3 mins read
AI can generate almost anything. Humans still have to decide if it should exist.
As AI lowers the cost of producing things, the real differentiators move upstream.
The rapid adoption of AI across design, development and marketing has fundamentally changed the economics of digital delivery. Tasks that once took days now take hours. Research, content drafting, code scaffolding, testing, reporting and optimisation can all be accelerated through automation. For digital agencies, particularly focusing on those like us working with large technology organisations, this shift is already reshaping how work is delivered. However, speed is not the same as value and output is not the same as impact.
As AI lowers the cost of producing things, the real differentiators move upstream: creativity, leadership, accountability... When average output becomes cheap and abundant, the question is no longer whether something can be produced, but whether it should be. Importantly, whether it will meaningfully move the business forward.
AI doesn’t just automate execution, it makes judgement the scarce skill
This isn’t the first time the digital industry has faced a moment like this. As far back as The Internet of Things (harking back to doing my degree 20 years ago!) the promise was that connected systems would automate data collection, decision-making and operational efficiency at scale. Machines would increasingly talk to machines, reducing human involvement in routine execution. What AI has done is push that same idea much further up the value chain. Instead of simply connecting systems, AI now participates in analysis, synthesis and creation, compressing not just operational cost but creative and cognitive cost as well. Execution is no longer scarce and producing something is easier than it has ever been.
That abundance creates a new challenge: judgement.
AI excels at generating options, patterns and possibilities. What it cannot do is be accountable for intent. It does not understand organisational nuance, brand truth, political context or reputational risk. It does not sit in leadership meetings or navigate the tension between competing stakeholders. It can propose however it cannot decide, nor can it be accountable for the consequences of those decisions. That responsibility and the authority that comes with it remains human, which is precisely why human judgement becomes more valuable as everything else becomes easier.
This distinction matters most in enterprise environments. For multi-million pound B2B organisations, digital programmes are rarely simple marketing exercises. Websites are platforms, products and internal signals all at once. We know they must balance compliance, technology constraints, sales enablement, investor perception and long-term brand equity. AI can accelerate the work, but it cannot resolve the complexity.
As execution accelerates, leadership (not efficiency) defines impact
As AI takes on more of the mechanical aspects of thinking, the human role shifts. Less time is spent producing repetitive artefacts and instead, more time is spent framing the right problems, curating direction and defining what ‘good’ actually looks like in context. The work becomes less about volume and more about judgement, so we see a shift from doing the work and towards directing it. This is where agencies face a defining choice, where AI can be used to simply do the same work faster and cheaper, or it can be used to create space for better thinking, deeper collaboration and stronger outcomes. Efficiency will quickly become a baseline expectation, however leadership will not.
I can see agencies excelling where they can combine aggressive use of AI with clear human ownership. Be that showing clients not just how quickly something was produced, but why it exists in supporting their business and what risks have been consciously addressed along the way. Value has never come from output alone, it comes from intent, coherence and accountability from knowing when to push forward, when to refine and when not to build something at all. This work remains human and in a world increasingly saturated with AI-generated output and it may just be the most valuable work there is.
Share this
