Freddie Turner is Managing Director, Chalice AI EMEA and a monthly columnist for NDA.
Advertising was supposed to fund the open web. Programmatic was supposed to make that fair. Somewhere along the way, we lost sight of that.
The jargon took over, and metrics like CPMs and CTRs became the success metrics, even though they rarely reflected what advertisers were actually trying to achieve.
Budgets grew, but confidence didn’t always follow. When budgets are scrutinised, programmatic is often the first place questions are asked. Not because it lacked scale, but because it struggled to explain its impact in a way businesses could rely on.
We’re at a fork in the road and AI – done right – might actually fix this.
A Shift Already Underway
Think about why your IT team blocks ChatGPT. Same data, same risk, different context. Across enterprise technology, we’re seeing a clear pattern: consumer AI tools are kept out of workplaces because they absorb sensitive data. In response, enterprise-grade AI is emerging that keeps data private, controlled, and commercially usable.
Advertising is having its own version of that reckoning. As AI moves closer to media decisioning, advertisers are asking where their data goes, how models learn, and whether the intelligence they pay for actually works in their favour. The market is slowly but clearly moving toward systems that keep control with the advertiser rather than handing it over to platforms.
This isn’t a sudden change. It’s a steady one. And that’s actually encouraging. The best market shifts don’t happen overnight – they happen when the logic becomes undeniable.
The Fork in the Road
When AI is applied well, it brings clarity. It helps identify which environments earn attention, which impressions create incremental value, and which decisions influence business outcomes. It reconnects planning and performance in ways the industry has wanted for a long time.
When it isn’t, it’s just another black box.
The difference increasingly shows up in results.
Cutting Through the Noise
There is no shortage of AI claims in the market. Almost every product now sounds intelligent. But performance doesn’t come from language. It comes from how systems are built and how they behave in practice.
Here’s how to cut through it. Ask these questions:
1. Data transparency – Is your data taken in, or do you retain control? 2. Learning scope –Is your data potentially training your competitors models, or is it silo-ed to only benefit you?
3. Explainability –Could you explain your AI’s optimisation rationale if you had to? Or is a black box on autopilot?
4. Cumulative learning – Does it improve over time in a way that remains proprietary? 5. Deployment flexibility – Can it operate wherever decisions are made, or only within one stack?
These questions don’t just protect against poor choices. They point toward AI that is more likely to perform, because it is grounded in real signals and aligned with real objectives.
Where Programmatic Is Heading
Programmatic advertising is already changing, most notably in how intelligence is applied. The direction is away from platform-owned optimisation and toward AI that is auditable, portable, and aligned with the advertiser’s commercial goals rather than platform incentives.
This is exactly why Chalice exists. Not because we predicted the future, but because we listened to what advertisers kept saying they needed: intelligence they could see, control they could exercise, and value they could prove. A private AI layer that works for you, not for the platform selling you inventory. Trained only on your data, designed to score every impression against your outcomes, and able to operate across platforms without taking possession of that data.
It’s actually quite simple: better decisions require better control.
As with other enterprise technology shifts, the advantage will accrue to those who align early with where the market is moving, rather than waiting for the transition to be complete.
