How to size new work with AI

How to Accurately Size New Work in Minutes with Socratic's Agentic PM

Every engineering leader knows the moment: the business team drops a new initiative — "this is what we want to get done" — and suddenly the clock starts ticking. Now it's up to engineering to estimate how long it'll take to deliver.

Usually that means huddling senior devs, recalling past projects, and arguing over what qualifies as "medium" or "large." It's part art, part educated guess — and almost always a time sink.

With Socratic's Agentic capability, that entire ritual can happen in minutes — and with far more accuracy.

---

From Guesswork to Ground Truth

Instead of relying on gut feel, Socratic draws on everything it already knows about your team: historical actuals by person, by team, and by type of work. Using those baselines, it predicts how long new work will take to complete — automatically.

Here's how it works.

I start with a simple scope of work: three epics, each with a rough T-shirt size (small, medium, large, extra large).
I assign them to teams already defined in Socratic, flag any dependencies, and indicate where the work will live in Jira.

That's it.

From there, the agent factors in:

• Historical throughput for those same people and teams
• Current capacity based on what's already on their plates
• The shape and complexity of the new work

Then it runs a Monte Carlo simulation — up to a thousand model runs — to forecast the most realistic completion range. You get an early case, a mid case, and a late case, all visualized in a clean, interactive chart. If multiple epics are involved, Socratic generates a Gantt view that shows sequencing and dependencies automatically.

---

"Can We Deliver by September?" Ask and You'll Know

Let's say the projected completion date comes back too late. Traditionally, that kicks off hours (or days) of debate about cutting scope or adding people.

With Socratic, you just ask:

"What would it take to get this done by September 1?"

The agent immediately models the trade-offs: how much scope you'd need to reduce or how many additional resources to add to meet that date. In seconds, you can simulate multiple scenarios and make a call based on real data, not intuition.

---

Why It Matters

For engineering leaders, this marks a shift from reacting to modeling.
Instead of waiting for a planning meeting to end in consensus-driven estimates, Socratic gives you probabilistic forecasts built from your own history.

You're no longer estimating — you're simulating.

This is what happens when estimation becomes an agentic process:
fast, factual, and collaborative. It's project sizing powered by the data you already have, guided by an AI that learns from how your team actually works.