Innovation: Are We Even Talking About the Same Thing?

That variance showed up clearly in how people talked about AI specifically.

When developers publicly discuss generative AI, it’s often positioned as a tool for exploration, rapid iteration and early-stage thinking, usually accompanied by the reassurance that “nothing ends up in the game.”

Here’s where I’ll offer a personal view, not just a summary of the room: that framing almost inverts the way many artists think about value creation. If there is one part of the process where I would most want talented creatives spending their time, it is exploring ideas, interpreting briefs, challenging assumptions and finding unexpected solutions, often where the greatest creative value is created.

That raises a real question: are we sometimes applying new technologies to the very part of the process that benefits most from human creativity, while overlooking opportunities elsewhere in the pipeline? Personally, I’m more interested in applications that remove friction from project planning, qualification, information management and the operational processes that consume time without benefiting from creative expertise.

That reflection led to what may have been the most important observation from both sessions.

When somebody referred to “AI use,” they might have meant final production assets, early-stage ideation, project planning, organising information, marketing copy, code assistance, localisation workflows, or simply upscaling imagery. Those carry very different implications, levels of risk and expectations of disclosure, yet they’re frequently grouped together under a single label.

I suspect that goes some way towards explaining why these conversations can feel confused. Two people can walk away believing they fundamentally disagree when, in reality, they’re discussing entirely different applications. Equally, people can find themselves apparently in agreement whilst carrying completely different assumptions about where boundaries should sit.

Throughout the discussions, I was struck by how often the challenge seemed to be less about conflicting opinions and more about a lack of shared language.

One observation has stayed with me since XDS: how different this conversation would have been even a few years ago.

Historically, when a client commissioned work, the focus was overwhelmingly on the outcome: did it solve the problem, meet the brief, achieve the desired quality? How the work was produced was generally considered part of the service provider’s expertise, rarely scrutinised in its own right. The pipeline was largely invisible, and provenance was, for the most part, assumed.

What feels different today is that the process itself has become part of the conversation. Alongside what has been created sits a growing interest in how it was created, what tools and workflows were involved, and in some cases whether those decisions can be evidenced. That’s a meaningful shift. Discussions that once focused almost entirely on outcomes are increasingly extending into process.

And it’s not only developers asking this question. Players want to know too. Whether art was hand-crafted, AI-assisted, or something in between has become a live topic in its own right, showing up in reviews, community discussion and, increasingly, platform policy. That pressure doesn’t stop with developers. It flows straight through to the partners who help make the work.

What was particularly interesting, however, was that this change was far from universal. For some participants, provenance, disclosure and governance were becoming central to evaluating trust and risk. For others, they remained relatively low on the list, particularly when weighed against commercial pressures and delivery expectations.

Perhaps the industry is still adjusting to the idea that process has become part of the product. Historically, the question was simply whether the work achieved the desired outcome. Today, there’s an increasing expectation that the journey can be explained as clearly as the destination, though how important different organisations consider that requirement still varies considerably.

In theory, flexibility is a strength. In practice, it can become a burden.

One theme that surfaced repeatedly was the burden that uncertainty places on service providers. Flexibility has become increasingly important. Partners are expected to adapt to different client preferences, different interpretations of policy and different levels of organisational readiness, often before those positions have been clearly articulated.

Many participants described a familiar balancing act: expected to demonstrate innovation whilst avoiding assumptions about where a prospective client might stand. Some organisations are actively exploring new technologies and inviting partners into that journey; others remain cautious; many sit somewhere in between. Until those positions become clear, service providers are left interpreting signals and adjusting their approach to avoid missteps.

During the discussion, I raised a concern about what happens if everyone approaches this challenge behind closed doors. If developers, service providers, legal teams and procurement functions are all independently forming policies but reluctant to show their hand too early, there’s a risk we gradually diverge without realising it, believing we’re moving in roughly the same direction whilst quietly developing very different assumptions about risk, transparency and acceptable practice.

What was encouraging was that some participants described ongoing, open conversations with their partners on these topics, and those discussions appeared to be working well. But those examples almost always existed where strong relationships, and a foundation of trust, were already in place.

That led to a question I kept returning to. If openness depends on trust, how do you build that trust with a new partner when neither side is entirely certain of the other’s position? How much do you disclose? How much do you ask? At what point does transparency build confidence rather than create concern?

Perhaps that’s where the real challenge lies. Existing partnerships have the benefit of history and established communication channels. New partnerships have to navigate the same uncertainty without that foundation, and the more fragmented industry positions become, the harder those first conversations risk becoming.

Several participants also questioned whether we sometimes avoid the most important conversation altogether. If speed, flexibility and cost efficiency are becoming more important buying criteria, should we be more explicit about how those outcomes are achieved, and what trade-offs are involved?

Project delivery has historically been understood through a familiar balance of cost, quality and time: optimise for two, rarely all three. Emerging technologies are beginning to challenge that assumption, or at least create an expectation that it should be challenged. What many service providers seemed to be wrestling with wasn’t whether efficiency should improve, but what sits behind expectations of greater speed and value, and how transparent both sides need to be about the methods used to achieve them.

More than anything, what I took away from both sessions was the sheer variance in how different organisations are approaching change. Some parts of the industry are significantly further along, with established policies and greater confidence discussing innovation openly. Others remain cautious, focused on different challenges, or still working through their positions.

What was remarkably consistent, however, was that assumptions around acceptable risk varied enormously, sometimes between organisations facing almost identical circumstances.

The challenge isn’t a lack of innovation. Nor, in many cases, is it a lack of willingness to engage with it. The challenge is that our language, expectations and assumptions haven’t evolved at the same pace.

Until they do, service providers will keep spending significant energy interpreting signals rather than responding to clearly stated expectations. Innovation was never the difficult part. The difficult part is learning how to have the conversations that innovation forces us to have, particularly when different parts of the industry are moving at different speeds, and nobody’s finished writing the playbook yet.

The tools themselves will continue to evolve. Whether trust evolves alongside them may depend on whether we can first agree on what we’re actually talking about.

 

Find Atomhawk’s service provider profile on XDS Spark and get in touch to continue the conversation.