Analysis Agent
7-step AI business analyst. Scores viability, revenue potential, and build effort for any idea.
What it does
The Analysis Agent is AppStudio's core feature. It takes your raw idea and runs it through 7 sequential research steps, each building on the last, to produce a complete business viability report — all powered by your local AI, with no cloud calls.
The 7 steps
Breaks the idea into 4 focused research questions and validates the core hypothesis. Sets the agenda for all subsequent steps.
Estimates market size, identifies the most likely customer segments, and gauges willingness to pay at solo-dev tool pricing levels.
Maps the competitive landscape, lists known direct competitors, and identifies your potential market gap — what makes this defensible.
Models revenue options (one-time, SaaS, API licensing), suggests a pricing strategy, and estimates realistic monthly revenue potential at solo-dev scale.
Surfaces the top risks, identifies key opportunities, names your biggest threat, and articulates your unfair advantage if one exists.
Synthesizes all previous steps into three numerical scores and delivers a final go/no-go verdict with reasoning.
For each risk identified in step 5, generates a concrete mitigation strategy with effort and impact ratings, plus a residual risk score after mitigation.
Understanding the scores
Each analysis produces three scores on a 0–100 scale:
- Viability score — How likely this idea can succeed as a real business, accounting for market size, competition, and execution complexity.
- Revenue score — Revenue potential calibrated to solo-developer scale. A 70+ means realistic path to meaningful solo income.
- Effort score — Build complexity. Lower is simpler. A score of 30 means a fast, focused build; 80+ means a long multi-month project.
The verdict
The agent delivers one of three verdicts based on the combined scores and analysis:
- BUILD Strong scores, clear path to revenue, well-understood market. Go build it.
- VALIDATE Promising idea but significant unknowns remain. Run a validation experiment (landing page, waitlist, interviews) before committing to a full build.
- SHELVE Low scores, a too-crowded market, or complexity that outweighs the revenue opportunity for a solo developer.
How to run an analysis
Click any idea card to open the detail view.
Requires your local AI server to be running. AppStudio will warn you if it can't connect — see the LLM Setup guide if needed.
Each step appears as it completes. Total time is typically 2–5 minutes with qwen3:14b on a GPU.
Results appear in the Analysis tab — scores, verdict, full step breakdowns, and risk mitigations.
Tips for better results
- More detail = better analysis. Fill in the "problem solved" and "target audience" fields before running. The agent uses these directly.
- Re-run anytime. Analysis overwrites the previous result — useful after you've refined your idea description.
- Revenue model bias. The agent is intentionally biased toward one-time tool sales and API licensing over recurring SaaS. This is by design for solo developers who want revenue without operational overhead.