Why AI use is outpacing transparency in creative work
As AI moves faster than the standards around it, creative work is being reshaped in ways neither clients nor agencies fully see, making clarity and boundaries part of the work itself.
Businesses are already integrating AI into their workflows to save time, generate ideas and move faster. But when it comes to clients, that use isn’t always visible. In some cases, that means work that looks fast on the surface, but still carries hours of decision-making behind it. There’s still no clear standard for how AI use should be explained, priced, or disclosed. Sometimes it’s unclear what needs to be disclosed. Other times, there’s a risk it will be seen as taking shortcuts.
That’s where the pressure builds. Clients expect faster answers, better insights and more personalized work, often for less. A first draft might arrive quicker, but the time doesn’t disappear. It shifts into refining, checking and making sure the work actually works. AI is already part of everyday creative work, but the rules around it haven’t caught up. Envato’s 2026 report shows that gap clearly: 58% of creatives say they’ve used AI in client work without disclosing it, while only 31% always do disclose. At the same time, expectations are split. Around 45% of creatives say clients ask for AI to save time or cost, while 15–19% still want non-AI work. So creatives are working between unclear signals.
What matters now is making the process visible where it actually affects the work. The parts that actually shape the outcome: where a tool was used or had an impact, and where judgment came in. The businesses doing well aren’t the ones using the most AI, but the ones that are clear about how they use it, where it helps and where it doesn’t.
A more transparent way of using AI
When AI use isn’t clear, it stops being a workflow choice and becomes a client risk. In client work, that quickly turns into a question of consent. TOML Collective is building a more open relationship with clients by being clear about how AI is used. The focus shifts from ‘are we using AI?’ to ‘how are we using it, and what does that change?’ That means being open about where AI is used, from early ideas to final campaigns, and what remains human-led. Early exploration can be AI-driven, but the work becomes more controlled as it moves toward delivery. This is supported by workflows that bring AI specialists together with filmmakers, writers and designers, using the technology to expand what’s possible while keeping the craft intact.
As Angie Obregon, senior manager, brand marketing & communications at Siebert Financial, one of TOML Collective’s clients, explains: “AI isn’t a replacement for creativity, it’s a tool that expands it. We work closely with our partners to define where AI adds value and where it should step back. The real impact still comes from human judgment, strategy, and refinement. AI doesn’t diminish the work; it removes limitations, giving us more room to create with intention.”
AI makes it easier to create, but also easier to get things wrong. It lowers the barrier to entry, but raises the stakes for how work is made, especially when it involves sensitive material such as brand strategy, unreleased campaigns or customer data. The tools used can quietly change the level of risk a client takes on. For example, using an open model for early exploration is one thing. Using it on unreleased campaign material or customer data is another. The same tools that speed things up can also flatten ideas or pull too closely from existing work. They make it easier to produce more, but not necessarily better. Which is why how we use them matters just as much as what they can do.
It also raises a simple question: what is the client missing that they would expect to know? Sometimes it’s as simple as not knowing a tool was used at all.
Why AI needs boundaries, not just better outputs
What’s required is a different way of building. Not pushing innovation at any cost, but setting clear boundaries from the start, around tools, data, and outcomes. In practice, this can be simple. AI might be used early on to explore ideas, but not in the final output. Or it means being clear with clients when a tool has shaped the work – not everything, just the parts that matter. For example, some agencies use AI to generate early campaign concepts or moodboards, but final assets are still refined and signed off by the creative team. Or deciding upfront that certain data, references, or client materials don’t enter those systems at all. Tools such as Adobe Firefly or Getty’s models point to a shift toward more controlled environments, where data is licensed, and outputs are easier to trace. What gets delivered is safer to use commercially and easier to stand behind.
AI needs to be treated as both technical and human at the same time. That’s why the risk sits less in ‘replacement’ as a big future scenario and more in the quiet erosion of the tasks where people build judgment. The small decisions, what to keep, what to discard, what feels right. And while ‘collaboration’ sounds generous, without boundaries, it can easily turn into substitution. So the focus shifts to the work itself. Not just what’s made, but the choices behind it, and where AI plays a role. What sets apart the teams getting this right is that they have a real sense of what’s possible. They’ve learned how to use the tools available now, and how to work with them properly. What they’ve done is turn trial and error into something more structured, something they can apply, explain and stand behind.
Right now, adoption is moving faster than understanding. The next phase belongs to those who can do this well. Who can use AI with intention, price it in a way that reflects the work behind it, and be clear about how it’s used. Clients, as well as leadership, need to understand what they’re using, how it works, and what role they play in it. The goal isn’t to slow things down, but to move forward with intent and accelerate responsibly. That means building with guardrails from the start, and updating them as the technology evolves. For that to work, systems need to be clear enough to trust, what they do, where they fall short, and what’s been left out.
Written by Debora Deva
Debbie is a writer, art director, and multidisciplinary creative at TOML Collective. With a background in advertising, she brings fresh perspectives to the journal — aiming to educate, question, and spark new ideas.
Get in touch with debora@tomlcollective.com