The Agency Doesn't Need to Know How to Build Everything. It Needs to Know What's Possible.

I have spent a significant part of my career working within agencies, often in the space between what has been imagined and what can actually be delivered. It is a useful place to be. You hear the ambition before it has been reduced to a scope of work. You also see what happens when an idea meets budgets, deadlines, existing systems and the people who will eventually have to use it.

Something I have noticed over the years is that the limits of a project are often established long before anyone begins building it. They emerge in the first conversations, when we decide which possibilities are worth considering. We tend to imagine solutions through the capabilities we already have. A design team reaches for design. A development team reaches for software. A campaign team reaches for a campaign. None of those instincts is inherently wrong. Each comes from experience. But the brief does not necessarily belong to the discipline that first receives it.

The real opportunity begins when we can examine the problem without being confined by the tools we are accustomed to using.

I don't think an agency needs to know how to build everything. That would be an expensive and probably impossible ambition. But I do think it needs to understand what could be built, why it might be useful and what would be required to make it real. There is a distinction between owning every capability and being able to recognise the right capability at the right moment.

The idea before the deliverable

A client might ask for a careers website. That sounds reasonably well defined until we ask what the site is intended to change. Is the objective to make the employer more recognisable? To help candidates understand where they might belong? To reduce friction in the application journey? To connect fragmented recruitment systems? Or to create an experience that changes as the organisation changes?

Those are related problems, but they lead to very different products.

In my work on careers platforms for organisations such as easyJet, the interesting decisions have sat across brand, content, user experience, technical architecture and the systems candidates eventually encounter when applying. A beautiful interface only takes us so far if the content cannot be maintained, the journeys are confusing or the platform is disconnected from the way recruitment actually works.

The question I like to ask early is simple: what would we do if we weren't already committed to this particular output?

It does not mean disregarding the brief. It means understanding the conditions that produced it. Sometimes the original solution is exactly right. Sometimes the smallest change to the question reveals a much stronger direction.

Bring technology into the conversation while the idea is still alive

Technology is often introduced as a constraint. A concept has been approved, screens have been designed and somebody is asked whether it can be built. At that stage the most inventive possibilities may already have been removed. The technical conversation becomes a negotiation about feasibility rather than an exploration of opportunity.

I prefer technology to be present while the idea is forming.

That does not require every creative to become a developer or every developer to become a strategist. It requires enough shared understanding for the disciplines to influence each other. A technical observation might make an idea more ambitious. A user insight might simplify a complicated system. A commercial constraint might lead to a more inventive interaction than an unlimited budget would have produced.

I sometimes think of this as a reasonable extreme: pushing an idea far enough that something genuinely interesting becomes possible while remaining honest about the conditions of delivery. There is very little value in a technically impressive solution that nobody can maintain, or a safe solution that never meaningfully addresses the problem.

The ability to move between those perspectives has shaped my own career. I began in creative and digital design, moved deeper into development and architecture, and increasingly found myself working across product direction, client conversations and delivery. I didn't follow that path because I wanted to collect disciplines. I followed the questions that needed answering.

Capacity is different from capability

Another thing I've learned is that the strongest team is not always the largest permanent team.

Agencies need continuity. They need people who know the clients, understand the standards and can make decisions without rebuilding the context each time. But they also need access to capabilities that fluctuate with the work. A particular programme might need technical discovery for a month, hands-on delivery for three, and occasional senior oversight after launch. Hiring around the peak of that requirement can create more structure than the problem deserves.

This is where long-term consulting and fractional, embedded relationships make sense to me. They sit somewhere between the distance of traditional outsourcing and the permanence of a full-time appointment. The relationship carries memory, while the involvement can change with the need.

I continue to work this way with Tonic, a UK employer-brand agency, managing and delivering work across several client projects, alongside relationships with other independent businesses. The value is rarely one isolated task. It is being able to move from a client question to a practical decision, support the team that implements it and remain close enough to understand what happens afterwards.

Of course, fractional arrangements are not automatically effective. If the person joining the project is kept away from strategy, excluded from important conversations or engaged only after decisions are made, the organisation gets an additional pair of hands but misses much of the value. Embedded should mean having enough context, trust and responsibility to contribute judgement, not simply attending meetings.

AI makes judgement more visible

AI has made the distance between imagining something and producing a first version remarkably short. We can explore concepts, write code, produce assets and test assumptions at a pace that would have been difficult to imagine a few years ago. I use these tools extensively. They have made me faster, but the more interesting change is that they allow me to investigate more possibilities before committing to one.

There is a temptation to interpret this as proof that specialist expertise matters less. I suspect the opposite is true in important ways.

When producing a prototype becomes easier, choosing what to prototype becomes more consequential. When several implementations can be generated quickly, architecture, accessibility, maintainability and judgement become differentiators. The cost of making something has fallen, but the cost of making the wrong thing has not disappeared. It may even become easier to accumulate the wrong things at speed.

An agency that understands these conditions can offer its clients more without pretending that every solution needs to be developed internally. It can combine its strengths with the right partners, test possibilities early and make better commitments.

The work continues after launch

I have always found the moment after launch particularly revealing. The presentation is over. The platform is now being used. Content teams need to update it, candidates encounter unexpected friction, stakeholders want new features and the business begins to learn what the product actually does.

A launch is an important milestone, but a product only becomes useful through its continued interaction with reality. Good relationships allow that learning to flow back into the next decision. They make maintenance, optimisation and product direction part of the same conversation that started with the creative idea.

That is the kind of relationship I value most: one where I can understand the thinking behind a project, help shape the possibilities and take responsibility for making them work.

The strongest agency, in my view, is not the one with every conceivable discipline in-house. It is the one whose imagination is not restricted by its current organisational chart. It knows its strengths, understands its gaps and can bring the right minds together around the problem.

The agency doesn't have to build everything. But it should be able to see further than what it already knows how to build.