The next investment needs fresh evidence

Originally published on LinkedIn on 22 October 2026, in Between the Boxes.

A project can still matter to the organisation while its current delivery approach no longer deserves further investment.

That is an uncomfortable distinction. People have worked hard. Suppliers have committed resources. Leaders have defended the programme. The original problem may remain urgent.

None of that removes the need to examine the next decision on its own merits.

For me, the question is practical: what evidence justifies the next commitment?

An ambition can survive a project

A recent public example comes from UWV, the Dutch employee insurance agency.

In its advice published on 2 February 2026, the Dutch Advisory Council on ICT Assessment, AcICT, recommended a different approach to UWV’s Bemiddelingsservice. It called for smaller steps and evidence of value from the matching functionality before further integration and optimisation.

UWV subsequently announced that it would stop development. Its explanation referred to data quality, technical integration and insufficient confidence that successful implementation was achievable within acceptable time and cost. The underlying ambition of skills-based employment matching remained.

This is a documented public case, not an engagement of mine. It does not establish that every troubled programme should stop, or that AI itself caused the outcome.

The useful distinction is between preserving an ambition and continuing a particular solution.

Test value before extending commitment

Consider an illustrative example.

An organisation is introducing a new customer-onboarding platform. The programme has delivered interfaces and configured workflows, but operations staff still correct incomplete applications manually.

The next request is to integrate more systems.

Before approving it, I would want to understand whether missing integrations are actually the main constraint. The problem might instead be unclear instructions, inconsistent acceptance criteria or information requested too late in the process.

A bounded test could compare the current process with clearer instructions and a simpler manual workflow. A limited version of the proposed platform could then be assessed against that improved alternative.

The service owner would define the intended outcome. Frontline staff would identify exceptions and practical workarounds. Technology colleagues would establish which dependencies must be tested safely.

Measure time to a usable application, correction work, customer effort and service quality. Include the effort required to operate the new approach.

If simpler process changes produce a comparable result, the case for more technology needs revision. If the platform adds demonstrable value, that evidence can support a carefully scoped next step.

The aim is to reduce uncertainty before increasing commitment.

Give continuation a clear boundary

“Continue” should describe a decision with conditions.

For example: continue the bounded test until the agreed review date, with an explicit cost limit and a named decision owner. Specify what would justify expansion and what would require a different approach.

Choose the evidence before seeing the result. Otherwise, it becomes too easy to redefine success around whatever has been delivered.

There are legitimate reasons to continue despite weak early results. Some benefits depend on integration. A poorly designed trial may understate value. Mandatory service or regulatory requirements may limit the available alternatives.

Those are reasons to improve the assessment, not to assume either success or failure.

A useful review therefore includes the strongest credible explanation for disappointing results, alongside evidence that could disprove that explanation.

Stopping needs an owner too

Stopping a project does not automatically release its people, remove its obligations or preserve what is useful.

Someone must own the transition.

That includes maintaining essential services, resolving outstanding commitments, retaining usable knowledge and deciding which components are genuinely worth keeping. Reuse can create value, but only if the receiving team needs the asset and accepts the cost of maintaining it.

People also need a clear explanation of what the organisation has learned and what happens next. Ending an approach should not require pretending that the work or the people involved had no value.

The business objective may continue through a smaller intervention, a different supplier, a redesigned process or a later decision supported by better evidence.

Bring one decision to the next review

For the next programme review, select one upcoming commitment rather than debating the entire transformation at once.

State the outcome it is expected to enable. Identify the uncertainty that matters most. Then establish the smallest credible way to resolve it before the commitment becomes harder to reverse.

Record who will decide, when that decision will be made and what evidence would change it.

Previous expenditure matters for accountability. Future obligations, transition costs and reusable assets matter for the next decision. The decision still needs a defensible view of what happens from here.

If you are facing a continuation, redesign or stop decision, I am happy to discuss one concrete programme and the evidence needed to move it forward.

Milo Izzinosa — Enterprise Transformation & Execution Lead | eXcelleRes

Sources

AcICT — Advice on Bemiddelingsservice, 2 February 2026

UWV — Decision to stop development of Bemiddelingsservice

Sources checked on 22 October 2026. The onboarding example is hypothetical. The recommendations are management analysis, not measured client outcomes.