See what we found.
Question what we assume.
brighterdesk uses OpenAI and live web search to produce a provisional assessment. A polished report must not disguise missing evidence.
Two methods. Two kinds of evidence.
Start: AI research uses external web sources and source links. Check important claims with the original source and test demand with customers.
Business: AI business reviews use confirmed financial observations, operational context, calculations and dated snapshots. Uploaded reports are extracted by OpenAI and checked by you before use. Every figure retains its source, period, currency and accounting basis. Arithmetic runs in application code; AI explains the evidence, alternatives and priorities. Missing figures remain unknown. Business usage is separate from Start credits.
Review records track what changed after you acted. The app supports accounting report connections when the provider is activated; new imports require confirmation. Companies House filings are historical, not a current management-accounting feed. An assessment is not audited accounts or verified due diligence.
Start with founder fit
Budget, available hours, skills, target market and selling preferences shape the research. Discovery also asks about customer access and exclusions. A theoretically attractive idea can still be a poor choice for a particular founder.
Use sources for factual claims
The research instructions prefer official product pages, published pricing and authoritative primary sources. Reports include source links and a research date. A response without usable web citations is rejected rather than presented as completed research.
This check establishes that citations were returned, not that every interpretation is correct. Sources may be outdated, incomplete, inaccessible or misunderstood. Open the important ones before making a decision.
Keep estimates separate
Setup costs, time requirements, conversion rates and financial scenarios may be assumptions or estimates. They are not supplier quotes or forecasts. If a price cannot be established, the research should say so rather than fill the gap with a plausible number.
What we do not establish
- Exhaustive competitor coverage or verified market size.
- Customer willingness to pay, unless you independently test and record it.
- A probability of success or guaranteed revenue.
- Professional legal, tax, investment or regulatory advice.
Your plan evolves; the original stays available
Saved projects keep the original research and source register. Your working plan and evidence notes are editable. Those edits are user-provided and are not independently verified or automatically re-researched.