AI work redesign · Belgium & the Netherlands

We redesign how you work with AI, and help you put it into practice.

The human capability behind your AI transformation. Independent. Done with you. Owned by you.

Working in Dutch, English and French.

The problem

AI tools are already in the business. The operating model is not.

Your people use AI every day. The way work is organised hasn't caught up, so the gains stay incremental and nobody is quite sure who owns what. Most companies bolt AI onto workflows that were already broken. We redesign how the work runs around AI, then stay to make it real.

Why now

Anyone can build AI now. Owning it is the hard part.

Generative AI and agents put building within reach of every function. A finance or HR team can stand up an agentic workflow with its own credentials, so the data team can no longer be the sole owner, and ownership has to be redistributed on purpose. Done deliberately, that is an operating model: the business owns the workflow and the value, IT runs the platform, risk owns the guardrails, HR owns skills and roles, leadership owns the ambition and the boundaries. Left to itself, it is shadow AI. If everyone owns it, no one does.

~5%

of companies capture AI value at scale. The difference is rarely the tools.

BCG, 2025
40%+

of agentic AI projects are expected to be cancelled by 2027, most for missing operating-model foundations.

Gartner, 2025
CEO

oversight of AI governance is among the factors most correlated with AI's impact on the bottom line.

McKinsey, 2025
What we do

We redesign the work, build the system, and build the capability to run it.

Redesign

How the work runs

The operating model around AI: which steps agents handle, where humans stay in the loop, who owns each decision, and what it costs.

Put it into practice

The capability to run it

The human capability to direct agents, judge their output, own the decisions, and adopt new ways of working, so the redesign holds.

Foundation

Designed and built, with your team

Platform and guardrails, agents as managed digital workers, and the business context they rely on. Built with your team, vendor-neutral, and handed over.

Measurement runs throughout, with a baseline from day one and value tracked to outcome. See how we work

The method

Four moves per workflow.

Two questions run underneath every move. Where should AI sit in this work: automate it, keep a human in the loop, or keep it human-led? And given an AI output, how hard should you check it before acting? The answers are placed on purpose, step by step, not left to habit.

Move 1

Map

Document the workflow as it runs today. Score every step on cost, time, and risk. Flag the steps that shouldn't exist. Set the baseline.

Move 2

Redesign

Which steps agents handle, where a human stays in the loop, who owns each decision, where the checkpoint sits, and the autonomy level per step.

Move 3

Enable

Build the team's capability to run it, with role playbooks and output checks they keep. Build the agents with them, not for them.

Move 4

Measure

Metrics agreed up front, read before and after. The value is proven, not claimed, and you own the answer.

Governance, made concrete

Every agent is a managed digital worker.

Not a person, not a script: a role with a named owner, permissions, performance metrics, an audit trail, an escalation path, a cost profile, and a kill switch. That turns an abstract governance conversation into a register your leaders can act on, and it answers the question every team eventually asks: who owns this?

Calibrated, not maximal

How far you go is your call.

Not every company wants to take AI all the way, and that is the point. We calibrate the redesign to the ambition you choose and write down what you deliberately won't automate. The result is a stance your board can defend, not a demo.

The capability we build

Four muscles, not soft training.

A redesign only holds if the people running it can. So capability is part of the system we deliver, built inside the workflow, with playbooks your team keeps.

Direct

Brief, orchestrate, and redesign work around agents. Using AI well, not just prompting it.

Judge

Evaluate output, see where it fails, avoid overreliance. Critical thinking lives here: knowing how hard to check before you act.

Own

Decision rights, accountability, and cost. Who owns which agent, and who answers for its output.

Adopt

Change behaviour and ways of working so the redesign holds, rather than dying as a pilot.

What you get

Three ways in, low commitment to high.

Every engagement is fixed in scope and leaves working artifacts your team owns. Prices on request; the first step is designed so one sponsor can say yes.

