Batch Acquires Moonfish AI to Accelerate Its Agentic CRM Roadmap

Today we're announcing our first acquisition in 11 years. The deal is set to accelerate our Agentic CRM offering, kick off our European M&A strategy and position Batch as the sovereign alternative to legacy CRM systems.

If you run marketing at a serious B2C brand right now, you have a problem you can't say out loud at your own offsite.

You don't know how big your team should be in three years.

You don't know which parts of your job will still belong to humans.

You don't know whether to bet on AI tools that flatter your existing process, or on platforms that will eventually replace large pieces of it.

Anyone who tells you they have a clean answer is selling you something. Including, often, your software vendors.

So let us tell you what we actually believe, and what we just did about it.

The news, in one paragraph

Today, Batch is acquiring Moonfish AI, a Paris-based predictive AI company we've been partnering with for the better part of two years. It's our first acquisition in 11 years of building Batch. 

The Moonfish AI team and their 15 production-grade machine learning models are now part of Batch AI, our integrated suite for Agentic CRM and AI Decisioning, running on top of the Batch Data Platform.

The deal closed in February and the technology shipped to our 350 enterprise customers in March.

That's the press release version.

Now here's why we did it.

Two futures for customer engagement. We're building for both.

Every CMO we talk to is being pulled between two stories about where customer engagement is heading.

Story one is the augmented marketer.

The human stays in charge. AI handles the grunt work, scales the ideas, surfaces the patterns hiding in years of customer data. The marketer keeps the brand, the taste, the call.

In a world where AI-generated content is becoming wallpaper, human authenticity becomes the moat.

The teams that win are the ones that move faster and more creatively than ever.

Story two is the autonomous marketer.

The human steps out of the loop. AI runs orchestration on its own. Humans set goals and guardrails, and they manage agents the way a portfolio manager manages strategies.

The job changes for good.

The honest answer is that nobody yet knows which of these futures will dominate, and the truthful answer is that they're not mutually exclusive. 

A luxury house won't hand its tone of voice to an autonomous agent the way a mass-market retailer will. 

A loyalty program won't run win-back campaigns the way a brand-image campaign runs. 

Sensitivity to brand equity will vary sharply between marge established brands and small pure players

Industry, segment, and brand sensibility will dictate the answer, and the answer will be different across teams inside the same company.

So here's our bet. The platform that wins this decade is the one that's natively built for both modes. Copilot for the marketer who stays at the wheel. Autopilot for the team that wants to delegate entirely. 

Same data. Same infrastructure. Same brand voice. Two ways to operate it.

That's the bet behind the Moonfish AI acquisition. And it's the bet behind everything else we're shipping this year.

Why Moonfish, specifically

Building production-grade predictive machine learning for marketing at scale is one of the harder problems in our industry.

Models have to score hundreds of millions of customer profiles, in real time, with stable performance across seasons, geographies, and product cycles, while staying explainable enough that a CMO can sign off on what it does.

Most teams that try to build this in-house spend years and end up with something brittle. 

Most vendors that sell it ship a black box that works in the demo and disappoints in production. 

There's a narrow middle path, and it requires a rare combination: experienced data scientists, a long time horizon, and real customer data to train on at scale.

Moonfish AI had all three. The company spent years inside Claravista, a data marketing consultancy itself recently absorbed by Jakala, a global Data and AI consulting leader with 3,500 people worldwide. 

That's not the usual venture-backed origin story. It's better. It means the algorithms were stress-tested over years with brands like L'Occitane, Aroma-Zone, Oscaro, Conforma, Printemps, who had no patience for lab demos and every incentive to demand results.

The introduction came from a mutual customer in early 2024 who wanted to plug Moonfish's propensity scores into their email campaigns. We started a partnership. We watched the numbers come in. They were honest. The kind you can show a board:

• Up to +150% incremental revenue on A/B-tested campaigns running with versus without Moonfish scores

• Up to +50% lift in post-campaign customer satisfaction

• Time-to-market on new segments effectively eliminated: marketers stop touching segmentation, the algorithms generate the scores, the platform wires them into the customer profile automatically

In January 2025, we put Barthélémy Lancelot, the Moonfish AI founder running the project, on stage at our annual Batch Keynote. The room was set up for fifty people. More than 120 customers showed up and stood along the walls to listen. That's the moment a partnership turns into something else.

By mid-2025, the partnership was hitting a ceiling. To unlock its full potential, the technology needed to live inside the platform, not bolted on. Jakala was looking to spin out the IP from its AI lab toward partners who could scale it. Our customers were asking for tighter integration. 

The math was simple: own the predictive layer, double down on Batch AI, execute our first M&A.

What this means for the platform

This acquisition is the moment we're refreshing the architecture and naming of Batch AI, our integrated suite for Agentic CRM and AI Decisioning. 

Four building blocks, one platform.

Batch AI Assist

The marketer's copilot. Generates and tightens CRM content in seconds. Speeds up Journey building. Audits campaigns. Surfaces what to do next. Includes Home AI, a conversational agent that reads across the Batch Data Platform and gives you straight answers about your customers.

Batch AI Predict

Predictive intelligence, native to the platform. Fifteen production ML models out of the box: purchase propensity, churn risk, engagement score, best send time, best channel, reactivation likelihood. Plug them straight into segments. No data scientists. No MLOps pipelines to maintain. This is where Moonfish lives.

Batch AI Decide

The autonomous decisioning layer. Batch AI Decisioning calls the shot in real time inside existing Journeys, channel by channel, message by message, with reinforcement learning under the hood. Batch Autonomous Operator runs orchestration around the clock and reports back transparently on every move. You set the goals and the guardrails. The AI does the rest. This is where most of our internal lab is focused right now.

Batch Data Platform

The foundation. Customer profiles, product and inventory data, behavioral signals captured by our SDKs, AI-enriched scores, all unified, all native to leading cloud data warehouses: BigQuery, Snowflake, Databricks, ClickHouse. No serious AI initiative survives without that layer. We've been building it for eleven years.

What this means for you

If you're a Batch customer today: the Moonfish AI algorithms are live on the platform. If you were already using Batch AI Predict (formerly Customer AI), there's nothing to migrate. There's just a much bigger toolkit available. 

→ Your Customer Success Manager can walk you through it.

If you're evaluating Batch: you're now looking at one of the most complete propositions in Data, CRM, and AI on the market. A natively omnichannel CEP across Email, SMS, Mobile and Web Push, In-App, and Conversational. A Data Platform processing 800 million unique visitors a month, wired natively into the major cloud warehouses. And a proprietary AI suite with eleven years of compounding behind it.

→ Request a demo

And if you've followed our story since 2015: partners, customers, friends, the bet hasn't changed. A modern, competitive, sovereign European customer engagement platform is possible. It can stand toe-to-toe with the U.S. cloud giants. We're more convinced of that today than we were on day one.

→ Meet the future of Agentic CRM

One last thing

I want to acknowledge something most CEOs skip in posts like this.

We don't know exactly how the marketing job will look in five years. Nobody does. We're making bets every quarter, and some of them will be wrong. What we do know is that the platforms that survive this transition will be the ones that take the change seriously, build for multiple futures rather than betting on one, and treat their customers as partners through the uncertainty rather than test subjects.

That's the standard we're holding ourselves to. The Moonfish AI acquisition is one move in a longer plan. There will be more. We'll tell you about them the same way: clearly, with the numbers, and with a point of view.

Welcome to the team, Moonfish AI. And thank you, as always, to our customers and partners for trusting us with the work.