Feb 29, 2024 · 40m · in-depth
Scaling and selling AI products for enterprise | May Habib (Co-founder and CEO of Writer)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this First Round Capital podcast episode, Writer co-founder and CEO May Habib outlines her strategic playbook for building, positioning, and scaling an enterprise-grade generative AI platform. She explains how proprietary domain-specific LLMs, rigorous champion qualification, and deep operational integration enable sustainable growth and extraordinary customer retention in a rapidly shifting technology market.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Brett holds 1.4% of the talking time here. How this is scored →
Brett as informed peer 3.3 Guest teaching 4.9 Guest disagreement 1.1 Brett pushing back 0.1
speaking balance: gold is Brett, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 30:20 Dismissing hype around future foundation model capabilities
May forcefully pushes back against industry anxiety regarding future OpenAI releases, stating that transformers cannot bypass internal enterprise data boundaries.
Hardest push from Brett ▶ 23:57 Pressing on how to identify innovation fundsTodd directly interrupts and challenges May to provide a concrete diagnostic test for distinguishing funny money from real operational budgets.
Biggest teaching moment ▶ 12:24 Differentiating RAG knowledge graphs from traditional searchMay clearly educates the host and listeners on retrieval-augmented generation mechanics and how appending structured data to LLMs delivers net-new work product.
Brett holds their own ▶ 15:41 Synthesizing execution discipline versus technological market shiftsTodd demonstrates sharp domain grasp by framing the fundamental tension founders face between maintaining disciplined product execution and adapting to rapid AI paradigm shifts.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Brett as informed peer | Guest teaching | Guest disagreement | Brett pushing back | Why |
|---|---|---|---|---|---|---|
| May Habib's Background and the Origin of Writer | 3 | 3 | 0 | 0 | Todd introduces the conversation and asks foundational questions about May's entrepreneurial background and the transition from Cordoba to Writer. May explains her motivations and the technical origins of Writer in a collaborative, reflective tone. | |
| Deliberate Market Validation and Early Research | 3 | 4 | 0 | 0 | Todd asks how May validated early market demand. May details her deliberate customer discovery process interviewing enterprise buyers of competing tools like Grammarly. | |
| Evolving to a Full-Stack Enterprise Generative AI Platform | 4 | 5 | 1 | 0 | Todd inquires about the expansion across multiple departmental use cases. May breaks down how Writer evolved into a full-stack platform and how selling shifted from marketing leaders to enterprise IT and AI executives after ChatGPT. | |
| Adapting to RAG and Maintaining a Zero-to-One Mindset | 3 | 6 | 0 | 0 | Todd prompts May to define RAG for the audience. May delivers a thorough technical breakdown contrasting basic enterprise search with retrieval-augmented generation on structured and unstructured knowledge graphs. | |
| Filtering Market Noise and Establishing Strict Product Boundaries | 4 | 5 | 2 | 0 | Todd asks how May balances roadmaps against rapid market evolution. May adamantly delineates what Writer refuses to do, rejecting consumer chatbots, ticket deflection, and SMB segments. | |
| Qualifying Enterprise Champions and Demonstrating Massive ROI | 3 | 5 | 2 | 0 | Todd asks how May established such sharp focus. May explains how Writer disqualifies prospective buyers who are merely chasing shiny AI toys rather than delivering measurable business ROI. | |
| Enterprise Readiness and Deep Operational Embedding | 3 | 5 | 1 | 0 | Todd asks how a startup can successfully close major enterprise accounts early on. May explains the necessity of heavy enterprise plumbing like SCIM, SOC2, and embedded solution architecture. | |
| Tactical Advice: Selling Real Business Value Over Innovation Hype | 4 | 6 | 2 | 1 | Todd prompts May for actionable advice for founders entering enterprise AI and actively asks how to spot vanity budgets. May strongly warns against selling into innovation POCs and funny money. | |
| Defining Product-Market Fit and Enterprise Retention Metrics | 3 | 5 | 1 | 0 | Todd asks about the metrics indicating true product-market fit. May details Writer's 209% net retention rate and emphasizes evaluating deep user action over shallow login activity. | |
| Specialized LLM Strategy and Data Sovereignty | 4 | 6 | 3 | 0 | Todd invites May's contrarian hot takes on the AI landscape. May dismisses fears around generalized models like future GPT iterations, arguing enterprise data privacy and fine-tuning outweigh raw scale. | |
| The 2024 Roadmap: Large Reasoning Models and Knowledge Graphs | 3 | 5 | 1 | 0 | Todd asks about Writer's upcoming roadmap. May outlines their transition toward large reasoning models for agentic enterprise execution and graph-based RAG architectures. | |
| Founder Resilience and the Power of Team Cohesion | 3 | 4 | 0 | 0 | Todd wraps up with questions on founder resilience and co-founder alignment. May shares how her culture of transparent, open debate with her co-founder prevents internal politics. |
Statements from this episode (24)
Habib: Unequal AI access will cause 100x more inequity than language barriers
“Being able to use AI or not, Is likely to create a hundred X more inequity in society than all of us speak in different languages.”
