The short version: upload anything that already contains real thinking about your project — briefs, research, specs, decks, financials, meeting notes. PA-AI parses these to auto-populate fields throughout the modules, so the more grounded source material you provide, the less you'll have to type from scratch.
Supported file formats
PA-AI accepts the following formats:
PDF (.pdf) — best for finalized documents, reports, and proposals
Word (.docx) — best for briefs, narratives, and editable drafts
PowerPoint (.pptx) — best for decks and presentations
Excel (.xlsx) — best for financials, pricing models, and structured data
Images (.png, .jpg, .jpeg, .svg) — best for mockups, sketches, and reference visuals
Plain text (.txt) — best for raw notes or transcripts
Most documents you'd attach to an email or share in a kickoff folder will work.
What to upload, by category
Design and product documents
Anything that defines what you're building or how it should look and feel.
Industrial design briefs — what the product is, who it's for, the design intent
Product requirements documents (PRDs) — features, constraints, success criteria
Spec sheets — dimensions, materials, technical parameters
Style guides or brand guidelines — visual standards, voice, brand rules
Mockups, renders, sketches — what the product looks like or might look like
Wireframes and prototypes — interaction flows and UI
Why these help: They give PA-AI a clear picture of the product itself, which feeds directly into Definition, Branding, and Constraints.
Research and discovery documents
Anything that captures what you've learned about the market, users, or problem space.
User research reports — interview summaries, survey results, usability findings
Market research — sizing, trends, audience studies
Competitive analyses — teardowns, benchmarking, positioning matrices
Customer personas — segments, jobs-to-be-done, pain points
Discovery synthesis decks — themes, insights, opportunity areas
Why these help: PA-AI uses these to ground Demographics, Psychographics, and Competitive Landscape in real evidence rather than assumptions.
Strategy and business documents
Anything that frames the why, the business case, or the path to market.
Strategy docs or one-pagers — vision, positioning, north star
Go-to-market plans — launch sequencing, channels, target segments
Business cases or investment memos — opportunity sizing, rationale
Pricing models — tiers, costs, margins
Roadmaps and timelines — milestones, sequencing, dependencies
Why these help: These shape Definition (especially the go-to-market and timeline questions), Operations & Expansion Strategy, and Financial Planning.
Technical and operational documents
Anything that captures the hard constraints or system realities.
Technical specifications — system requirements, integrations, dependencies
Cost analyses or COGS breakdowns — manufacturing, operating, unit economics
Manufacturing or supply chain docs — sourcing, production constraints
Compliance or regulatory requirements — standards, certifications, legal limits
Engineering or architecture documents — how it's built or will be built
Why these help: These populate Constraints and feed Operations & Expansion Strategy and Policy/IP/Compliance Risk.
Meeting notes and decisions
Often the most underrated source of context.
Kickoff meeting notes — what was decided, who's involved, what's expected
Stakeholder interview transcripts — internal POVs from leadership, sales, support
Decision logs — what was chosen and why, what was rejected
Workshop outputs — affinity maps, alignment notes, prioritization exercises
Why these help: Meeting notes carry the "soft" context that polished documents often strip out — disagreements, open questions, who pushed for what. PA-AI surfaces these into more nuanced analysis.
What not to upload
Some files are technically supported but won't help (and may dilute the analysis):
Generic templates with no content filled in — empty briefs, blank decks
Files unrelated to this project — content from other projects, old archives
Personal or sensitive data — financial records, HR documents, customer PII not relevant to the work
Marketing collateral from competitors unless you're using it as a competitive reference (and even then, prefer your own analysis of it)
Very large reference documents you haven't read — uploading a 400-page industry report PA-AI will struggle to extract focused signal from. Pull the relevant sections into a separate doc first.
Tips for great inputs
More context beats less. If you're not sure whether something is relevant, upload it. PA-AI is better at filtering noise than guessing at gaps.
Name your files clearly. "Design Brief v3 Final" is more useful than "doc-1.pdf". File names give PA-AI extra context about what each file is.
Include versions when relevant. If a project went through multiple drafts, upload the final version plus key earlier drafts — the evolution itself is signal.
Upload at the project level, not the module level. Files uploaded during initial ingestion get parsed once and reused across all 7 modules. Module-level uploads only inform that specific module.
Re-upload if your inputs change. If you swap in a new brief or get fresh research mid-project, add it to Project Documents and re-run any affected modules.
Still not sure what to upload?
If you have something and aren't sure it'll help, upload it — worst case it gets ignored. If you have nothing, the Suggest answer button on each question field lets PA-AI draft starting points based on general patterns, which you can edit. But the more you give it upfront, the sharper the output.
For specific questions about what to upload for your project, hit the chat bubble and we'll help.