivered | Email, paid social, search |
April 17 | First demo day | Try it in ten minutes | In-store, ambassadors, paid social |
April 27 | "What your commute costs" send | The car is the expensive option [confirm figures] | Email, search |
May 8 | Second demo day | Riders' own commute stories | Ambassadors, paid social, in-store |
May 17 | Trials close | Last days to book | Email, paid social |
The first draft had nothing between April 17 and May 8. Three weeks of silence in a six-week campaign was a hole, and the April 27 beat was added to fill it.
The traceability audit reported four assets with no message: a general brand video, a "meet the mechanics" reel, a tote bag and a spring service discount flyer. Each was a reasonable idea and none of them told a car commuter to try their commute for 48 hours. All four were cut, taking the list from 22 to 18.
It also reported the opposite problem. The proof point "the bike is delivered" had no asset carrying it. One paid social variant was rewritten around delivery. That gap would not have been visible in an outline.
Every read and every change through the connector counts as one action. In this illustrative run the count was about 15: two to find the project and board, one to show the board, one for the mind map, seven for the sections, one to read the board for the audit, and three edits afterwards. An invited free account gets 25 actions a week, so a map like this fits with room to revise. On Plus, Pro and Max there is no connector limit.
Replace the bracketed parts. The prompts assume Storyflow is connected and that you have a project for the campaign. They are written to be pasted one at a time, in order.
| Stage | Prompt to paste |
|---|---|
Ground rules | I am going to map a marketing campaign with you. Use only facts I give you. Where a date, budget, baseline or name is missing, write it as a slot in square brackets and keep a running list of open slots. |
Inputs | Here is what is true about this campaign: [product, price, budget, fixed dates, last results, constraints]. Tell me which of the seven layers (objective, audience, message, channels, moments, assets, measurement) you cannot fill from this. |
Interview | Interview me one layer at a time, starting with the objective. Ask one question at a time. Do not move on until the objective has one number and one date. |
Objective check | Here is the objective: [text]. Work backwards from it to the number of leads, visits or bookings it needs, using my conversion figures. Show the arithmetic as a table. |
Audience | For each audience segment, write what they believe today about [category] and the one barrier that stops them acting. Three segments at most. If I gave you more, tell me which to merge. |
Message | Write five candidate core messages as sentences a customer could repeat, none containing "and". For each, write the strongest objection from the audience. Recommend one. |
Tree | Write the campaign map as an indented outline: campaign name at the centre, seven branches, at most five sub-branches each, every label three to eight words. Do not build anything yet. |
| Stage | Prompt to paste |
|---|---|
Overview | Using Storyflow, open the project "[project name]". Show me the board, then build the outline we agreed as one mind map with "[campaign name]" at the centre. |
Sections | To the right of the mind map, build seven sections in this order: Objective, Audience, Message, Channels, Moments, Assets, Measurement. Use notes for the first three. Use tables for the last four with the columns I list next. |
Channels table | Channels columns: channel, job, audience segment, budget, share of budget, owner. Split the budget of [amount] and show the total row. If I gave no budget, write "[allocate]" in every budget cell. |
Moments table | Moments columns: date, moment, message emphasis, channels, owner. Write rows in date order. Use my fixed dates [dates] and propose the beats between them. Flag any gap longer than ten days. |
Assets table | Assets columns: asset, format, channel, message or proof point it carries, owner, due date. List only assets that carry a message or proof point from the Message section. |
Measurement table | Measurement columns: metric, role (primary or diagnostic), baseline, target, source. One primary metric. At most four diagnostics. Write "[baseline needed]" where I gave no number. |
Open slots | Add a note at the top of the board titled "Open slots" listing every bracketed item on the board, each with who should answer it. |
| Stage | Prompt to paste |
|---|---|
Traceability | Read the whole board. List (1) assets that carry no message, (2) messages or proof points that no asset carries, (3) channels with no asset or no stated job. Do not change anything yet. |
Budget cut | If the budget dropped by a third, which channel and which assets would you cut first, and what would the objective have to become? Show the revised arithmetic. |
Overlap | Which two channels reach the same segment with the same job? Recommend which to keep and why. |
Apply | Apply the cuts we agreed to the Assets and Channels tables. Tell me exactly which rows you changed. |
What changed | Show me what changed on this board in the last week and who changed it. Then re-run the traceability check against the board as it is now. |
Client version | Build a slide deck on the board from this map: one slide per layer, objective first, with the moments table as the timeline slide. Use only what is on the board. |
Two notes on the prompts. "Do not change anything yet" matters in the audit: you want the lists before the edits, so the cuts are your decision. And the last prompt is where a map becomes a pitch. The fuller method for that is in our guide to presenting a campaign concept built with ChatGPT or Claude.
