March 2026  ·  David Richards  ·  Market Assessment

Assessing Flint's Opportunity

An assessment of Flint's market opportunity, themes from user research, ideas for educational content, and open questions worth exploring.

00
The Opportunity

For two decades, B2B marketing tools have supported one primary distribution model: help marketers get human buyers to visit a webpage, understand the value, and convert. SEO, paid search, ABM, and more are all designed to win attention in the traditional funnel.

While that model isn't going away anytime soon, a second one is emerging alongside it.

AI agents are beginning to discover, evaluate, and recommend software on behalf of users. Gartner projects that organic traffic will fall by as much as 50% by 2028. At the start of 2025, roughly 1 in 200 website visits came from an AI agent. By the end of the year, it was 1 in 31. The web's next power user will be an agent.

The required infrastructure is already under construction. Cloudflare launched "Markdown for Agents" in February, automatically converting HTML into clean Markdown so AI agents can read web content efficiently. Google published the A2A protocol for agent-to-agent transactions. Microsoft is building a publisher content marketplace for the agentic web. Tavily, a search engine built specifically for AI agents, was acquired for $400M. These projects are early infrastructure for a web that will eventually serve two audiences: humans and AI agents.

Flint's MCP and API are an important part of its GTM strategy.

Traditional
Awareness → Trial → Purchase
Months of marketing to get someone to visit your site, sign up, try the product, and convert. You need the buyer to know your name. Flint already excels here — ABM pages, comparison pages, campaign landing pages, all on-brand and fast.
Emerging: Agent-Era
User asks AI → AI picks the tool
The agent discovers, evaluates, and acts. The buyer might never visit your site directly. This model is just starting to take shape, but the companies building for it now will have a real head start.

The MCP is a new distribution channel for Flint. Today, it enables discovery: a marketer asks Claude to build a landing page, and Flint can be the tool Claude reaches for. Down the road, it could enable end-to-end autonomous use: agents spinning up pages without a human in the loop.

The API opens a different, more familiar channel: integration with the established marketing stack. CRM platforms, ABM tools, campaign orchestration — these are the systems marketers live in today. An API that lets Clay, Relay.app, or a CRM workflow trigger Flint page creation is distribution through existing infrastructure. That's a motion Flint can run today.

The meta opportunity

The emerging agent-era model isn't just important for how Flint gets distributed. It's important for Flint's customers too.

Every company using Flint will soon need those landing pages to work for two audiences:

1. The human buyer who lands on a comparison page and needs to be convinced.
2. The AI agent evaluating tools on behalf of someone who asked "what's the best solution for X?"

Flint already has a GEO offering, but the opportunity goes further: structured data that agents can parse, content signals that help agents understand value propositions, Markdown-friendly output that works with infrastructure like Cloudflare's. Flint could become the tool that helps its customers win in both distribution models, not just the traditional one.

This raises two key questions that marketing must eventually determine: how does Flint become the default tool that AI agents reach for? And how does it build a strong, memorable brand that makes human buyers choose it too?

01
User Research

Given the MCP and API launched just weeks ago, I spoke with three marketers in my network to get a feel for how they think about a solution like Flint: a growth marketer, a product marketer, and a GTM ops lead. I wanted to understand who would actually evaluate and buy a programmatic landing page API, what would make them trust it, and how they discover new tools in the first place. Newsletters like MKT1 and influencers like Kyle Poyar and Adam Robinson have outsized influence on what these people try.

📈
Growth Marketer
Paid Acquisition Lead
B2B tech company
🎯
Product Marketer
AI Launch Lead
Enterprise SaaS
⚙️
GTM Ops Lead
Revenue Infrastructure
Series A AI startup
Theme 01
The bottleneck is production, not ideas.
Campaign windows close before pages go live. Launches slip because web work is stuck in someone else's sprint. Copy takes a day. Getting it designed, approved, and deployed takes weeks. When a tool solves this pain, they pay attention fast.
Theme 02
The evaluator and the buyer are different people.
The GTM engineer tests and integrates. The marketing leader owns the budget. The API has to work for both. At larger companies there's another layer: even with budget, a PMM might not have permission to ship without brand review and web team approval. One product marketer at a large enterprise told me the issue is coordination overhead. He doesn't have the autonomy to implement a new page on his own. That suggests Flint's sweet spot is growth-stage companies (50-500 people) where the marketer can say yes and ship.
Theme 03
Trust is earned through output, not promises.
One person wanted conversion data before putting ad spend behind AI-generated pages. Another wanted messaging fidelity. The third wanted clean docs and consistent API results. But they all said the same thing: show me it works.
Working Hypothesis
The GTM engineer is the entry point. The marketing leader is the champion. The sweet spot is growth-stage companies where marketers have the autonomy to ship. Comparison pages and ABM microsites are the wedge use cases that get both roles excited.
02
Educational Content

Three content ideas. Each ties to a real Flint capability and a different audience.

