Unusual raises $3.6M to shape how AI models talk about brands – News

Unusual, a startup that focuses on changing how artificial intelligence models talk about brands, today announced it has raised $3.6 million in early-stage funding.

Investors in the round included BoxGroup, Long Journey Ventures, Y Combinator, Instacart co-founder Max Mullen and Phosphor Capital.

Founded in 2024 by Chief Executive Will Jack and President Keller Maloney, Unusual helps customers gain transparency into how AI models and services “talk” about their companies. Jack, a second-time founder, has a deep background in AI research dating back to 2014 and hails from the Massachusetts Institute of Technology and Space Exploration Technologies Corp., better known by its shortened name SpaceX. Maloney has a background working with AI models since 2019 with companies such as Gatsby Inc. and Eight Partners LLC.

“These models are not like search engines — the devil’s really in the details,” Jack told News in an interview. He argued that brands now have to manage how they’re characterized across long, consultative chat sessions — not just whether they show up for a keyword.

Unusual focuses on “brand alignment” and visibility for AI models. By working with brand owners and understanding the inner workings of AI models, the company can assist companies with building their online presence so that they are represented accurately.

Jack said the key is that modern models increasingly show their work: “They reveal their reasoning and the research that they do every time they respond.”

That’s because of the proliferation of reasoning models that “think out loud” as they process queries. This includes how the model approaches a specific type of question, the information it gathered and the sources it used.

Unusual uses that to run large-scale brand attribute probes, inspect what sources models rely on, and translate misperceptions into content, proof points and third-party validation strategies. This vast body of research knowledge can be translated into actionable content changes and external sources the brand could influence, such as case studies and media coverage.

The current market has multiple startups that assist companies with AI brand alignment; however, many of them tout themselves with approaches that mimic search engine optimization. Both answer engine optimization and generative engine optimization — emerging new strategies for bringing content into AI-driven search results — focus on delivering concise, direct answers that can be picked up readily by Google LLC’s AI Overviews, OpenAI Group PBC’s ChatGPT and Perplexity AI Inc.

The company goes beyond this with its AI model interrogation. Where AEO/GEO tends to interpret success and results based on whether a company or brand gets mentioned high up in results, Unusual examines the multi-turn portrayal of companies to make certain it matches how a company or service wants to represent themselves.

“Other folks have come in treating AI models like the next generation search engine, and we’re treating them a bit more like human influencers,” Maloney said.

The shifting landscape of customer perception

A survey from McKinsey & Co. this year revealed that 50% of Google searches already have AI summaries, a figure expected to rise to 75% by 2028. Customers also continue to flock to AI models for research and discovery for new products.

Jack said brands have reported a significant shift in organic search traffic, which he characterized as a “hair on fire” problem. This trend has meaningfully affected companies with sophisticated or niche products where buyers want consultive guidance.

Because of the appearance of humanlike conversation, people treat AI models as a de facto source of truth, the same way they would treat another person. This ongoing fundamental shift in perception by the public at large, and the rapid adoption of AI models into search and everyday life, means that brands cannot ignore their influence on awareness.

Unusual works with multiple customers, primarily brands, but also public relations and marketing firms. Customers include AI-driven document parsing company Reducto Inc. and budgeting app Monarch Money Inc.

In the case of Reducto, Unusual discovered that AI models perceived the company as high-quality but “growth-stage” and not enterprise-ready. This meant that decision makers could be quietly steered away by AI-driven answers, leading to poorer marketing opportunities and a less accurate representation in the industry.

Jack said Unusual believes the future requires is bringing the best possible information about companies to AI models so that answers properly reflect brands and services. As AI models become a primary source of information, this will affect every part of the industry from discovery to customer support.

“You want to ensure that you know no matter what question is asked about you, no matter how detailed it is, no matter what nuance or concern one of your customers has, the model is equipped to represent you appropriately,” Jack said.

Image: News/Microsoft Designer