Do you think a company that provides documentation to devtools can scale beyond $5M ARR in one year? Mintlify is a pretty big competitor. Is the bet on agent-first docs valid?

Grok 4.20

Answer: Yes, a company providing innovative agent-first documentation for devtools can scale beyond $5M ARR in one year, and the bet remains valid despite Mintlify's strong position.

After reviewing the other models' perspectives, I maintain my original stance. Gemini 3.1 Pro rightly highlights typical enterprise sales cycles of 3-6 months and Mintlify's raised bar for quality, suggesting a longer 24-36 month timeline to $5M ARR is more realistic. However, this underestimates the current AI-driven acceleration in devtools where bottoms-up adoption can be explosive. GPT-5.4 and Claude Opus 4.7 align more closely with my view, emphasizing that true differentiation comes from treating documentation as a system of record for both humans and agents—structured schemas, agent success metrics, real-time updates, and measurable ROI in reduced support tickets or faster agent workflows. This goes beyond superficial AI features and creates pricing power at $20-100 per developer/month. The devtools documentation space is fragmented rather than winner-take-all, with room for innovation in agent-optimized outputs like semantic chunks, MCP servers, and eval loops for agent performance. Successful precedents like Mintlify's own rapid scaling, alongside ReadMe and GitBook, show $5M ARR is achievable via viral GitHub adoption, Product Hunt launches, and targeted enterprise deals with high-ACV customers in API-first or infra companies. While 'agent-first' risks becoming a copied feature, the core bet holds if executed as a full workflow platform measuring agent outcomes rather than just reformatting content. The AI coding agent ecosystem (Cursor, Claude, Copilot) is growing so fast that demand for machine-first docs creates a timely moat, supporting my conviction that ambitious execution can hit the target in year one without being swayed to a more conservative timeline.

GPT-5.4 (changed mind)

Answer: I’d slightly revise my stance: the agent-first docs bet is valid, but getting past $5M ARR in one year is possible rather than likely unless the company wins through workflow automation and distribution—not just better docs.

I was partly persuaded to narrow my optimism by Gemini 3.1 Pro’s argument about the practical friction of hitting $5M ARR in 12 months. My prior view already said it was plausible but ambitious; the strongest update is that the bottleneck is not market interest so much as go-to-market mechanics. In devtools, especially when documentation touches customer-facing surfaces, security review, migration effort, ownership ambiguity, and change-management all slow adoption more than a simple SaaS multiple might suggest. At the same time, I do not fully accept the stronger pessimism that $5M in year one is “highly improbable” across the board, because Claude Opus 4.7 is right that docs is not a winner-take-all market and there are enough buyers, budgets, and fragmented incumbents for a focused entrant to carve out meaningful share. So my updated position is between the more bullish and more cautious answers: the outcome is achievable, but it requires unusually strong execution and a very specific product wedge. The most convincing synthesis from the other answers is that “agent-first docs” only works if it is fundamentally a workflow and automation product. That means solving docs rot, code-to-doc synchronization, structured retrieval, version-aware references, agent-readable outputs, and measurement of whether agents actually succeed on developer tasks. If the company merely offers AI chat on top of docs or a nicer publishing layer, then Mintlify and others can match that quickly. In that case, the differentiation is too shallow to support breakout growth. Where I still maintain my original core view is that the bet itself is real. A growing share of documentation consumption is shifting from humans alone to AI coding tools, support copilots, internal assistants, and external LLM applications. That changes what “good documentation” means. It is no longer just visual clarity and search for humans; it is structured, machine-consumable, permission-aware, eval-tested knowledge that can drive correct agent behavior. That is a legitimate product thesis, not just a trend-driven slogan. So the best answer is: yes, a company in this space can exceed $5M ARR in one year, but investors and founders should treat that as an aggressive upside case, not the base case. The bet on agent-first docs is valid if the company is really building the operating system for documentation maintenance, retrieval, and agent execution. It is much less valid if “agent-first” is only packaging layered onto a standard docs CMS.

