Lantern's Pivot Shows GEO Is a Business Model, Not Hype
Lantern spent 2024 as a loyalty-tech startup before relaunching in July 2026 as a generative engine optimization (GEO) platform for e-commerce brands, according to Business Insider. The pivot rides a real trend — AI-driven shopping traffic is up 4,700% year over year and converts 5-8x better than Google organic, even though 60% of it never clicks through. Founder Andrew Lissimore raised a $3.1 million seed round led by Salesforce Ventures and hired ex-Amazon engineers to build the visibility-scoring model behind it. For PMs, the lesson isn't about Lantern specifically — it's that GEO tooling is becoming a required line item, and you need to vet the measurement claims before you buy.
If your product touches e-commerce discovery in any way (DTC brand, marketplace seller, retail media, even B2B catalog search), treat this week's action item as an audit, not a purchase decision. Pull your last 90 days of traffic and separate AI-referred sessions (ChatGPT, Perplexity, Copilot, Gemini) from search-engine sessions, even if that number looks small today. It won't stay small: Lantern's own figures put AI-driven e-commerce traffic up 4,700% year over year, with 75% of new product searches now starting inside an LLM. You don't need to buy a GEO tool this week. You need your baseline before every vendor in this space is pitching you a dashboard. Then stress-test any pitch you do take. Ask the vendor, Lantern included, how they isolate a real visibility improvement from ordinary model drift, since an assistant's answer to the same prompt can change week over week with no code change on your end. If they can't explain that in plain language, they're selling you astrology with better charts. To be fair, this is a young category and nobody has published a clean measurement standard yet. That's not disqualifying, it's just where the industry is. But "young category" is exactly when you get the best pricing and the most vendor attention. Move now, audit hard, and don't let a visibility score replace your P&L as the thing you actually manage to.
Lantern pivoted from a 2024 loyalty-tech startup to a generative engine optimization platform, launching its Agentic Commerce Performance product on July 1, 2026, after a $3.1 million Salesforce Ventures-led seed round.
Lantern spent its first year as a loyalty-tech company. Almost nobody remembers that now, because on July 1, 2026, the startup relaunched as an "Agentic Commerce Performance" platform — software that predicts how ChatGPT, Claude, Gemini, and Perplexity describe your products, then rewrites your catalog so the answer favors you. According to Business Insider, the company just pivoted its entire go-to-market from loyalty tooling into generative engine optimization, or GEO.
My verdict: the pivot is the smartest move this company has made, the category is real and growing fast, and you should still be suspicious of anyone selling you a "visibility score" they can't back with a receipt.
The Loyalty Startup That Found Religion
Here's the part of the story the press release skips. Andrew Lissimore, who built Headphones.com into a real business using classic SEO, co-founded Lantern in 2024 with Kyle Peatt and Dominic McPhee, two former Shopify designers who helped build the Polaris design system, to sell loyalty tooling to e-commerce brands, per TechCrunch's earlier reporting on the company. Loyalty is a perfectly fine business.
It is not a business that gets you a $3.1 million Salesforce Ventures-led seed round retold as a growth story eighteen months later. Generative engine optimization is.
It's not that Lissimore spotted a market nobody else saw. Plenty of marketers have watched AI assistants swallow search traffic over the past couple of years. It's that his original company needed a narrative its loyalty product couldn't generate on its own, and GEO handed him one, fully formed, with a growth chart and a plausible villain.
The Numbers That Justify the Rebrand
Give Lantern this much: the underlying trend isn't manufactured. Traffic to e-commerce sites from AI-driven sources is up 4,700% year over year, and roughly 75% of new product searches now happen inside a large language model instead of a search engine, according to figures the company reported to Business Insider. That traffic converts 5 to 8 times higher than Google organic, and 60% of those interactions end without a single click back to the brand's site.
Sit with that pair of numbers for a second. Five to eight times the conversion rate, on 60% of the volume you can't even see land. That's not a rounding error in your attribution model.
That's most of your highest-intent traffic disappearing into a black box before it ever touches your analytics stack.
