Programmatic Advertising: Will AI Save the Open Web?

By Max Legrand · GreyFox Ads · A French perspective on programmatic advertising

Will artificial intelligence save the Open Web? Let's get straight to the point: No. 🙂‍↔️ At least, not on its own.

The plumbing metaphor: pipes, plumbers and water 🔧💧

Here's my favourite way of looking at the AI revolution in programmatic advertising. AI could replace the pipes. It could replace some of the plumbers. But it won't magically improve the quality of the water flowing through them.

That distinction matters because discussions about AI often focus on making media buying faster, easier or cheaper. All useful improvements, of course. But advertisers ultimately buy an opportunity to reach an audience with a message that actually works.

1. Replace the pipes: can AdCP reshape programmatic infrastructure?

The AdCP protocol points towards a world in which an AI interface can help plan media and book advertising inventory. The ambition is attractive: simpler workflows and potentially fewer intermediaries than today's complex OpenRTB supply paths. Players such as Scope3 and PubMatic are exploring this direction.

But let's not confuse a promising protocol with a fully operational replacement for the entire ecosystem (the scope of supported transactions and integrations varies).

OpenRTB does more than move a booking request. The programmatic ecosystem also supports inventory discovery and auctions, audience activation, delivery controls, measurement, billing and interoperability across multiple partners. Recreating all of those functions at scale will take time.

AI may make the pipes more efficient. That is a meaningful advance — but a more elegant pipe does not make an ineffective advertising placement suddenly effective.

2. Replace some plumbers: AI agents and the future of media trading 👨‍🔧🤖

Media planning, campaign set-up, reporting and optimisation all involve processes that can increasingly be assisted by AI — provided the systems have reliable data, access to the right tools and appropriate controls.

On the buy side, companies such as Amazon, Google (DV360), The Trade Desk and Concord.ad are part of this evolving landscape. On the sell side, players including Equativ and Supply Finder are exploring automation opportunities.

Does this mean the end of media traders and yield managers? I expect their roles to change significantly. Routine execution may shrink, while strategy, supervision, troubleshooting and commercial judgement become more important (the speed and extent of workforce changes remain uncertain).

AI can help decide how to buy a placement. It cannot guarantee that the placement was worth buying in the first place. That's why the human understanding of supply quality and SPO remains essential.

3. The real problem: AI won't improve the quality of the water 🚫💧

Advertisers choose media environments because of two fundamentals: how effectively the format delivers their message, and how effectively they can reach the right audience.

And where are advertisers directing significant investment? Towards environments with a clear and compelling proposition for users and brands.

Why social video, CTV and sponsored formats attract advertisers

These environments aren't flawless, and attention or engagement will vary by campaign and creative. Still, their strengths are easy to understand. Compare that with much of the traditional Open Web: sound-off video players outside the main content, decades-old IAB banners and fragmented first-party audiences that are difficult to activate in the Open Auction.

That is the uncomfortable competitive reality. Better automation alone does not address it.

What should French Open Web publishers actually change?

Here in France, publishers have valuable editorial environments, trusted audiences and considerable first-party signals. The opportunity is to turn those assets into a stronger advertising product, rather than simply automate the sale of existing inventory.

First, rethink formats. Make video genuinely visible and engaging, particularly on mobile. We've already explored how French publishers could win back video advertising budgets and why fullscreen formats deserve more attention.

Second, simplify audience activation. First-party data must become easier for agencies to use across buying environments, while respecting applicable privacy rules. This includes exploring interoperable identifiers and 1.5-party data approaches.

Third, clean up supply. Reduce unnecessary reselling, questionable placements and opaque fees. Our work on MFA inventory and the Open Web's attention problem illustrates why media quality cannot be an afterthought.

AI can accelerate these efforts — but publishers, broadcasters, ad tech companies and buyers still need to make the product decisions.

Conclusion: what will save the Open Web isn't AI. It's the quality of its water.

AI will probably make programmatic advertising easier to operate. It may reduce friction, streamline workflows and change the daily work of traders and yield managers. Great.

But if the industry keeps selling the same underwhelming formats with the same targeting limitations, faster pipes won't rescue it.

Want to save the Open Web? Improve what advertisers are actually buying. The water, not just the plumbing. 😉

Based on the original LinkedIn post by Max Legrand. The outlook for AI, AdCP and future job roles reflects the author's analysis, not a guaranteed forecast.