How we work
Every engagement starts with a baseline and a test set, so you can see whether the work helps before you rely on it. You work with And directly, from the first call to handoff.
Five steps, from the first call to support
- 01
Intro call
Free, 30 minutes
What happens
Talk through the process or problem and the goal. Check fit honestly.
What you get
A yes or no, and a suggested first step.
- 02
Discovery
Paid, fixed price
What happens
Map the process and the data, measure the baseline, list options (including "AI isn't the right tool here") and agree on success metrics.
What you get
A written plan with architecture, risks, success metrics and a fixed quote for the build. [TODO: whether the fee is credited to the build]
- 03
Build
Fixed scope
What happens
Working software early, a short demo every week, a test set built with you and approval steps where judgment matters.
What you get
A working system in your own accounts, measured against the baseline.
- 04
Launch and handoff
What happens
Go live, train the people who use it and document it.
What you get
A runbook, documentation, a recorded walkthrough and [TODO: days] of fixes included.
- 05
Ongoing support
Optional, monthly
What happens
Monitoring, re-running the test set when models or data change, improvements and new workflows.
What you get
A short monthly report on what ran, what broke and what improved.
Four ways to work together
[TODO: And decides whether to publish starting-at figures or ranges]
Discovery only
For teams that want a plan before committing.
Fixed-scope project
One process, agent, assistant or pipeline, quoted after discovery.
Monthly support
Month to month, after a project.
Team training
Workshops on AI workflows or data foundations, drawing on And's graduate courses.
What you can count on
Measure before and after.
Every engagement starts with a baseline and a test set, and results are reported against them.
Your accounts, your data, your code.
Systems run in your accounts where possible, and you own everything delivered.
People approve what matters.
Agents act only within agreed permissions, and sensitive steps wait for a person.
No lock-in.
Everything is documented so your team can run it without Innervu.
Plain language.
Weekly updates a non-technical owner can read in two minutes.
An honest no.
If a simpler tool or no AI is the better answer, we say so.