Habib: Writing is the last unstructured enterprise business process
“Writing is super strategic, but is the last unstructured business process that a company does.”
Habib: Writer's largest models match GPT-4 quality at lower hosting cost
“Now we've got 18 models. The biggest ones are GPT-IV level quality, but at a fraction of the cost of hosting. Super, super fast for the enterprise use cases that That our customers run them on.”
Habib: Goldman, Vanguard, and UnitedHealthcare use Writer's industry-specific LLMs
“We now have a Palmyra financial services model, you know, customers like Vanguard and Northwestern Mutual and Goldman use a medical model that United Healthcare and other healthcare customers use.”
Habib: Most impactful enterprise AI sits between raw APIs and generic copilots
“And the most impactful use cases really sit in between kind of those two poles, but it is really hard, expensive, and time consuming to connect large language models with customers' data, with AI guardrails, with business users and business logic.”
Habib: Post-ChatGPT, Writer shifted to selling directly to IT and AI leadership
“Over the past year, we have really been selling to IT and the office of the head of AI. And then they really bring us into that first sort of big impactful use case.”
Habib: Enterprise generative AI adoption concentrates in CX and expert assist
“And it really tends to be one of two flavors, either a customer experience Use case. So that can be marketing or digital or a we call them expert assist. So a knowledge management type of use case.”
Habib: AI capabilities will outpace most organizations' capacity to absorb them
“And I do think the capabilities that are possible are gonna outstrip most organizations' capacities to absorb them, for most of their employees to absorb them without a lot of effort.”
Habib: Writer deliberately refuses to build chatbots or ticket deflection tools
“We don't do chatbots. We don't do ticket deflection. No one's gonna build something that sits on a website.”
Habib: Enterprises should use GPT-4 for zero-shot tasks without proprietary data
“If it's zero shot, they don't need any data, right? And it's not medical, it's not financial services, and it doesn't need to be on-prem, they should use GPT-IV.”
Habib: Writer deliberately refuses to work with agencies, SMBs, or mid-market
“We don't work with agencies. We don't work with SMBs. We don't work with the corporate segment.”
Habib: Writer deployed across 21 brands and 30 use-case families at L'Oréal
“We rolled out to 21 brands at L'Oreal, right? We know the thirty-ish use case families that we did across that whole company.”
Habib: Enterprise champions target $100M in cost reductions using Writer
“Showing people that we've got champions who are signed up with procurement to take a hundred million dollars of costs out of their organizations completely built on writer.”
Habib: Writer qualifies out enterprise prospects with weak champions, even CIOs
“And so our team is really good at politely declining to engage, right? Not just because the use case doesn't fit or it's not ICP, but we qualify the champion out, even if it's a CIO.”
Habib: Enterprise innovation and POC budgets are easily cut distractions for startups
“I think, you know, the deals around POCs and innovation and, you know, sort of, I call it funny money. You don't really want that stuff, actually. I think that is a distraction, because it's easy to cut. Your team will have spent a lot of time, and it'll be fo…”
Habib: Scaling GenAI in SMB and mid-market is hard due to switching
“And I think it is just so much easier in mid-market and SMB from a functionality perspective to just switch, right, to something else, especially in generative AI where it'll be confusing for a while. It's gonna be hard to scale a big business if your buyers a…”
Habib: Writer's smaller enterprise accounts maintain over 160 percent NRR
“I like to strip that away from deals that are bigger than a certain amount and then look at the NRR on just smaller accounts, right, who've got lots of optionality. That's still a 160 something percent NRR, which is amazing.”
Habib: Writer has the potential to become a 400 percent NRR company
“I think this should be a 400% NRR company.”
Habib: Most enterprise buyers reject data sharing and Azure AI vendor lock-in
“And for us, the vast majority of enterprises we talk to, that stuff can't leave their environment, and they're not excited about the kind of lock-in that an Azure AI requires, right?
And the way that that scales the amount of engineering effort that's required…”
Habib: Current transformer architectures fail at enterprise agentic workflows
“You know, our kind of tests on inference on quality for agent use at work. It's just transformers. Just, it was current capabilities, really not able to do what we wanted them to do. And so new architectures that we're working on are super, super promising.”
Habib: Vector databases fail at RAG retrieval beyond 50,000 pages
“Like, it's fine when I am, I've indexed 50 pages, but when it goes to 50,000, you know, I can't get this needle in a haystack out of my kind of VectorDB approach.”
Habib: Founders debating openly in leadership meetings reduces company politics
“It also really reduces politics for folks to see that, right? We have an executive meeting every other Monday, and then on the alternate weeks, it's a team leads meeting. So smaller meeting, a big meeting, but all leaders in the company. And, you know, we don'…”
Habib: Writer runs without regular one-on-ones to preserve asynchronous culture
“And so we're able to have a really low meeting culture as a result because, you know, we are able to do things as a group. That meeting is 90 minutes. Sometimes it goes two hours, a little bit longer. But, you know, as a result, I don't really have one on ones…”