Idea 01
The GEO Playbook: How to Get Cited by ChatGPT, Perplexity, and Google AI Overview
Nobody owns this topic yet. The answer to "how do I show up in AI search?" is "produce more structured content, faster." That's what Flint does. This is the HubSpot "inbound marketing" play: define the category, teach the market, become the solution. Gate it lightly, repurpose into a blog series, promote through MKT1 and influencer channels.
Idea 02
From 1 Page to 341: The ABM Landing Page Playbook
The proof already exists. Amigo.ai generated 341 ABM pages in one click. Tandem tripled paid media conversions. Start with those stories, teach the framework, show how to measure whether personalization actually works. Product-led but genuinely educational. Growth marketers would share this.
Idea 03
MCP for Marketers: How AI Agents Are Changing Marketing Workflows
Most marketers probably don't know what MCP is yet. That's an opportunity. Explain what happens when Claude reaches for tools on your behalf, show it in practice, and connect it to why marketing teams should care. Riskier because the audience is smaller, but positions Flint as the company that saw this shift first.
03
Open Questions

Questions worth exploring.

On the API GTM motion
The API is enterprise-only today. Is there a self-serve version that unlocks bottom-up adoption? Or is the wedge MCP and pre-built integrations (Clay, Relay.app) rather than raw API access?
On positioning around conversion, not creation
"Make your perfect landing" sounds like the value is design. But the real value is knowing what converts: analytics, GEO rankings, A/B tests across thousands of pages. Getting the right pitch is the how. Converting the right champion is the why. How do we lead with that?
On category ownership
Segment owned "CDP" by repeating it for years. If Flint wants to own "autonomous websites," what's the content engine? GEO thought leadership seems natural since it connects directly to why companies need more pages, faster.
On integration priorities
Clay, Relay.app, and MCP are live. How were these prioritized? I imagine we're tracking where current Flint users are building automations. We should continue to prioritize those that are most popular with Flint's user base. I also couldn't find Flint in the list of Claude connectors -- is this in the works?
04
Reference: API Precedents

Four API companies that got sticky.

StripeDeveloper experience IS the product
$5.1B rev$95B valuation50% of Fortune 100

Seven lines of code replaced a months-long sales cycle. Developers became the sales force. Uber went from tiny to paying $30M/year on one integration. Now has a Claude MCP connector, positioning for AI agent distribution.

Lesson for Flint

First API call needs to take under 5 minutes. If a GTM engineer can generate a page faster than filing a Jira ticket, Flint wins.

TwilioWin the users, the buyers follow
$4.1B rev4% annual churnAcquired SendGrid $2B + Segment $3.2B

Went straight to developers, not CTOs. Once integrated, nobody ripped it out. 4% annual churn. Later acquired SendGrid and Segment to own more of the stack.

Lesson for Flint

Don't start with the CMO. Let GTM engineers find Flint through MCP, Clay, or docs. Once they've built something, they'll fight for budget.

VercelOwn the framework, own the ecosystem
$200M+ ARR$9.3B valuationBacked by Accel

Built Next.js open-source, then became the best place to deploy it. Now v0 generates code from prompts that deploys instantly on Vercel. Already has a Claude MCP connector. AI generates output, Vercel provides infrastructure.

Lesson for Flint

Flint's MCP follows the same logic. Stripe and Vercel are already in Claude's Connectors marketplace. If Flint isn't there yet, that's a quick win worth exploring.

Recall.aiEarly-stage infrastructure playbook
$10M ARR25 people$250M valuation

Closest analog to Flint right now. Universal meeting recording API. Self-serve free tier, great docs, strategic investors who were also customers (HubSpot Ventures, Salesforce Ventures). Once integrated, nobody switches.

Lesson for Flint

Get embedded into 3-5 recognizable GTM stacks. Once Flint is wired into someone's CRM pipeline, it's part of the infrastructure.

PatternStripeTwilioVercelRecall.ai
Self-serve first call✓✓✓✓
< 5 min to value✓✓✓✓
Docs as marketing✓✓✓✓
Bottom-up adoption✓✓✓✓
Claude MCP connector✓—✓—
Category ownedPayments APIComms APIFrontend infraMeeting bot API

The common pattern: Make it easy to try. Let users become advocates. Embed into workflows. Become the thing nobody wants to rip out. MCP compresses that whole sequence. Flint's MCP integration is probably the single highest-leverage thing to invest in right now.

This is independent research.
It is not affiliated with or endorsed by Flint.