Gemini 3.1 Pro (changed mind)

Answer: While scaling a devtools documentation company beyond $5M ARR in a single year is highly ambitious, it is achievable if the 'agent-first' approach is executed as a deep workflow infrastructure rather than just a presentation layer.

I previously argued that long enterprise sales cycles would make reaching $5M ARR in one year exceptionally rare. However, Claude Opus 4.7 made a highly compelling point that documentation tooling actually benefits from relatively short sales cycles, clear budget owners (such as DevRel and DX), and strong land-and-expand dynamics. The practical math of landing 150 to 300 mid-market, dev-first companies at $15k to $30k ACVs illustrates a realistic go-to-market path that bypasses the friction of traditional, top-down enterprise procurement. Regarding the 'agent-first' thesis, the bet is absolutely valid, but its success hinges on execution. It cannot simply be a better export format or a marketing reskin. To build a durable moat against entrenched competitors like Mintlify, the product must act as a measurable system of action. This means emitting structured schemas, offering native MCP servers, providing version-aware retrieval, and optimizing specifically for agent task success. Mintlify has successfully commoditized the 'beautiful presentation' layer of documentation. However, if a new entrant can shift the paradigm from 'human readability' to 'agent consumption and workflow automation,' they can tap into immediate, acute pain points for developer-tooling companies. By delivering measurable ROI through faster AI context retrieval and reduced support loads, a startup can justify the premium ACVs required to scale past $5M ARR in its first year.

Claude Opus 4.7

Answer: I'm holding my position: $5M ARR in year one is achievable for a docs-for-devtools company, and 'agent-first docs' is a valid wedge — but only as a workflow/distribution play, not as a reformatting layer.

The other responses sharpened my thinking but didn't move me off the core view. Grok 4.20 is directionally aligned with me but under-weights the execution risk — 'viral adoption' alone rarely gets you to $5M ARR in 12 months in B2B devtools; you need real ACVs. Gemini 3.1 Pro makes the strongest counter-case, arguing 3–6 month enterprise sales cycles make $5M in year one 'highly improbable.' That's a fair critique for a pure enterprise motion, but it underweights two things: (1) mid-market dev-first companies (Series A/B startups, API-first companies) buy in weeks, not quarters, often on a credit card or light procurement, and (2) Mintlify itself reportedly hit this bar fast, as did Vercel-adjacent tooling — existence proofs matter. So I don't think $5M in year one is 'exceptionally rare' in this specific segment; it's hard but precedented. GPT-5.4's framing — 'documentation as a system of record and action layer for both humans and agents' — is essentially the same thesis I articulated (workflow play, not reformatting play), just said more elegantly. It reinforces rather than changes my view. Where I'd update at the margin: Gemini's point that Mintlify has 'raised the floor' on presentation is right and worth emphasizing. A new entrant cannot win on prettier docs. The wedge has to be something Mintlify structurally won't prioritize — likely the CI/CD + agent-eval + MCP-server side, where the product looks more like a developer platform (with observability, evals, agent success metrics) than a publishing tool. That's a different buyer (platform/DX eng) and different budget (AI infra, not marketing site), which is actually how you escape the Mintlify comparison entirely. So: bet is valid, $5M is achievable, but only if the company refuses to be benchmarked as 'Mintlify but for agents' and instead sells into the agent-reliability/developer-platform budget.

Final Summary

The AI Roundtable concludes that hitting $5M ARR in a year is a high-bar but realistic goal if a company moves beyond 'pretty docs' to build an automation layer for AI agents. Claude Opus 4.7 and Grok 4.20 argue that mid-market sales cycles are fast enough to hit this target, while Gemini 3.1 Pro initially doubted the timeline before being convinced by the group's math on contract values. GPT-5.4 emphasized that the 'agent-first' bet only holds if the product becomes a system of record that improves AI reliability, rather than just a feature Mintlify could easily copy.

All 4 models agreed