Lantern's answer is to build instrumentation for that black box: hire ex-Amazon engineers, train a proprietary model that runs a brand's prompts daily across the major assistants, score presence, position, and sentiment, and surface a "Predicted Visibility" number alongside the specific citation gaps where a competitor gets recommended and you don't. Pricing starts at $99 a month for smaller brands, with custom deals for bigger catalogs. Fixes get applied to product pages and catalog structure, with a human still approving the change before it ships.
To Be Fair: The Question Nobody's Answered
To be fair to Lissimore: this isn't a rebrand slapped on the old loyalty dashboard. Standing up a model that ingests four different assistants' outputs daily and turns them into a per-SKU visibility score is real engineering, and hiring people who've built recommendation systems at Amazon scale is a credible signal that someone here understands the problem, not just the marketing opportunity around it.
But every gatekeeper shift produces the same swarm, and this one is no different. When Google's algorithm became the front door to the internet in the late 1990s, a wave of directory-submission consultants rebranded overnight as "SEO agencies," most of them years before anyone could prove a stable ranking correlation. A handful of those agencies became Moz and Ahrefs.
Most became a line item some CMO cut two budget cycles later. GEO is that same swarm, one generation on, and analysts covering the category are already flagging the same weak spot: without auditable, cross-model attribution, "we got you cited by ChatGPT" is a claim, not a receipt — because model outputs drift week to week, and nobody in the category has published a measurement standard that isolates a vendor's fix from that noise.
What This Means for Your Stack
If you run product, growth, or marketing for anything that sells online, GEO tooling is headed for the same fate SEO tooling had by the 2010s: not optional, not exotic, just another line in the martech budget. The 60%-no-click number is the one to internalize first, because it kills last-click attribution as a way to judge whether any of this spend is working. You need a brand-lift or citation-tracking layer that doesn't depend on a visit landing in your analytics tool, full stop.
Before you sign anything (Lantern or a competitor), make the vendor show you their measurement methodology, not their dashboard. Ask how they separate a real improvement in how your product gets described from ordinary week-to-week drift in the underlying model. If the answer is a confidence score with no error bars, you're buying a report card graded by the same student who wrote the test.
The deeper question isn't whether Lantern's technology works. It's whether "visibility inside a system you don't control and can't fully audit" is a metric you can defend in a board meeting — or whether GEO turns into pay-per-click's estranged cousin: a gatekeeper tax with no refund policy and no receipt.
Get this in your inbox. AI Rundown Daily delivers original briefings every morning — free. Subscribe →
Frequently Asked Questions
Not entirely — the company started as a loyalty-tech startup in 2024, not an SEO tool, so this is a full pivot rather than a relabeling. Founder Andrew Lissimore built his earlier company, Headphones.com, through traditional SEO, which gives him direct experience with how search-driven discovery breaks down as a business model. That said, the skepticism is fair to apply to the category broadly: plenty of "GEO" vendors are legacy SEO or content-marketing shops that bolted on a ChatGPT dashboard. Lantern's move to hire ex-Amazon engineers and train a proprietary visibility-prediction model is a stronger signal than most, but it's still a pivot story, and pivots deserve scrutiny until they show revenue traction, not just funding headlines.
Lantern's tooling starts at $99 a month for smaller e-commerce brands, with custom pricing for enterprise catalogs, according to reporting on the company. Implementation centers on connecting your product catalog so the platform can run daily prompts across ChatGPT, Claude, Gemini, and Perplexity, score how your products are described, and propose fixes to product pages and catalog structure. Changes to your storefront still require human approval before they go live, so the rollout is closer to a review workflow than a fully automated pipeline. Expect the real cost to be internal time spent validating whether the suggested changes move any business metric, not just the subscription fee.
Measurement. Analysts covering the category have pointed out that GEO vendors, broadly, lack an auditable way to prove that a visibility improvement came from their fix rather than ordinary drift in how the underlying LLM answers the same prompt week to week. That makes it hard to connect any 'citation' or 'visibility score' back to actual revenue. A secondary risk is dependency: you're optimizing for interfaces controlled entirely by OpenAI, Google, Anthropic, and Perplexity, none of whom owe you a stable ranking algorithm or advance notice of a model update. Budget for this as an experiment with a defined review point, not a locked-in annual contract.
