Salem Handoura and Romain Darrieus have launched Laier, an agentic platform built for the daily work of campaign optimization. Both come from French agency group Heroiks, where Handoura was managing director. Their starting point is a frustration most digital buyers share: ad platforms surface more data every year, and little of it turns into a decision on a live campaign.
"Data visualization tools made campaigns easier to track without necessarily making the decisions better," said Handoura. Time spent optimizing campaigns fell from 22% of the working week in 2016 to 11% in 2026, according to a survey of media buyers Laier ran.
Reporting and dashboard building went the other way, from 12% to 18%. Automation freed up hours, and those hours went into watching campaigns and data feeds instead of strategy.
"Media buyers spend half as much time optimizing as they did ten years ago. Laier exists to reverse that."
The founders built the company in that gap: more data than anyone has time to use. Their bet is that the daily optimization work teams keep cutting is where the next performance gains sit.
"Advertisers have made the big calls – shifting budgets to digital, journey studies, attribution models, measurement. That work is behind them. With budgets under pressure, the extra performance won't come from another strategic decision. It will come from optimizing every live campaign as well as possible, every day," said Handoura.
Moves is the feature built for that. A move is an optimization action the platform puts in front of the user.
The basics are there – excluding queries that don't convert, pausing a campaign or an audience. Laier goes after the rest, the optimizations teams know they should run and never get to.
Shifting budget from a Meta campaign to a Google or Pinterest campaign performing better on the same goal. Catching creative wear-out before acquisition cost climbs. Spotting which audiences, creatives, dayparts and channel combinations deliver, and carrying that into the next campaign.
"The big change in how media gets bought is tempo," said Handoura. "Campaigns are shorter and more fragmented across platforms. A two-week push on four channels leaves no room for a mid-flight review."
In a two-week window, speed is the lever. "Reallocating a budget on day three rather than day 10 means going after performance where nobody had time to go," Handoura said.
Recommendations land on the home screen. "Each one has to come with a hypothesis and the data behind it," said Romain Darrieus, Laier's other co-founder.
The user accepts, rejects, or adds context – a spend commitment, a commercial priority, a creative that stays live whatever its numbers.
Approved moves are pushed straight into the ad platform. Rights sit at workspace level, one workspace per client, so an agency can limit approval to certain profiles and give the advertiser read-only access.
Every decision lands in a ledger, a timestamped history. Laier comes back to it an average of 14 days later to check whether the change did what it was supposed to, and to feed later recommendations. The trail doubles as the agency's evidence with its client.
The company says it has coded around 120 analysis routines: reading rules specific to Performance Max, creative fatigue detection, pacing control, benchmarking an acquisition cost against its recent baseline.
An LLM comes in after that, to read the context, write the analysis and handle the back-and-forth. Deterministic calculations frame what it can say, rather than the model running the analysis end to end.
Analysis, reporting and campaign launch, all through chat
Around Moves, Laier pulls in functions usually spread across separate tools.
"Explore" queries the data in plain language: where a budget will land, what drove a drop in ROAS, or how worn a creative is.
Charts and answers worth keeping get pinned to a Canvas, a dashboard that refreshes when the period changes. An exploration can also become a move and go through the same approval and tracking loop.
"Reporting" builds the weekly or monthly review from a template the user defines: refreshed data, comparison with the previous period, generated commentary, and a recap of the optimizations run.
"Launch" creates a campaign from a plain-language instruction. The user sets one goal per campaign at setup: a click or a conversion at the bottom of the funnel, reach or completion at the top. "Branding campaigns aren't steered on CPA, but they are steered. We're moving fast on that part, especially with DSPs coming into scope," said Handoura.
Laier reads the creative assets it is given and works out which formats and environments they fit. It then proposes a media plan based on what earlier campaigns showed.
The agency's and the advertiser's own documents feed the same engine – decks, media plans, post-campaign reviews, naming conventions, internal rules, creative – so recommendations match each client's context and working methods.
A RAG architecture handles that retrieval, letting the model consult the document base before it answers.
The tool compares campaigns that share the same goal, whatever the platform: an acquisition campaign on Meta against the same one on Pinterest, a TikTok acquisition push against a Google Demand Gen. Google Analytics is integrated and can serve as the common yardstick.
Laier connects to the ad platforms mainly through APIs. The founders said they prefer them to MCP servers for now, because APIs offer "more functional depth and more predictable read and write permissions."
Around 10 connectors are live, including Meta, Google and Amazon. The target is some 40 platforms by the end of October. DV360 and Amazon DSP arrive by the end of September, which brings open web programmatic into scope.
Media agencies are the first target, since they roll the tool out across several clients, share the analysis and document what they optimized. "Part of the goal is to make the work between an agency and its clients easier," said Darrieus. Advertiser acquisition teams come second, since they can tie spend to a conversion quickly.
About 10 pilot partners before the September launch
Laier said it had about 10 partners at the time of our interview, direct advertisers and agencies, and close to 15 tests since development started.
Media agencies Haiku and Oconnection are the early adopters, both deploying the platform across several clients. Those rollouts test the product at portfolio scale: workspace-level rights, moves approved by the teams, reviews shared with the advertiser. Direct advertisers round out the group.
Laier ruled out a commission on media spend from the start, the standard arrangement in ad tech. "Getting paid on budget creates a conflict with optimization. You can't be paid to spend and stay credible about spending better," said Handoura.
Pricing is a flat all-inclusive subscription, set by the size of the organization rather than by ad spend. The rate card sits on the site, with several tiers.
Every feature is included at every tier, with no cap on the number of markets and a seven-day free trial. Above 10 workspaces, agencies get custom pricing that starts under €2,000 a month.
Laier is aiming at the whole campaign cycle, from launch to post-campaign review, with daily optimization in between. "The platforms automated the bid. What's left to automate is everything around it: the arbitrage between channels, reading the signals, executing the decisions and remembering what worked. That's what Laier does, under human control," said Handoura.
The platform goes on sale in France in September, with the rest of Europe and the U.S. to follow. The team has also started talks on a funding round.
The real test is already running. Laier has to show its moves are worth more than smart alerts, and that teams stop opening the interfaces they monitor today. The pilots haven't proved that yet. On productivity the gain is already banked: reporting, launch and the execution of decisions all run themselves.
