Storyflow has one AI assistant for the whole app. The same assistant follows you from the dashboard to a whiteboard to a document, and it can see whatever you have open, so you rarely have to explain where you are or paste your own work back to it.

This article is the short answer to "what can it actually do".

Where it is

Click the AI button in the top navigation to open the assistant panel. It stays with you as you move around: open it on a board and it can build on that board, open it in a document and it can write in that document, open it on the dashboard and it can answer questions across your projects.

Answering questions about your own work

It reads your real content rather than guessing: your projects, the boards inside them, the elements on a board, your documents and your Tactics. Ask what is on a board, what a wall covers, what is still unfinished, or what two documents disagree about, and it will go and look before answering.

Reading things you give it

• Files on your board or attached to a message: PDFs, spreadsheets, CSVs and documents. It reads the actual contents, so it can summarise them, pull rows out of them, or turn them into a table or a task list.
• Links: paste a URL and it opens the page and works from what is really there.
• Recordings: audio on your board can be transcribed, so a voice note becomes text you can build from.
• Images: select images on the canvas and it can see them, describe them, label them, or use them as the basis for a new picture.

Researching the web

It can search the web live and cite the sources it used, which matters for anything with a date on it: prices, current tools, recent events, statistics. Source links appear alongside the answer so you can check them.

Building on a canvas

Ask in plain language and it builds real elements you can then move, edit and delete yourself: notes, headings, to-dos, tickets, walls, tables, kanban boards, colour swatches, documents, folders, and arrows between things.

Generating and editing images

Describe a picture and it is generated straight onto the board. It also edits: give it an existing image as a reference to restyle it, change something in it, or combine several images into one. Point it at an empty image frame and it fills that frame rather than leaving a duplicate beside it.

Making visuals and structured documents

Beyond plain cards, it can produce charts and dashboards, mind maps, flowcharts, sequence diagrams, storyboards, shot lists, character sheets, planners, progress trackers, moodboards, and small self-contained interactive widgets.

Editing what is already there

It is not only for making new things. Ask it to rewrite a card, expand one with more detail, split one card into several, merge several into one, shorten something, recolour a wall (which re-themes everything inside it), add labels, caption an image, reorder items, or move things between walls.

Writing in documents

With a document open it writes directly into the document: adding a section, rewriting one section without touching the rest, replacing a table, retitling it, or rewriting just the part you have highlighted.

Workflows: jobs it already knows

Many jobs have a workflow — a recipe the AI already knows, so you pick it instead of describing the whole thing. Shot lists, character sheets, campaign briefs, colour palettes, mind maps, dashboards, competitor teardowns and many more.

Reach them three ways, all reading the same catalogue: the browse panel with filters, the slash palette (type / in the chat), or the quick tools and starter buttons. Twenty of them were distilled from boards actually built in Storyflow, so the structure is a real one rather than a generic grid.

Controls worth knowing

• The scope bar names the board and project it is working in, and the scope follows the turn rather than whatever page you navigate to. You can point it at a different board.
• The Effort dial sets how much a single turn takes on, from a tight step-by-step reply to a full build in one pass.
• Memory keeps what it knows about how you work, filed with a path and a record of where each part came from. You can browse and edit it as a graph.
• A read on your work. Ask it to review a board, a document or a selection and it leaves notes on the cards themselves, each with a next step.

What it will not do

• It works inside Storyflow. It cannot sign in to other tools or push files into them. It can produce the content, and you export or copy it out.
• It does not invent links or statistics. If it has not read a page or researched a claim, it will say so instead of producing a plausible URL.
• It asks rather than guesses when a request is genuinely ambiguous, using a short form in the chat — but only for two rounds per turn. After that it proceeds and states in a card what it assumed, which is the cheapest place to redirect it.
• Its work is yours to undo. Everything it creates, edits or removes behaves like your own editing: one undo reverses a turn, and deleted items sit in the board's trash for 7 days.

Tips and tricks

Select first, then ask. Whatever you have selected on the canvas is what "these", "this" and "it" refer to. Selecting three notes and asking to expand them edits those three rather than adding new cards beside them.

Ask it to look before it builds. For anything factual, prompting it to research first produces noticeably better content than letting it work from memory, because it builds from specifics rather than generalities.

Point at an empty frame. Make an empty image frame where you want a picture, select it, and ask for the image. It fills that frame in place.

Give it your files. A board that already holds the brief, the transcript and the spreadsheet gives much better answers than a prompt describing them.

Good to know

AI usage counts against your plan's allowance, so what scales with a higher tier is how much you can generate, not which features you can reach.