Both work, and the Storyflow connector gives both the same tools. The differences are in setup and plan, not in the map you end up with.
Claude allows one custom connector on its Free plan, so you can connect Storyflow and build a map without paying Anthropic anything. Claude's pricing page lists Pro at $17 a month on an annual subscription or $20 monthly, which raises usage limits. Adding the connector is a settings screen and a sign-in.
ChatGPT needs Plus or above with Developer mode. After that the connector behaves the same way. If your campaign thinking already lives in ChatGPT, with projects and past conversations about the brand, that context is worth more than a simpler setup. Stay where the thinking is.
Map in the assistant that already knows the campaign. If you are starting cold and have no paid assistant plan, start in Claude. A longer comparison for planning work is in Claude vs ChatGPT for creative project planning.
Prices are the vendors' published annual-billing rates in October 2026. The assistant's own subscription is separate and the same in every row, so it is left out.
| Tool | Solo, per month | Team of five, per month | How it is billed |
|---|---|---|---|
Storyflow Plus | $7.99 | $7.99, with four invited collaborators joining free | Flat, per account |
Storyflow Max | $39 | $39, with a team workspace, roles and permissions | Flat, per account |
Miro Starter | $8 | $40 | Per member |
Whimsical Pro | $10 | $50 | Per editor |
Notion Plus | $10 | $50 | Per member |
Canva | See Canva's pricing page | See Canva's pricing page | Varies by plan |
Miro, Whimsical and Notion each have a free plan you can start on today, and Storyflow does not. If the test is "can one person try this for nothing this afternoon", Miro's Free plan with 3 editable boards and its connector included is the honest answer.
The table also hides connector limits. Miro caps tool calls per day by plan. Storyflow caps an invited free account at 25 actions a week and puts no connector limit on paid plans. Neither cap is likely to matter for a single map; both matter if the assistant is reading and revising boards every day.
One person on a paid plan can invite the rest of the team to the campaign board, and those people join free. An invited free account can also connect an assistant, within 25 actions a week. For a small team with one planner and several contributors, that makes the real cost one Plus subscription.
Use Storyflow. The map, the references, the content plan and the pitch deck end up on one board the assistant can read. This is the situation the tool is built for.
Use Miro. If eight people are going to argue about the audience on a call with sticky notes, that is Miro's home ground, and its connector reads the comments afterwards.
Use Whimsical. It is the fastest setup, and it builds mind maps and flowcharts well. Move to something with tables when you need dates and owners.
Use Notion for the assets and moments databases, and accept that the overview is a page. Some teams keep the visual map on a board and the task database in Notion, and that split works if one person owns keeping them in step.
Use Canva for the slide, built from a map that lives somewhere structured. Do not make the Canva design the source.
For a two-week campaign on one channel with one audience, write the seven layers as seven lines in a document. A map pays for itself when there are enough channels and assets that you can no longer hold the connections in your head. Our guide to why mind maps break past 100 nodes covers the opposite end of the scale.
One prompt gets you the average campaign for your category: three personas, five channels, eight KPIs. It looks finished and contains nothing you decided. The interview in step 3 is slower by fifteen minutes and is the only part that puts your campaign on the board.
A mind map node holds a label. If you ask for audience descriptions as nodes, you get either truncated labels or an unreadable web. Labels in the mind map, paragraphs in the sections.
A plausible budget nobody agreed to gets approved by people who assume someone else checked it. Give the number or get a slot. The same goes for launch dates, baselines and any statistic in a proof point.
A map without dates cannot be sequenced, and the team will sequence it by whichever asset is finished first. Even three beats (open, peak, close) change how the assets get prioritised.
Assets are fun to list. If the message changes after the asset table exists, every row needs rechecking, and nobody rechecks. Build the layers in order. The audit exists because this mistake is common.
When something changes, the instinct is to ask for a new map. A second mind map lands next to the first, and now there are two. Ask the assistant to read the board and change specific rows, and remove outdated elements yourself.
