Human + AI: Where Control Still Matters in Programmatic

Your AI just shifted half your budget overnight. Can you explain why?

If the honest answer is “not really,” you’re in good company. The real programmatic debate in today’s time isn’t whether to use AI. It’s where humans should stay firmly in charge. 

AI is everywhere. Autonomy isn’t.

IAB Europe’s 2026 study, reported by Digiday, found that 86% of respondents use AI for marketing. Among those asked about specific workflows, 59% use it for programmatic optimisation, while agentic buying and selling sits lower at 38%. Of 47 respondents asked about autonomy, 36 reported either no agentic system in daily use or one where people still direct or plan alongside it. 

The takeaway: the industry trusts AI to optimise, but it’s still deciding how far to let it transact. The sample is small (50 respondents), so read it as a signal, not a verdict.

The risk is AI you can’t see inside

McKinsey’s June 2026 research found that more than 90% of surveyed US advertisers use AI to plan media, set budgets, optimise targeting, and generate creative. Yet 42% cite reliance on “black box optimisation” as a key risk.

The governance gap is just as telling. In the IAB Europe study, 78% of organisations have an AI governance owner, but only 48% have marketing-specific AI guidelines. Among those offering training, just 24% cover how to check autonomous actions and when to intervene.

We’ve handed teams powerful tools, then skipped the lesson on when to take the wheel.

Where human control still matters

🔹 Goals and guardrails: AI optimises what you ask for. Humans define “good”: KPIs, brand-safety limits, floor prices, frequency caps.
🔹 Approval gates: Set spend limits, and require sign-off for big budget shifts or new supply sources.
🔹 Supply quality: AI can chase low CPMs into weak inventory. Review where impressions actually come from.
🔹 Explainability: If a system can’t show why it made a call, treat it as a suggestion, not a decision.
🔹 Brand context: Nuance, culture, and tone still need human eyes.

A simple test before you hand over more control

Ask four questions:

🔹Can we see what AI changed, and why?

🔹Can we pause or reverse it quickly?

🔹Who is accountable if spend goes wrong?

🔹Does our team know when to step in?

Any “no” means a human checkpoint stays in the loop.

What the balance looks like in practice

Media planning is a good example. In a tool like Voise Agentic Planner, AI does the heavy lifting: real-time inventory discovery, audience matching, budget split recommendations, and KPI forecasts for reach, CTR, viewability, and CPM. VoiseTech says this can cut the time from brief to IO by up to 50%.

But the planner still decides which audiences matter, whether the budget split makes sense, and which plan goes to IO. Machines do the legwork. People make the call.

The bottom line

Winners won’t be the most automated teams. They’ll be the ones who know what to automate, what to review, and when to step in.

That’s the thinking behind how we build at VoiseTech: AI-driven infrastructure across planning, buying, and monetisation, with transparency and control at every stage. Whether you’re an agency, advertiser, or publisher, we’re always open to working with teams that want automation they can trust. 

How ready is your infrastructure to capture higher yield through transparent programmatic channels? Let’s talk.