AI scores 7 dimensions 1–5 → complexity band → indicative timeline + cost + retainer
Describe one task the way you'd explain it out loud and let AI score each of the 7 things that drive build cost — the data, the cleanup, the sources, the decision, storage, the cost of being wrong, and checking. The scores blend into a complexity band (Simple / Medium / Hard / Very hard) and an indicative build timeline, cost, and ongoing-retainer recommendation. Adjust any score yourself; everything updates live.
Replay the last time you actually did this task: what you started with, the systems you touched, the calls you made, what happens if it's wrong. The more detail, the sharper the score. (Or skip this and rate the seven stages yourself.)
What lands in front of you when the task begins?
How much fixing/tidying before the data is usable?
Which systems do you pull from — and do they have ready connections?
How much real judgment, and how many edge cases?
What has to be saved — including half-finished work?
If it's late, wrong, or inaccurate — how bad is that?
Who can tell if the result is good, and how?
Scope it properly and keep a human in the loop while it beds in.
Indicative only. Most of the calendar time is turnaround on your side — billed effort is roughly a third of it (active dev) plus communication and info-prep, at a 0.3 Senior AI Architect · 0.5 Product Manager · 0.2 AI Engineer blend (≈$1,600/day). A scoping call firms it up.
Recommended for a medium build: Silver.
Basic upkeep — keep libraries and scripts current (excludes third-party tooling costs).
Observability + debugging of edge cases as they surface in production.
Active iteration — stacking new functionality and complexity layers over time.
Not a department — one task you'd want off your plate, like "after a call, write up the notes and update the CRM." Replay the last time you actually did it: what you started with, the systems you touched, what happens if it's wrong. The more detail, the sharper the score.
AI reads your description and scores each stage 1 (simple) to 5 (complex) with a one-line rationale: how messy the data was, how much cleanup it needed, how many sources you pulled from, how much judgment the decision took, what had to be stored, how bad it is if it's wrong, and who can judge the output. Each stage shows a concrete simple-vs-complex example, and you can override any score — the blended total updates live.
Your scores roll up into a band — Simple, Medium, Hard, or Very hard — and an indicative build timeline (2–4 weeks for Simple up to 15+ for Very hard), an estimated cost (a blended ~$1,560/day team), and a recommended retainer. We aim for a proof-of-concept in 3–4 weeks and a first production run within ~2 months, then stack complexity in iteration cycles.
The band maps directly to build effort. Drop it into the AI Roadmap Generator as the complexity input and it becomes the scheduling weight — so a backlog of scored tasks turns into a realistic, capacity-aware plan.
A free tool gives you a hypothesis. The 30-minute diagnostic is where we pressure-test it against your actual workflows — and decide whether the project is worth building, buying, or skipping.