2–3 weeks

AI Work Redesign Scan

The low-risk door. Your ambition, a current-state and context read, five to ten workflow opportunities scored, the gaps, and a 90-day plan. You leave knowing where to start and what not to automate.

Flagship
4–6 weeks · fixed fee

AI Work Redesign Sprint

One real workflow redesigned end to end: agents placed, humans in the loop, owners named. Built with your team, measured against a baseline, and yours when we leave.

  • For software, product and data teams: the AI SDLC Redesign Sprint, from intake to delivery.
Multi-workflow

AI Work Redesign Programme

The follow-on the Sprint earns. Several workflows plus the operating-model and capability work across a function or the organisation.

What you hold at the end of a Sprint

Six artifacts your team keeps.

1

Ambition statement

One page: how far this team goes with AI, and what it deliberately won't automate.

2

Baseline & opportunity map

The current workflow mapped and scored on cost, time, and risk, with where to start.

3

Redesigned blueprint

The to-be workflow, step by step: AI role, human role, control, decision rights, owner.

4

Agent spec & register

Per agent: purpose, owner, permissions, cost, kill switch, autonomy level, acceptance criteria.

5

Capability & playbooks

Your team trained to run it, with the role playbooks and output checks they keep.

6

Value & adoption plan

The agreed KPIs, baseline to target, and a 90-day tracking sheet with an owner.

You can hold it

Every deliverable is an artifact, not a slide of advice.

You can measure it

A before and an after, in numbers you agree up front.

It stays

When we leave, you own the result and can run the next one without us.

Who it's for

Teams already using AI that never redesigned the work.

First wedge

Software, product & data teams

Already on AI coding and productivity tools, but delivery, review, testing, governance, and ownership were never redesigned around them. The impact is biggest and most measurable here.

The shape

Mid-sized companies and departments of larger groups

Enough complexity to have real workflow pain, some internal tech capability, AI already used informally, and a sponsor who can decide. Belgium and the Netherlands first.

The screen

Three filters, all three

AI is already entering real workflows. Nobody can name who owns the agent, the output, the cost. And one sponsor can start without a procurement marathon.

Also a fit: media and knowledge work, regulated professional workflows, operations-heavy industrials, and sector federations. Not sure? Take the 90-second check.

Why us

Five things the big firms won't say.

1

Workflow-first, not strategy-first

We start with one real workflow, not an enterprise-wide programme. Proof first, then scale.

2

Agent ownership as a signature

Each agent is a managed digital worker: an owner, permissions, a cost profile, a risk profile, a kill switch.

3

Capability as infrastructure

Direct, Judge, Own, Adopt is part of the redesigned system, not soft training bolted on afterwards.

4

Done-with-you build

We build with your team and transfer ownership. No dependency, no lock-in. When we leave, it's yours to run and extend.

5

Vendor-neutral

No model, platform, or tool of our own to resell, so we build on what's best for you and can tell you which demos are just wrappers.

What we are not

Not a model or platform reseller. Not a generalist transformation group. Not a vendor of prompt training. We design how humans and machines work together, build it with you, and hand it over so it stays yours.

Judgment stays sovereign Always a stance Measured in impact Practical over polished At the frontier
Who you'll work with

Senior in the room, not a bench behind a partner.

The person who scopes your engagement is the person who delivers it, end to end: the redesign, the capability building, and the measurement. No hand-off to a junior team after the pitch.

Sensemakers is an independent, Belgium-based firm with a background in data and AI strategy, business analysis, and adoption and change, built in media and telecom. We work in Dutch, English and French.

Where would you start?

In six weeks: one workflow redesigned, measured, and owned by your team.

Tell us about one workflow that carries real cost, time, or risk. We'll show you what redesigning it around AI would look like, and whether you're ready for it.

What to expect: a 30-minute first call
1

We listen

You walk us through the workflow and the pressure behind it. No pitch.

2

We test fit

A quick read on whether a redesign is the right move now, or whether you're too early. We'll say so either way.

3

We propose a start

If it fits: a Scan or a Sprint, with a clear scope and a number you can hold.