Digital product personalization startup Aampe Inc. said today it has closed on an $18 million Series A funding round that brings its total amount raised to date to $27.3 million. Theory Ventures and Z47 were the only investors named in the round.

Aampe in the creator of an agentic artificial intelligence infrastructure that’s designed to make digital products more personalized and boost users engagement. The basic way it works is that a unique AI agent is deployed to follow each user that continuously learns how that person interacts with the product. Through this monitoring process, it slowly learns their changing preferences, so it can decide what to show each user, when to show it, or whether or not to show anything at all.

In this way, Aampe’s AI agents can help to drive more optimized interactions for users, delivering personalization for products that serve millions of individual users.

Aampe says its AI personalization agents can serve all manner of digital products, ranging from sports and fitness applications to fintech and entertainment services, and it believes they are far more effective than conventional approaches to personalizing digital products. The agents can individualize various aspects of the application experience, ranging from marketing and messaging to the interface layout and feature discovery, so it can evolve to suit each user’s preferences over time.

Traditionally, developers have always had to manually create different rules and segments to determine what each product users might see at different moments. This approach has been relatively unchanged in over a decade, and requires a big investment in teams of workers who can manually orchestrate a product’s message or surface to best serve individual user’s interests.

As a result of this, most consumer applications today look almost identical to everyone who opens them, with personalization limited to a few narrow recommendation feeds, says Aampe co-founder and Chief Executive Paul Meinshausen (pictured center, alongside co-founders Schaun Wheeler and Sami Abboud).

“We’ve designed and developed infrastructure that enables every aspect of an application to adapt to each user’s context and preferences, continuously,” he said. “Our mission is to fundamentally improve the way users experience digital products.”

Instead of using generative AI or traditional machine learning algorithms, Aampe relies on a subset of AI called reinforcement learning to deliver continuous, parallelized experimentation for each user. The agents can learn and adapt in real-time, managing the attention of each user so they can make more complex choices in real-time in terms of the material and content they will see. In addition to monitoring users’ behavior, they can also tap into existing data platforms, marketing delivery and product analytics tools to further enhance and individualize user’s experiences.

Aampe reckons this approach has already made a huge splash, saying that it has deployed more than 100 million agents across hundreds of consumer applications in four continents, including food delivery apps. Those agents help those applications to make a staggering 150-200 billion decisions each week that determine how specific individuals will interact with those apps.

Alexander Beresford, chief marketing officer at Taxfix GmbH, said his company is already using Aampe’s agents to personalize user experiences on its popular tax accounting application. He said it has become essential for the company to respond to each individual user’s needs.

“Customers expect brands to know what they want and respond instantly, and so the future of engagement in owned media lies in AI systems that learn from each customer’s behavior so they can adapt automatically to deliver personalized experiences,” he says. “Unlike older systems that follow rigid rules, these agents evolve with the customer, keeping every interaction relevant without extra effort from the business.”

Of course, Aampe understands that some users might be uncomfortable with the thought of being continuously monitored. But it does its best to reassure those users, saying it adopts a privacy-centric approach that involves keeping zero personally identifiable information and anonymizing behavioral patterns.

Aampe said its plan going forward is to double its employee base, enabling it to focus on helping new customers integrate its agentic infrastructure and workflows into their applications.

Theory Ventures Partner Andy Triedman said Aampe’s AI agents have the advantage of being able to scale decision-making to a level that’s impossible for any human or human-led team. “Aampe allows customer engagement teams to craft experiences for diverse user bases, versus just one or two flows targeted at the typical person,” he said. “This new type of infrastructure will be transformational for companies looking to provide personalization driven by data.”

Photo: Aampe