Ask an AI assistant for a plumber and it will name one. The Y Combinator startup Cheers is building the machinery behind that answer - and it starts with the person standing in your kitchen.
Type "best HVAC company near me" into ChatGPT and it will give you a name. Not a page of blue links to sift through - one confident recommendation, maybe three. Somewhere in that answer a decision was made about who deserves your money, and most of the businesses in your zip code had no idea the vote was even happening. Cheers is a San Francisco startup that noticed the vote, and decided to run for office on behalf of the local plumber.
Founded in 2024 and part of Y Combinator's Summer batch, Cheers calls what it does Generative Engine Optimization, or GEO. The pitch is simple enough to fit on a business card: search is moving from Google's ten links to a single AI answer, and the businesses that get named are the ones with the deepest, freshest, most trustworthy pile of reviews. Cheers helps local service companies build that pile - and, in a twist most reputation software skips, it traces every review back to the exact employee who earned it.
Strip away the acronym and Cheers is a loop with three moves. First it audits: it checks where a business is recommended across ChatGPT, Google's AI answers, Perplexity and Gemini, then shows the operator who is getting named instead of them and why. Second it fixes the plumbing that AI crawlers read - schema markup, citation errors, location pages, the machine-readable files that tell an assistant what a business is and where it works. Third, and this is the part competitors tend to leave out, it puts a tool in the hand of the frontline worker.
That last tool is the piece people remember. A technician finishes a job, the customer is standing there pleased, and instead of hoping they'll remember to leave a review three days later, the tech pulls out an NFC tap card. One tap on the customer's phone opens a Google review. The same card can collect a referral or a lead. The friction that kills most review programs - the gap between the good feeling and the follow-through - collapses to a single gesture at the moment it matters.
Cheers sells to local and multi-location service businesses - the trades and home services where a recommendation is worth thousands of dollars and reputation travels by word of mouth. Think HVAC, plumbing, roofing, solar and smart home. The company points to customers including Vivint Smart Home and Sunrun, and reports a footprint of more than 900 locations and 15,000 frontline employees across its clients.
The Vivint number is the one that gets repeated: 70,000-plus five-star reviews in five months. What made it possible wasn't a clever campaign, it was operationalizing the frontline across roughly 400 locations - every install crew turned into a review-collecting node, every review stamped with a name.
Two problems, really, that turned out to be the same one. The first is visibility: as customers ask AI assistants instead of scrolling Google, businesses that spent a decade tuning their websites for search suddenly can't tell whether the machine even knows they exist. The second is the frontline data gap. Managers rarely know which employee actually drives the reviews, referrals and repeat business that feed reputation. Praise and blame get spread evenly because nobody can measure it.
Cheers closes both by making the review the connective tissue. Because each review is attributed to a specific person and place, a manager can finally see which tech, which branch, which behavior produces results - and coach, reward and scale it. The same reviews that answer "who is my best employee" are the ones that teach ChatGPT to answer "who should I hire."
Reputation software isn't new. Podium, Birdeye and Weave have sold review-collection tools to local businesses for years, and Cheers' own founders come partly from that world. The difference Cheers is betting on runs in two directions. Looking forward, it treats AI visibility as the goal rather than Google stars alone - auditing how assistants recommend businesses and optimizing for them directly. Looking down, to the frontline, it insists on per-employee attribution instead of treating reviews as a faceless pool.
| Capability | Traditional review tools | Cheers |
|---|---|---|
| Collect & respond to reviews | Yes | Yes |
| NFC tap capture at point of service | Rare | Core |
| Per-employee attribution | No | Core |
| AI-visibility auditing (ChatGPT, Perplexity) | No | Yes |
| Done-for-you GEO optimization | No | Yes |
Comparison reflects publicly described positioning, not a feature-by-feature audit of every competitor.
Cheers is small - three people at last count - and the founding pair is an odd, useful pairing. Dylan Allen-Arnegård, the CEO, spent time at two earlier reputation-software companies that also came through Y Combinator, Podium and Weave, plus a stint at Microsoft Azure. He knows the category from the inside. Amadeus Peterson, the CTO, arrived from a different planet entirely: he worked on missile systems at Northrop Grumman before turning to startups. One founder understands why reviews matter to a plumber; the other builds systems that have to work the first time.
There's a small tell that the team believes its own thesis. Cheers publishes an llms.txt file on its own website - a machine-readable summary meant for AI crawlers. It practices the AI-readable web presence it sells.
The model is B2B SaaS with a services wrapper: a recurring subscription that bundles the software - visibility tracking, dashboards, attribution - with done-for-you optimization and the physical tap tools. Early on, the founder described monthly recurring revenue climbing from around $1,400 to roughly $5,000 in a two-month stretch, the kind of small, real number that early-stage companies live and die by. Backing so far comes through Y Combinator's S24 batch.
The larger question is whether "GEO" becomes a category or a footnote. If AI assistants really do become the front door for local hiring decisions, then someone will build the layer that helps businesses show up there, and the frontline - the tech at the door, the roofer on the roof - is a plausible place to source the trust that AI rewards. Cheers is a three-person company making that argument early. Whether it owns the frontline of AI search or simply proves the idea for a bigger player, the loop it's drawn is worth understanding.