The map shows what the campaign is and why each piece exists. It does not assign Tuesday's tasks. From the map, build the working plan: a content calendar, a production schedule, or a launch plan like the one in our guide to planning a product launch with Claude or ChatGPT.
Storyflow is our product, and it is the wrong pick in these cases.
Storyflow is paid-only during early access, and its Free plan launches before the end of 2026. Unless a paid member invites you to their board, you pay from day one. Miro, Whimsical and Notion all have free plans now.
The Storyflow connector does not read comment threads. If the review process is thirty comments on a board that the assistant should triage and answer, Miro's connector does that and Storyflow's does not.
Storyflow does not publish or schedule posts, does not pull performance data, has no calendar and does not sort tables. Dates live in table cells and tickets. If you need a campaign calendar that reschedules itself or a dashboard fed by ad platforms, you need a marketing or project management system next to the map.
There is no approve button and no versioned approval trail. For regulated categories or clients who require recorded sign-off on every asset, use a proofing tool for that step.
Storyflow runs in the browser, with no desktop app, no mobile app and no offline mode.
Adding Storyflow to ChatGPT uses Developer mode, and adding it to Claude uses a custom connector. If your IT policy only allows connectors from an approved directory, check before you plan around it. Our guide to connector permissions and safety covers what to ask.
Paste the table below into ChatGPT or Claude with the instruction "Fill this for my campaign using only what I have told you; leave unknowns as bracketed slots", then run the build prompts.
| Layer | Fill in | Test it must pass |
|---|---|---|
Objective | [What changes] by [how much] between [start date] and [end date] | One number, one date, no "and" |
Audience | Segment: [who]. Believes now: [belief]. Barrier: [what stops them] | Three segments at most, each with a belief |
Message | Core message: [one sentence]. Proof: [point 1], [point 2], [point 3] | A customer could repeat it; every proof point is true and sourced |
Channels | [Channel]: job [reach, prove, convert or remind], segment [who], budget [amount] | No two channels share a job and a segment |
Moments | [Date]: [moment], emphasis [which proof point], channels [which] | No unexplained gap longer than ten days |
Assets | [Asset]: format, channel, carries [message or proof point], owner, due | Every asset carries something from the Message layer |
Measurement | Primary: [metric], baseline [n], target [n], source. Diagnostics: up to four | One primary metric that restates the objective |
For a campaign with a budget under about $20,000, a healthy map is small: one objective, two audiences, one message with three proof points, three to five channels, four to six moments, and fifteen to twenty-five assets. Those figures are a rule of thumb from the layers' own tests, not a benchmark. If your map is much larger, the audit will usually find the excess.
You can build the same seven layers by hand, or with Storyflow's built-in agent, starting from the AI marketing campaign planner or the AI mind map generator. The method in our tool-agnostic guide to mapping a marketing campaign applies unchanged.
A campaign map is seven layers and the lines between them, and its job is to show you which assets have no reason to exist before you pay to make them. An assistant is good at drafting the layers and bad at keeping them anywhere useful. Put the two together: do the thinking in ChatGPT or Claude, and have the assistant build the map on a board it can read back.
For that job we recommend Storyflow, with the limits stated above: no free plan yet, no comment reading through the connector, and no execution features. If one of those limits is your requirement, Miro is the alternative we would pick first. Either way, run the audit. To set it up, start at the Storyflow MCP page.
Storyflow is the simple AI tool we recommend for mapping out marketing campaigns. You connect it to ChatGPT or Claude once, describe the campaign in the chat, and the assistant builds a mind map and seven sections on a board your team can edit. Miro suits teams already running workshops there, Whimsical suits flowchart-heavy maps, and Notion suits plans that must be databases.
Yes, in two ways. On its own, ChatGPT writes the map as an outline in the chat, which is useful for thinking and hard to share or keep current. With a connector such as Storyflow's, ChatGPT builds the map as a mind map, notes and tables on a board. Adding Storyflow needs ChatGPT Plus or above with Developer mode.
Yes. Claude can interview you through the seven layers and, with the Storyflow connector added, build the map on a board and read it back for an audit. Claude's Free plan allows one custom connector, so you can connect Storyflow without a paid Claude plan. Usage limits on Claude's Free plan are lower than on Pro.
The seven layers are objective, audience, message, channels, moments, assets and measurement. The objective is one number and one date. Audience and message define who should believe what. Channels and moments define where and when. Assets are what gets made, and measurement is how you will know. Every asset should trace back through the layers to the objective.
