How we work
A short method, with a result in every phase
We don't start with a hundred-page AI strategy. We start with one real person's working week and end with agents in production and a team able to maintain them.
The Innomanagers method has three phases: role assessment, pilot agent, and rollout with enablement. Each phase delivers something usable and decides whether to move to the next.
Phase 1
Role assessment
We work with the person in the role, not with an org chart. We record their week task by task: what they do, with which data, how long it takes, what interrupts them and where errors appear.
At the end of the phase you have a map of the role, a shortlist of candidate tasks ranked by recoverable hours and risk, and a concrete recommendation on where to start. That deliverable is yours, whether or not you continue.
- A map of the role's real week, task by task.
- An inventory of the data and systems each task touches.
- Candidate tasks ranked by impact and by risk.
- A definition of what will be measured, against which baseline.
Phase 2
Pilot agent
We build the agents for the prioritised tasks, connected to your data and your tools, and put them to work in the real process with the person in the role in the loop.
The pilot is measured against the baseline defined in the previous phase: time per task, errors caught, rework. If an agent doesn't improve the process, it is adjusted or dropped — not dressed up as a success.
- Agents in production on your systems, not an isolated demo.
- Explicit limits: what runs alone, what needs approval.
- Logging and traceability of every action from day one.
- Weekly measurement against the previous way of working.
Phase 3
Rollout and enablement
We formalise what worked: documentation, controls, data governance and review criteria. We train the team to run, adjust and extend the agents, and we write down how the next one gets built.
From there we extend the model to neighbouring roles, which usually share data and processes with the first one and move much faster.
- Hands-on training for the team that will run the agents.
- Documentation, controls and review criteria in writing.
- A plan to extend to neighbouring roles.
- Full handover: the capability stays in-house.
Frequently asked questions
- How long does each phase take?
- The role assessment takes two to three weeks. The pilot agent takes four to eight weeks until it is in real use. Rollout and enablement depend on team size, usually a few more weeks.
- Do we have to commit to the whole programme up front?
- No. Each phase ends with its own deliverable and a decision to continue or not. Starting with the assessment alone is common.
- What do you need from our side?
- Access to the person in the role, access to the data and systems they use, and someone with the authority to approve the agent's limits. Nothing else.
- Which technology do you use?
- We choose the model and tooling based on the task, the risk and the firm's data-processing requirements, and we document the choice. We don't lock clients into a specific vendor.
Which role should you augment first?
In a 30-minute first call we can identify the role with the most recoverable hours in your firm — and tell you frankly whether starting there makes sense.