An autonomous website for lean marketing teams. A marketer plugs in their brand, ad accounts, and CRM; Flint builds and continuously updates brand-faithful landing pages — and now, with our agents layer, watches ad campaigns and ships the pages those campaigns need. The first agent is the Google Ads agent: it closes the loop between ad spend and the landing page (generate → host → iterate), which single-surface tools and agencies can't do natively. (A common mix-up worth clearing early: Flint is not an AI search or SEO company — the agents exist to help marketers build and ship more landing pages and ads.)
The data advantage: because customers connect their ad accounts, we see both sides of the loop — the inputs (what pages we build and ship) and the outputs (conversions and revenue). So Flint doesn't just generate pages; it runs experiments, attributes conversion lifts to the specific changes that caused them, and measurably improves a customer's marketing performance over time. Almost every other tool touches only one side of that loop and can never prove its impact.
Product demo — Get Started with Flint
Product demo — Get Started with Flint
Growth, with product enablement — not heavyweight implementation. The bet is a self-serve + product-enablement motion like Unbounce or Leadpages, not the SAP-style "forward-deployed engineer" model where you run four-to-six-week integrations for each customer. The most a customer should need is a ~30-minute enablement call with a solution engineer to deepen usage. The goal is to make the acquisition-to-activation loop automated enough that revenue isn't capped by GTM headcount — the way Cursor scaled.
That said, we are layering on sales for the accounts that need a little bit more sales assist, but even so we see the majority of these requiring standard customer implementation and not FDE.
The golden ICP is venture-backed B2B SaaS with horizontal products — many SKUs, industries, and personas to address (think tools that have to sell to dozens of use cases at once), run by lean marketing teams of 0–2 people scaling paid programs. This includes health tech and PLG productivity companies. Growth agencies running paid for many clients are a second, high-leverage vector — they need new landing pages per campaign per week, and Flint ships at the cadence the ad spend demands. The buyer is usually growth/marketing.
The closest incumbent is Unbounce, Leadpages, and Webflow — drag-and-drop landing-page and website builders for marketing teams who can't ship site changes fast enough. Flint tends to win those deals head-to-head. The broader market is split: AI CMO tools sell point solutions to SMBs with no platform piece; AI SEO tools tell you what to write but don't ship, host, or maintain the page; CMS-only players like Webflow are bolting agents onto a data model that wasn't built for them. Flint owns the full loop — CMS + context graph + agents — inside one system. There's also a large untapped outbound opportunity: sizable companies sign up on their own today with zero acquisition effort, and nobody's been working that demand yet.
Yes — it's gotten bigger. Early on the focus was human-in-the-loop prompting: a marketer asks, Flint generates a page. As the models got better, the strategy shifted from generating pages on request to building a system and a data moat around the customer — figuring out, on their behalf, what pages they should build next from their ads, SEO, and competitive data. Two consequences: (1) it's now critical that customers connect their ad accounts, because that proprietary feedback loop is what compounds over time; and (2) we're far less attached to being the one that generates the code — over time, customers can bring, edit, or export their own — and far more focused on owning the "marketing harness" that decides what to build and ship. Same underlying mission (help lean marketing teams win), much bigger surface area.
A fair risk, and one we actively think about. The ad platforms are already encroaching on adjacent territory (Google's Stitch is one example), so it's plausible one of them ships website generation. Our bet is that whoever builds it can't be neutral: Google will always steer you toward more Google spend, Meta toward Meta, and neither can advise reallocating budget across channels. It mirrors what's happening with model providers — teams are realizing they don't want to be locked to one (a Cognition gives you Claude, Codex, and Gemini; lock into a single tool and you're cut off from the rest) — and we think marketers will similarly want a neutral, cross-platform layer that sits above the ad platforms rather than inside one. As the space matures, the market should converge on wanting an independent, cross-platform execution and observability vendor.
Our long-term moat is the data coming from a closed feedback loop: we can influence the inputs (what pages get built and shipped) and measure the outputs (conversion and revenue impact). Because we see both sides, we can run experiments, attribute lifts to specific site changes, and turn those performance improvements into durable learnings. As more customers connect ad accounts and run more iterations, those learnings compound into proprietary, cross-account insight a single-surface tool can’t build.
We're a painkiller for some teams and a nice-to-have for others, and we're upfront about it. People don't adopt when: they're locked into an ecosystem like Shopify and would lose its integrations; they have an in-house engineering team already shipping marketing pages fast; or the pain of stale/slow landing pages just isn't acute enough yet to warrant a separate platform. Knowing where we win head-to-head — and where we don't — is part of the open commercial roles.