Partly. A mind map is the right shape for the overview, because each node holds a short label and the whole campaign fits in one glance. It is the wrong shape for budgets, dates and owners, which need columns. Use a mind map for the overview and sections with tables for the working detail, on the same board.
Plan for about an hour for a small campaign: ten to twenty minutes answering the assistant's interview questions, a few minutes for it to build the board, and the rest reading the result and running the audit. That is an estimate, not a measured figure. The interview is the part you should not shorten.
Tell it at the start to use only the facts you give and to write anything missing as a slot in square brackets, such as "[set date]" or "[baseline needed]". Then ask for an "Open slots" note listing every bracket on the board. You get a map with visible gaps and an agenda for filling them.
A traceability audit checks that every asset carries a message, every message has an asset, and every channel has a job. With the map on a Storyflow board, you ask the assistant to read the board and report those three lists. Assets that carry no message are candidates to cut. Messages with no asset are gaps to fill.
Not during early access. Storyflow is paid-only today: Plus is $7.99 a month billed annually, Pro $14 and Max $39. Its Free plan launches before the end of 2026. Anyone a paid member invites to a board joins free now, and that invited account can use the connector for up to 25 actions a week.
No. The connector runs on your assistant's own model in ChatGPT or Claude, so it does not spend Storyflow AI credits. Paid Storyflow plans have no connector limit. A free account, which today means one invited by a paid member, gets 25 connector actions a week, counting every read and every change, resetting on Monday UTC.
Yes. The connector reads the board as it stands, including edits your team made by hand, and it can report what changed over the last day, week or month and who changed it. It can then edit notes and table rows in place. It cannot read comment threads, so paste comments into the chat if you want them acted on.
Yes. Miro's MCP server works on every Miro plan, including Free, and its documentation lists Claude and ChatGPT as supported clients. It creates diagrams, documents and tables and reads comments. Its tool documentation does not name a mind map tool, and tool calls are capped per day by plan, starting at 100 on Free.
We found no connector for Milanote, and its plans page lists no AI features. It remains a pleasant board to arrange by hand, with a free tier of 100 notes, images or links and 10 file uploads, per Milanote's plans page. To use it for a campaign map, you would build the map yourself from the assistant's outline.
A brief is the ask: the problem, the objective, the audience and the constraints, written for whoever will make the work. A map is the answer's structure: how the message reaches each audience through channels, moments and assets. Write the brief first when a client or stakeholder commissions the work, then map from it.
Yes. A Storyflow board has a presentation mode that runs fullscreen, and the assistant can build a slide deck on the board from the map, one slide per layer. The client can also open a view-only link with no account and signs in to comment. For a formal pitch, build the deck; for a working session, present the board.
As few as have distinct jobs. Give each channel one job (reach, prove, convert or remind) and one primary segment. If two channels share both, cut one. For a small budget, three to five channels is a sensible range, because every added channel needs its own assets and takes a share of the spend.
No. The connector has no image generation tool. It can place pictures on the board from web addresses, and it can give you an upload link for a photo, because an image attached in the chat does not reach Storyflow. Image generation in Storyflow belongs to the built-in agent on the Pro and Max plans.
Table of Contents
Plan the whole campaign on one board: brief, audience, channels, and assets connected, with an AI that reads all of it. Open a template and start from real structure.
A visual AI workspace where every feature lives inside one canvas. No tab-switching, no context lost.
Build your entire board from a single message
Type what you need in the AI chat at the bottom of your canvas. The AI adds cards, headings, and structure directly onto your board.
Use expert frameworks as AI context
Type @ in the AI chat and choose any Tactic. The AI tailors every response to that framework instead of giving generic advice.
Turn your board into a mind map in seconds
Ask the AI to restructure your canvas as a mindmap. It connects your ideas into a visual hierarchy so you can see how everything relates.
Storyflow actually began as a personal tool while working on creative and research projects.
We kept running into the same problem: ideas were scattered everywhere: notes, documents, and whiteboards.
Nothing helped us see how everything connected.
So we started building a workspace designed around how ideas actually grow.
→ Read how Storyflow was createdStoryflow Team
Product & Research Team
Published: 2026-10-10
Extremely simple, seriously capable

Watch a real project come together, from a blank board to a finished plan, clicked through step by step.
Transform your creative workflow with AI-powered tools. Generate ideas, create content, and boost your productivity in minutes instead of hours.