Gushwork raises $9M betting on AI search as a source of customers

For decades, digital marketing has revolved around a single axis: Google's search results. Entire industries grew up around the skill of landing in the top ten results, and SEO and contextual advertising budgets ran into the billions. But what if that familiar order is already coming apart at the seams? Gushwork, a startup that has just raised $9 million in a seed round led by SIG and Lightspeed, says it sees the first outlines of a new reality, and it is building its business on the idea that AI search tools like ChatGPT are becoming a full-fledged channel for acquiring customers.

Gushwork's idea sounds both simple and provocative. The company noticed that more and more users turn to language models for recommendations rather than to traditional search engines. Instead of typing a query into Google and scrolling through links, people ask ChatGPT, Perplexity or Gemini a question and get a ready answer with specific recommendations. For businesses this raises a fundamentally new question: how do you get an AI assistant to mention your product when a user asks about a solution to their problem? Gushwork reports early but quite tangible results: part of its customer base came specifically through such AI recommendations.

To grasp the scale of the shift, it helps to recall the context. Over the past two years, AI search has gone from experimental toy to everyday tool for hundreds of millions of people. According to various analyst firms, the share of search queries that start in a language model interface rather than a classic search engine keeps growing steadily. OpenAI built web search into ChatGPT, Google responded with AI Overviews in its results, and Perplexity is built entirely as an AI-first search engine. All of this creates a new ecosystem in which link-ranking algorithms give way to a very different logic: the logic of answer generation.

This is where the main intrigue lies. Classic SEO relies on understandable, if constantly changing, rules: keywords, backlinks, page load speed, content structure. Optimizing for AI search is territory where there are almost no rules yet. Language models form answers from training data, dialogue context and, increasingly, current information from the internet. How exactly a model decides which product to recommend is a question that even the developers of these models cannot answer transparently. Gushwork, apparently, is trying to find patterns in this black box and turn them into a repeatable strategy for businesses.

The participation of Lightspeed and SIG in the round is a signal worth taking seriously. Both firms are known for early bets on transformational trends, and their willingness to put $9 million into a seed-stage company suggests investors see AI lead generation not as a niche story but as a potentially large market. If AI search tools really do take a significant share of commercial queries from Google, it would mean a reallocation of advertising budgets measured in tens of billions of dollars a year.

Skepticism is warranted, though. Early results are not yet a lasting trend. It is still unclear how stable AI recommendations are, and whether they can be influenced systematically without manipulation that will sooner or later provoke a response from model developers. OpenAI and Google are unlikely to watch calmly as third parties learn to "hack" their recommendation algorithms. There are also serious transparency questions: if an AI assistant recommends a product, the user should know whether the recommendation is organic or paid. Regulators, already concerned about native advertising in traditional search, will almost certainly turn their attention to this area too.

Still, Gushwork has tapped into the moment. Business always goes where the customers are, and customers increasingly start their path to purchase with a conversation with an AI. The startup that first learns to work with this channel systematically will gain an enormous advantage. Nine million dollars is a modest sum by industry standards, but enough to test the hypothesis and, if it holds, raise the next round on very different terms. We are watching a new market being born, and for now it looks like the Wild West: no maps, no rules, and huge opportunities for those willing to go in first.