Our approach
An engineer with your teams. Agents in production.
Three steps, each with its deliverable. You can stop after any of them, and everything we build belongs to you.
The three steps
Diagnostic, build, run.
Step 1: Diagnostic
We meet your teams, map processes, data and systems, then size each case: expected gain, difficulty, risk.
- Interviews with the teams and observation of the real work
- Inventory of data, systems and access
- Each case sized: gain, difficulty, risk
- Deliverable: a roadmap and a first case ready to build
Step 2: Build
A forward deployed engineer builds with your teams, in your tools. Staged rollout: first the agent suggests, then it acts on what you authorize.
- A forward deployed engineer alongside your teams
- Connections to your systems, with your IT team
- Evaluation sets built with the business
- Staged, measured rollout to production
Step 3: Run
We operate, measure and improve: monitoring, evaluations, model updates, new features. You stay in control of permissions.
- Monitoring and alerts
- Evaluations before every update
- New models tested, adopted if they do better
- Changes and new cases, as needs arise
Forward deployed engineer
The person who builds is the person who listens.
A forward deployed engineer is neither a consultant who hands over a report nor a developer waiting for a specification. They spend time with your teams, understand the work as it is really done, and build right away.
Every week, something runs and gets measured. What does not work is fixed while the context is fresh.
More about this model ›- Monday
A day with the customer service team: listen, watch, note where the time goes.
- Tuesday
Access to the systems, with your IT team. First tests on real requests.
- Wednesday
First agent in suggestion mode: it prepares, the advisors decide.
- Thursday
Review with the advisors. What is wrong gets fixed the same day.
- Friday
Check-in with the sponsor: the week’s measures, decision on next steps.
What you receive
Everything you need not to depend on us.
The code, in your repository
Readable, tested, documented. Your teams can take it over.
The evaluation sets
Real cases approved by your teams, replayed before every change.
The dashboards
Volumes, quality, costs and business indicators, tracked over time.
The operating procedures
What to do when an agent hesitates, gets it wrong or a system goes down.
Engagement
A fixed fee to build. A subscription to run.
Every project is quoted after a first call. You know the price before we start.
Diagnostic
Fixed fee
Agreed before we start, based on the scope.
Build
Fixed fee per scope
A price per case, with its deliverables and success criteria.
Run
Monthly subscription
Operations, maintenance and changes, billed monthly.
On your side
What we need from you.
Few people, but the right ones. We take care of the rest.
- A sponsor who makes decisions and follows the measures
- One or two people from the business available each week
- Help from your IT team with access and security
Questions
About how we work.
Do we have to start with the diagnostic?
It is what we recommend: it avoids building the wrong thing. If you already know what to build and your systems are ready, we can go straight to the build.
Does the engineer work at our offices?
As often as the project needs. The first weeks are largely spent with your teams; after that, the rhythm follows the work.
What do you need from our teams?
A sponsor who makes decisions, one or two people from the business available each week, and help from your IT team for access. We take care of the rest.
What happens if we stop the run?
You keep everything: the code in your repository, the documentation, the evaluation sets and the operating procedures. Your teams or another provider can take over.
Can the agents act on their own?
Only on what you authorize. At first they prepare and a person decides; then you open up autonomy task by task. Sending, publishing or paying always follows the rules you set.
Let’s start with a diagnostic.
A first call to understand your goals and propose a scope.