How does a sprint work?
A sprint is a fixed-scope engagement that results in a working AI assistant deployed in your operation. Each sprint has a defined start date, a defined deliverable, and a defined end date — no open-ended consulting.
The sprint process
Week 1: Discovery and design We map your workflow in detail — the inputs, the rules, the exceptions, the systems involved, the people affected. We design the assistant's behaviour and get your sign-off before building anything.
Weeks 2–3: Build and integration We build the assistant and connect it to your systems. We test it against your real documents and data. We iterate with your team to tune the behaviour before go-live.
Week 4: Deployment and training We deploy the assistant to your live environment. We train your team on the dashboard and monitoring tools. We monitor performance in the first weeks after launch.
Post-launch: 30 days of support All sprints include 30 days of post-launch support. If anything behaves unexpectedly in production, we fix it.
What you need to provide
- A named internal contact who can attend discovery sessions
- Sample documents for training and testing the assistant
- Access to the systems the assistant will integrate with
- 3–4 hours of your team's time in Week 1 for discovery workshops
We handle everything else.
Sprint tiers
There are three tiers based on scope — Proof of Concept, Process Transformation, and Multi-Process Rollout — with timelines ranging from 2 weeks to 45 days. See the pricing page for full details.