A planner can accept a recommendation from a machine. The harder question is whether that planner can defend it when a client asks why.

In Out-of-Home, a recommendation does not stay on a screen. It gets attached to a budget, a campaign and, ultimately, a client conversation about whether the money worked.

OOH planning has moved well beyond choosing a billboard based on traffic counts and available inventory. Audience mobility data, geospatial analytics, inventory intelligence and increasingly sophisticated measurement have given planners a richer picture of where audiences are, what inventory can do and how campaigns can be evaluated.

But more data creates its own problem. Someone still has to connect the information, decide what matters and turn it into a recommendation a client can understand.

That is the human gap Chen Chen, Global Head of AI at billups, has been thinking about as AI becomes more embedded in OOH planning. Her view is less about adding another layer of technology and more about making technology understandable, useful and accountable.

Inside billups, its very own agentic AI platform, audrai™, offers one example of how that thinking is being tested in practice.

From selling inventory to selling confidence

“For a long time, we mainly focused on planning and selling inventory, not so much on the data and measurement side,” Chen said.

That shift changes the planner’s responsibility.

Measurement is no longer simply something that happens after a campaign. It increasingly influences how objectives are defined, how media is selected and how success is discussed with a client.

The planner’s job, then, is not simply to find available inventory. It is to connect audience, location, inventory, objectives and measurement into a coherent recommendation.

This is where Chen believes AI can help not by becoming the answer, but by helping planners work through information more efficiently and with greater transparency.

Chen sees transparency as an important part of how AI should support planners. If a system can show where its information comes from, she believes planners have something they can evaluate rather than simply accept.

The distinction matters. A recommendation that arrives without context asks a planner to trust the system. One that gives the planner something to examine creates room for judgment.

The first sign of trust is skepticism

Trust in AI is rarely created by a successful demo.

One of the clearest signs that planners were engaging seriously with audrai came when the system did not work perfectly.

Chen recalled planners pushing back when the system crashed early on, checking its output line by line and double-checking material before it went into a client deck.

That skepticism was not a problem. It was part of the process.

A polished demo shows what a tool is supposed to do. A failed output reveals what happens when people have to decide whether the tool deserves a place in their actual workflow.

This matters particularly in OOH, where planning involves multiple variables: market, audience, inventory, location, timing and campaign objective. Even with increasingly sophisticated data, someone still has to decide what matters and how it should influence the plan.

Over time, repeated use helped build confidence. Planners could flag responses with thumbs-up, thumbs-down, emojis, or notes, giving the system more information about their preferences and reservations.

“It’s really about learning from our users, understanding their preferences, and even their reservations,” Chen said.

Trust is not installed with a product launch. It develops through use, feedback and the willingness to question the output.

The adoption gap is human

Technology is only one part of the equation. Chen sees a gap between planners who are eager to experiment with AI and those who remain cautious.

“We see a gap,” Chen said. “There’s real worry about job security.”

Those concerns cannot be solved simply by adding more features. They require a clearer conversation about what AI is actually for. That conversation is particularly relevant in media planning, where more sophisticated data can make decisions more precise while also creating more systems, information, and processes for teams to navigate.

The opportunity for AI, then, is not simply to provide more information. It is to reduce the friction between information and action.

For Chen, that does not mean replacing people. It means giving planners more time for the parts of the job that still require judgment, creativity and client understanding.

“It’s never about replacing people with audrai,” she said. “It’s about using the time we save to do work we enjoy, and serve clients better.”

A planner does not embrace a tool simply because it exists. The tool has to make the work better.

Work first, tool second

That idea shapes Chen’s approach to adoption.

“A lot of companies bring in a tool and then want people to use it,” Chen said. “Instead, we’re trying to understand how people already work, find the friction, then find the model that fits.”

For OOH, that means looking at the planning process as a chain: understanding the brief, defining objectives, identifying audiences and markets, evaluating inventory, building a recommendation and determining how the campaign will be evaluated.

AI can potentially support different points in that chain. But the starting point should be the work itself.

Inside billups, audrai is being used as an AI platform to support that work, including applications across planning, market intelligence and creative optimization. The team has also used it across pitches and client work, cutting some tasks from weeks to hours.

The important question is what happens to those recovered hours.

If they simply produce more output, the benefit is operational. If they create more time for strategy, client conversations, critical thinking and better recommendations, the benefit becomes more meaningful.

That is the distinction Chen’s approach points toward: AI should fit the work, rather than forcing the work to fit the technology.

The work behind the intelligence

Caution around AI is not necessarily rejection. It can also reflect responsibility.

OOH is becoming more measurable and data-driven, but richer data raises the standard for the recommendation built from it. Clients increasingly expect media plans to connect activity with measurable objectives. Planners, in turn, have to translate increasingly complex information into something coherent and defensible.

Human responsibility does not disappear when AI enters the workflow. If anything, it becomes more important.

A recommendation attached to a client budget has consequences. The planner has to understand it, explain it and recognize where judgment is still required.

That is why transparency is more than a technical feature. It is part of the working relationship between a tool and the person using it.

An AI system can make mistakes and still become useful if people can question the output, understand what happened and help improve the system. What matters is not blind confidence. It is informed confidence.

Key takeaways

• OOH planning is becoming more data-driven, but planners still have to turn that information into decisions clients can understand and defend.

• Trust in AI is built through transparency and use, including the ability to question, challenge and improve its outputs.

• AI adoption is ultimately a human issue, shaped by concerns around jobs and changing workflows.

• The best place to start is the work itself: identify friction in the planning process, then determine where technology can genuinely help.

• AI should create more room for human judgment, giving planners more time for strategy, creativity and stronger client relationships.

The Performance Age of OOH will bring more data, more measurement and more AI into the planning process. None of that removes the human from the decision.

It raises the value of the human who can look at the information, understand the context and answer the question that still matters most:

“Why this recommendation?”

The answer cannot simply be, “The AI said so.” The planner still has to know why.

And every planning team has different friction points: measurement, data confidence, workflow or simply knowing where AI belongs in OOH. The useful starting point is understanding where the work gets harder than it should, and where technology can genuinely help.

If that’s a question your team is asking, billups can help you find the answer.