// replaces: planning-poker guesswork
Story estimates that learn from your history, not your gut.
Estimo will train a neural net on your resolved tickets and predict the size of new ones — story points or hours, your call — with a confidence band calibrated to your team instead of a generic benchmark.
TEAM-1482 — Add SSO to the admin panel
predicted 8 pts · 80% band 5–13 · or hours, your call
nearest shipped stories: #1203 (8 pts) · #1377 (13 pts) · #1051 (8 pts)
How Estimo will work
Planning poker is a vibe check. Your estimates are wrong in the same direction every sprint and nobody trusts the burndown.
Straight up: Estimo isn't built yet — it gets built the moment enough teams raise their hand. Here's the plan. Estimo learns from your resolved Jira stories — titles, descriptions, labels, components — and predicts the size of a new ticket. Pick the unit your team actually speaks: story points for relative sizing, or hours if you log real effort. No history yet? You start on a model pretrained across public datasets, and it calibrates to your team's reality as you close (and log) work. New story comes in, you get a grounded estimate and a confidence band — before anyone argues about points. It ships as a single Claude skill: no integration project, drop it in and you're estimating in minutes. Founding teams get it trained on their data with me at the keyboard — and pick what it plugs into.
what it does
- ▸Estimate in story points or hours — pick the number your team runs on
- ▸Gives a confidence band, not a single false-precision number
- ▸Calibrates to your history; warm-starts from a pretrained model on day one
- ▸Trained on your resolved tickets — titles, descriptions, labels, components
- ▸Ships as a single Claude skill — drop it in, estimating in minutes
// honest answers
Questions you should be asking
Is this built?+
No — and the page says so everywhere, because pretending otherwise would be a bad way to start a relationship with your team. Estimo gets built the moment enough teams raise their hand. Your email is a vote, not a commitment.
What happens after I vote?+
If enough teams want it, I build it and you're first in line with the founding terms. If not enough do, I tell you that too — no drip campaign reviving it every quarter. Either way you hear from me once, with the answer.
What data would it need from us?+
Your resolved tickets — titles, descriptions, labels, components — from Jira to start. Story points work from day one (warm-started from public datasets, calibrated to your history as you close work). Hours mode needs worklogs, so it's for teams that log effort. Cross-team model sharing would be strictly opt-in, sharing model updates rather than your tickets.
Want Estimo to exist? Raise your hand.
Drop your email to vote. If enough teams want it, I build it — and you're first in line, with the same beta deal: six months free and the founding price locked.
- 01
Your vote is counted
one email, one tally mark — that's the whole commitment.
- 02
I report back either way
built or not built, you hear the answer once. No drip.
- 03
If it gets built
you're first in, on founding terms.
Joe Christianson
the developer building this — you email, I reply. No team hiding behind this form.
// different pain?