The Capability That Comes Too Late

When Design Is Treated as an Afterthought, Products Pay the Price

Over the past few years, working as a contractor across product, experience, and AI-enabled initiatives, I’ve found myself noticing the same pattern repeating across organisations. Teams are moving quickly to build new products, particularly those involving AI, driven by clear business goals and increasingly sophisticated technology. Yet despite that momentum, many of these products struggle to land in practice.

What’s striking is not a lack of effort or intent. It’s that the work is often framed too narrowly from the outset. AI initiatives are approached primarily as Technology and Business challenges, with an implicit assumption that if something can be built, deployed, and measured, its value will naturally follow. Design, when it appears at all, is frequently brought in later to support delivery or improve usability, rather than to shape the problem being solved in the first place.

That framing matters more than it might initially seem.

Signals from the field

Looking back across projects and conversations, two signals stand out clearly.

The first is the expanding scope of product roles, particularly those connected to AI. Product owners and managers are increasingly expected to do far more than manage backlogs or prioritise features. They are asked to translate between technical capability and human context, support adoption across diverse stakeholders, and hold decisions together in situations where outcomes cannot be fully predicted in advance. In effect, these roles are absorbing uncertainty that was never explicitly designed for.

The second signal is the renewed pull towards service design. This rarely shows up as a deliberate strategic move. More often, it seems to appear when initiatives start to falter. Products feel fragmented, adoption stalls, and teams struggle to explain how decisions connect end to end. Service design is then brought in to map journeys, diagnose breakdowns, and create a shared view of what already exists.

Taken together, these signals suggest that organisations are sensing a gap. Something is missing in how products and services are being shaped. But the response is reactive. Design is being relied on to retrofit clarity, rather than to help define value and intent from the start. This pattern is reflected more broadly across the industry. Despite the pace of investment and experimentation, only a small proportion of AI initiatives ever translate into sustained real-world value. The technology works, but the product often doesn’t.

The missing capability, not the missing role

At the heart of this pattern is a misunderstanding of what design actually offers.

In many organisations, design is still primarily associated with production. Visuals, interfaces, execution. Important work, but only one part of the picture. What’s often overlooked is design as a thinking capability: the ability to frame problems, surface assumptions, explore trade-offs, and make sense of situations where the right answer is not obvious upfront.

That capability becomes especially important in AI-enabled products and services. These systems influence decisions, shape behaviour, and affect how work gets done in ways that are not always immediately visible. While technology can generate options and insights at scale, people remain responsible for interpreting outputs, weighing consequences, and deciding what action to take.

When design is absent early on, those judgement-heavy questions don’t disappear. They resurface later, embedded in overstretched roles, fragmented services, and products that technically work but feel misaligned with the realities of use.

Why this keeps happening

The reliance on stretched product roles and retrofitted service design is not accidental. It’s a natural outcome of treating design as something that happens after key decisions have already been made.

When the early stages of product development focus primarily on feasibility and commercial logic, questions about relevance, impact, and experience are deferred. By the time they become unavoidable, the product is already in motion. Design is then asked to resolve issues that are structural rather than cosmetic.

This is where much of the frustration comes from. Not because design isn’t valued, but because it’s engaged too late and framed too narrowly to address the underlying problem.

A different starting point

If organisations want AI-enabled products to be more relevant and effective, the shift is not simply to add more designers or more process. It’s to involve design capability at the point where problems are framed and decisions are shaped.

That means treating design as a strategic partner alongside business and technology from the outset. Not to slow things down unnecessarily, but to ensure that what is being built is grounded in a clear understanding of who it is for, what it is meant to change, and how it will actually be experienced in practice.

When design is present early in this way, product roles don’t need to stretch to hold unresolved questions, and service design doesn’t need to be used as a corrective measure. Coherence is built in, rather than patched on later.

The signals are already there. The question is whether organisations are willing to interpret them differently.

What You Could Do Tomorrow

Small shifts in how questions are asked early on often do more to improve outcomes than changes made later in delivery.

Before the next AI-enabled product or feature moves into build, pause to ask a different set of questions:

• What outcomes is this meant to enable, and for whom?

  This grounds the conversation in intent and experience, rather than features or technology.

  
  

• What assumptions are we making about how this will be used, and how confident are we in them?

  This surfaces risk early, without slowing things down or assigning blame.

  
  

• Where does human judgement need to sit in this system, and how is it being supported?

  This ensures that decisions, responsibility, and trust are designed deliberately, not left to chance.

  
  

If design capability exists in your organisation, involve it earlier and more continuously. Not to polish solutions, but to help frame problems, surface assumptions, and support decision-making as complexity emerges.

If it doesn’t, notice where these questions land instead. That observation alone often reveals why products struggle to land, and where design is missing