When a team starts exploring AI, the conversation can quickly become a list of tools. A more useful starting point is the work itself: what question are you trying to answer, and which part of the process could benefit from another way of working?

Treat AI as a method to assess within a brief. The quality of the work still depends on the direction, the material provided and the decisions people make about the result.

Choose a specific task

“Use AI in our marketing” is too broad to guide a meaningful test. “Explore several ways to explain this service to a first-time buyer” gives the team a task it can review.

Other useful starting questions may involve creative directions, content variations or a repetitive step in a workflow. Keep the first test small enough that someone can inspect both the output and the process. Define what the result will be used for before generating it.

A concept exploration and customer-facing material require different levels of review. Make that distinction clear to everyone involved.

Set the boundaries before generating

Bring the audience, intended message, tone and relevant brand guidance into the brief. Identify the source material that can be used and the claims that require evidence. Decide how project information should be handled under the team’s requirements.

For visual work, describe what the image needs to communicate and how it will fit the layout. For writing, explain what a reader should understand or do. A long list of style adjectives is less useful than a clear purpose supported by concrete examples.

Keep approved material separate from explorations so a draft cannot be mistaken for a finished asset.

Review against the job, not the novelty

Generated options can be surprising, but surprise alone is not a reason to use them. Ask whether an option expresses the right idea, matches the brand and fits the context where it will appear.

Review wording for accuracy, relevance and unsupported claims. Inspect visuals for details that could confuse the intended message. Consider whether the people maintaining the work can make the necessary changes later. Where rights or usage conditions matter, include them in the review before release.

The responsible person should be able to explain why an option was selected. “The tool produced it” is not a creative rationale.

Learn from the test

After the work is reviewed, record what helped and what needed correction. Did the process clarify a direction? Did it produce useful variations? Did the review effort outweigh the benefit for this particular task?

Keep the answer specific. A method that helps with early exploration may not fit final production. Give the next test a clear reason, rather than expanding the process simply because a tool is available.

At NEXT, AI belongs within a connected conversation about strategy, design and technology. Start with a useful question, define the boundaries and keep people accountable for the work that reaches the audience.