The founding idea arrived as an absence. Chris Andrew was no longer opening so many websites. When he wanted to compare products or understand a subject, he had begun expecting ChatGPT to do the wandering and return with an answer. Twenty blue links suddenly felt like an errand. For a product executive trained to watch behavior, the missing browser tabs were more interesting than the chatbot itself.
Andrew called a few chief marketing officers he knew from his years in enterprise software. Their dashboards showed only a small amount of traffic arriving from AI tools, but those visitors behaved differently. They arrived informed and ready to act. The same brands also appeared differently across ChatGPT, Gemini and Perplexity because each answer drew on a changing mix of company pages, reviews, third-party articles and competitors. A familiar marketing promise had become unstable: companies could publish a neat story about themselves and still meet customers through a machine's untidy synthesis.
“I realized I was visiting fewer websites. I was expecting an answer from ChatGPT instead of 20 links from Google.”Chris Andrew, describing the observation behind Scrunch
In fall 2023, Andrew and Robert MacCloy started Scrunch. They built software to show companies what AI systems said about them, where those claims came from and what information was missing. The product launched publicly in November 2024. By June 2026, Scrunch had raised $26 million, worked with more than 500 brands and agencies, and been acquired by Sitecore. The business continued as a standalone platform, now attached to a much larger content-management machine.
A long apprenticeship for a fast market
The quick version of the Scrunch story begins with ChatGPT. The useful version begins much earlier, in Bow, New Hampshire, and at Grove City College. Andrew grew up homeschooled in the 1990s, an experience he later turned into a joke with decent timing: he graduated first in his class, which sounds more imposing until the class size is disclosed. At Grove City, he studied business economics, captained varsity cross-country and track teams, and spent a 2007 term studying economics at Lancaster University.
Distance running supplies an almost suspiciously tidy metaphor for startups, so it is best used sparingly. Still, the record says something. Andrew chose events where progress is repetitive, measured and occasionally unpleasant. His professional education developed the same way. He worked in product management at Intuit, then left to join Clara Shih and the young Hearsay Systems team. Hearsay helped large financial institutions use social networks while keeping the necessary rules, reviews and controls. Andrew was its first employee and moved across product, customer success, sales engineering, international expansion and, eventually, the chief product officer role.
MacCloy was there too, leading technology. The two would work together for nearly 15 years. At Hearsay, they watched a new consumer habit force itself into conservative institutions. Facebook and LinkedIn were not difficult because executives had never heard of them. They were difficult because personal, fast-moving networks collided with slow approval systems and regulated communication. Hearsay built the bridge: central rules with room for individual employees to speak like people.
That history became Scrunch's hidden advantage. AI search looked like a fresh category, but the organizational problem felt familiar. Consumers had adopted a new way to discover information. Marketing teams could see the movement and still lacked the controls, measurements and vocabulary to respond. Andrew and MacCloy already knew how enterprise buyers discuss a new channel after the novelty wears off. They knew that a persuasive demo eventually meets procurement, security, compliance and a vice president asking for proof.
The browser becomes a broker
Andrew's central claim is less about search rankings than delegation. Browsing means looking around without knowing exactly where the answer lives. An AI agent can perform that uncertainty on a person's behalf, gathering pages, comparing options and returning a recommendation. The person may visit a company website only at the end, after a short list has already formed inside a chat window.
Scrunch began by monitoring this invisible contest. It runs prompts across AI platforms, records mentions and citations, and shows where a brand disappears or is described inaccurately. Monitoring, however, identifies the bruise without moving the furniture. The company added an Agent Experience Platform that gives crawlers a simplified, machine-readable version of a site's information while leaving the human-facing pages intact. The design team can keep the graceful animation. The crawler gets clean hierarchy, context and facts.
The practical lesson is refreshingly unromantic. Publish explicit pricing, comparisons, definitions and evidence. Make important information accessible without relying on a person to click a clever interface. Track which sources appear in answers. Correct stale pages. Then run the prompts again. AI-search strategy may sound like a new branch of sorcery, but much of the work resembles good information architecture with a stricter reader.
Watch what you have stopped doing. A vanished habit can reveal a market earlier than a trend report, especially when its second-order effects create a new buyer problem.
People, judgment and a patient pivot
Scrunch was not the first version of the company Andrew and MacCloy set out to build. Investor Patrick Salyer backed the pair before the final product had emerged, after trying to recruit Andrew several times nearly two decades earlier. During roughly a year and a half of searching, Salyer and Andrew held biweekly conversations around a useful pressure: the team had found something, but was it large enough? Andrew later remembered the permission inside the challenge: “We invested in a team. Go.”
That history complicates the tidy founder myth in a healthy way. Scrunch was neither a lightning strike nor a straight road. It was a credible team with enterprise scar tissue, enough patient capital to reconsider its direction and a sharply timed observation when consumer behavior moved. The founders' friendship mattered because a pivot asks people to discard work without discarding trust.
Andrew speaks about company-building in similarly human terms. In a 2026 conversation about leadership, he emphasized learning agility, empathy, authenticity and enthusiasm. His shortest hiring instruction was “Hire for judgment.” When information becomes abundant and technology spreads quickly, he argues, the quality of decisions and the people making them become more important. Culture, in this view, lives in repeated behavior rather than framed vocabulary.
“Technology is less differentiated than people.”Andrew on what remains scarce as tools become widely available
There is a domestic version of the same preference for craft. Andrew is a certified Neapolitan pizza chef, as well as an enthusiast of trails, jazz, food and wine. Neapolitan pizza is governed by constraints: a narrow ingredient list, fierce heat and very little baking time. It rewards preparation and exposes clutter. The comparison to software should not be kneaded beyond recognition, but Scrunch shares the instinct to remove what its audience cannot use.
One podcast supplied a livelier artifact from Andrew's earlier years. He recalled a wild car ride with Gary Vaynerchuk around the launch of Wine Library TV and pitching him a cooking show. The proposal did not become Andrew's career. It does reveal that the certified pizza chapter and the internet-marketing chapter have been trying to share a table for some time.
The second audience
Sitecore's acquisition in June 2026 placed Scrunch beside the systems that companies already use to manage large libraries of content. The fit follows Andrew's argument. Seeing how a brand appears in AI answers is useful; connecting that observation to the workflow where teams can correct and publish information is more useful. Scrunch brings the view from the answer engine. Sitecore brings the content estate behind it.
Andrew remains CEO of Scrunch, now a Sitecore company. He has described the next phase as a closed loop: observe what agents say and do, understand the gaps, and deliver the information they need. The company has introduced tools including Signals and Explorer while maintaining that its standalone product and customer support will continue.
His longer ambition extends beyond making brands easier to cite. Andrew imagines two parallel surfaces of the internet: one arranged for human attention, another structured for agents that retrieve, reason and eventually transact. Scrunch wants to help companies serve both. The person still matters, especially when a decision is consequential. But a machine may conduct the research, assemble the choices and decide which names make the first cut.
The charming irony is that this machine-facing thesis came from close attention to one human. Andrew noticed what he, personally, had stopped doing. He tested the observation with people he trusted, then built for the organizations that would inherit the consequence. The next important website visitor may be software. Finding it required the old-fashioned habit of watching a person change his mind.