Documentation
Workflows, patterns, and usage guides for getting the most out of AI persona swarms.
Step 01
Start by creating a Market. Add competitors, watch items (trends to monitor), and a scan interval. This gives the system context for understanding your competitive landscape. Markets are the foundation — they inform persona generation and provide the backdrop against which your proposals are evaluated.
Step 02
Build your persona panel. Each persona has a demographic profile, psychographic traits, goals, pain points, and archetype tags. You can create them manually (custom), define internal stakeholders, or use auto-generation to fill gaps based on your market data. A good panel covers 4–6 distinct archetypes.
Step 03
A project encapsulates a business question or proposal you want to test. Link it to a market, write a manifest (hypothesis + success criteria), and select which personas will participate. Projects can have multiple sessions and branches.
Step 04
Submit your question to the persona panel. Each persona responds based on their profile — goals, pain points, and psychographic traits shape their reaction. The system produces a threaded conversation showing each persona's perspective, concerns, and level of support.
Step 05
Every completed session generates a structured artefact — a report summarising overall sentiment (positive/negative/mixed), per-persona breakdowns, key themes, and prioritised recommended actions. Reports can be viewed in report mode or presentation (slide) mode, and exported as Markdown.
Step 06
Share projects, markets, and personas with team members or make them globally available. Branch from any session to test variations — different personas, adjusted queries, follow-up questions. Each branch preserves the original context while exploring alternatives.
Type / in the chat drawer to access quick commands. These are available on every section.
/generate-reportGenerate a structured report from the current session data/create-personaStart creating a new persona (opens the form or guides via chat)/scan-marketTrigger a market scan to check for new competitor signals/helpShow all available commands and usage tipsReal strategy is not linear. When a persona raises a concern — say, Marcus flags a security blocker — you may want to follow up with just the enterprise personas, or re-ask with a modified question. Branching lets you do this without losing the original session.
How branching works:
Branches can also be created conversationally via the Project Assistant chat drawer — describe what you want to explore and the assistant will help set it up.
Swarm Lite is designed for teams. Every entity — markets, personas, and projects — can be shared.
Search for team members by name, email, or role. Select one or more people to share with.
Toggle global sharing to make the item available to your entire organisation.
When sharing a project, choose to share everything or scope it to specific branches and/or personas.
Shared projects surface pending actions to all stakeholders. Anyone with access can approve or reject.
Every section has a dedicated AI assistant available via the chat drawer at the bottom of the main content area. Each section maintains its own conversation thread.
Artefacts are the structured outputs of Swarm Lite. Every completed session can produce artefacts.
Full written analysis with sentiment scoring, per-persona breakdown, key themes table, and recommended actions.
Slide-style view of the same content — each section becomes a card. Useful for stakeholder reviews.
Download any artefact as a .md file for use in documents, PRDs, or version-controlled decision logs.