The Prompting Company Wants to Make ChatGPT Talk About Your Product

The Macro: Search Has a New Front Door

I spent the last year watching smart people argue about whether AI search would actually replace Google. That debate is settled. Not because Google lost. Google is still enormous. But because a meaningful chunk of product research queries now start with “Hey ChatGPT, what’s the best tool for X?” instead of a search bar. The behavior shift is real and it is accelerating.

This creates a problem that nobody in marketing has a good answer for yet. Traditional SEO is a known discipline. You optimize pages, build backlinks, create content clusters, and wait. There are hundreds of agencies that do this competently. But when someone asks ChatGPT for a recommendation, the answer comes from the model’s training data and whatever retrieval system is running underneath. There is no meta description to optimize. There is no keyword density to target. The ranking factors are opaque and different from anything the SEO industry has spent twenty years learning.

Some companies have tried to brute-force this by flooding the internet with branded content and hoping the models pick it up during training. That works sometimes, poorly. Others have ignored it entirely, betting that traditional search still matters more. Both approaches feel like they belong to 2024.

The gap in the market is clear. There is no established toolset for influencing how AI models perceive and recommend products. Surfer SEO, Semrush, and Ahrefs are built for Google. They do not tell you whether ChatGPT mentions your product when someone asks about your category. Nobody has owned this space yet, and the companies that figure it out first will have a real advantage because the window for establishing best practices in a new channel is always short.

The Micro: Former Founders Who Already Sold Companies, Now Selling Visibility

The Prompting Company was founded by Kevin Chandra, Michelle Marcelline, and Albert Purnama. All three are based in San Francisco and came through Y Combinator’s Summer 2025 batch. Their resumes are worth noting. Kevin and the team previously built Typedream, which was acquired by beehiiv, and Cotter, which was acquired by Stytch. Two exits before starting this company. That is not common for a team working on what is essentially a new category of marketing tooling.

Their approach breaks into three steps. First, they identify the questions users are asking AI models about your product category and measure whether your product currently gets mentioned. Second, they create and distribute content designed to increase the likelihood of AI model mentions. Third, they build landing pages optimized for the traffic that comes from AI-generated citations.

I find the measurement piece most interesting. Knowing that ChatGPT does or does not mention your product when asked “what’s the best project management tool” is genuinely valuable intelligence that almost no company tracks right now. The distribution and content creation pieces are more familiar territory, closer to what a content marketing agency does, but the targeting is different because the audience is a model, not a human.

The team is small, four people total, which makes sense at this stage. The real question is whether this becomes a product or stays a service. Right now it feels like a consulting engagement with software attached. That can work and generate revenue, but the scale dynamics are very different from pure SaaS.

The Verdict

I think The Prompting Company is early to a real category. LLM optimization (or whatever we end up calling it) is going to be a line item in every marketing budget within two years. The question is whether it looks more like SEO software or more like a PR agency. If it is the former, there is a massive business here. If it is the latter, it is a good boutique but not a venture-scale outcome.

In 30 days I want to see case studies with measurable before-and-after data on AI model mention rates. That is the proof point that matters. Sixty days, I want to know whether the content distribution methodology is repeatable or whether every client engagement is a custom project. Ninety days, the question is market timing. Perplexity, Claude, and Gemini all handle product queries differently from ChatGPT. A platform that works across all the major models is exponentially more valuable than one that only optimizes for OpenAI. The team has the pedigree to build something significant. Whether the market develops as fast as they need it to is the variable they cannot control.