Authority tools
Plan the AI project around the evidence required to release it
Choose the system you are building. Answer the evidence questions yourself or paste an existing brief for a reviewable head start, then receive a budget band, timeline, readiness score and production risks.
Choose a project track
One method, four production contexts
Each estimator shares a versioned scoring engine, then asks the track-specific questions that actually change the build.
New: brief-to-evidence assistance. Paste a project brief to prefill only the answers it supports. You still review every choice, and the versioned rules—not the language model—produce the range.
Generative AI Project Estimator
Estimate the build budget, timeline, readiness, evaluation burden and production risks for a RAG, assistant, document AI or controlled agent project.
Estimate this project 02Vision AI Project Estimator
Estimate the build budget, timeline, feasibility evidence and integration scope for visual inspection, safety monitoring and edge computer-vision systems.
Estimate this project 03AIoT Project Estimator
Estimate the build range, timeline, evidence and operational scope for sensor AI, predictive maintenance and locally intelligent device fleets.
Estimate this project 04Custom AI Project Estimator
Estimate the implementation budget, timeline, readiness and required evidence for a custom AI application across text, vision, sensor and multimodal workflows.
Estimate this projectWhat comes back
Enough structure to decide the next move
Readiness, not theatre
A 100-point score shows whether the outcome, data, interfaces, release gate and owner are ready for commitment.
A published band
The rules select one existing Pran engagement band. They do not multiply assumptions into a falsely precise price.
Production risks
The summary exposes the three constraints most likely to change the scope or invalidate a larger build.
A scope record
The emailed PDF preserves assumptions, phases, architecture pattern, acceptance criteria and evidence still required.
A deliberate boundary
This estimates the project required to get to production
It covers feasibility, implementation scope, integration complexity, acceptance criteria and build budget.
It deliberately excludes model-provider consumption, token economics, private-inference sizing and engineering-team rate cards. Those are different questions that need measured workload data.
See measured engineering reports