This page shares some of the thinking and tools behind how I help early-stage AI companies land their first enterprise customers: discovery templates, ROI models, close plans, and account planning tools I've built, plus a few resources I trust. Remix freely.
📋 Table of Contents
🗞️ Published August 5, 2026
For all the power of frontier AI models, and their increasingly adept open-source counterparts, they still struggle with a fundamental lack of consistency, owing in large part to their probabilistic architecture.
Much like biological systems, these models are "grown" rather than "built." And just as it's hard to know why your cat has suddenly declared its own tail an arch-enemy, it can be hard to know why a model capable of crushing the LSAT insists you walk, rather than drive, to the car wash.
For low-risk use cases where humans are closely in the loop – say, reviewing the draft of an email – these errors and hallucinations are no big deal. They're easily caught, and when one slips through, the consequences are generally bearable.
As we move up the adoption curve into higher-stakes workflows with less human review, however, the equation flips: a once-in-a-hundred hallucination becomes a real liability. Think of an agent issuing an erroneous refund, citing a hallucinated legal case, or serving as the vehicle for data exfiltration – errors that can end a customer relationship and earn a company a day in court.
Hence the large and growing ecosystem of tools built to "wrangle" these models into consistency: wiring telemetry into agents and observing how they behave, running large-scale evals across their outputs, and/or enlisting other LLMs as judges. All of it sits atop more research-driven, less immediately productizable work to peer inside the "mind" of these models, through techniques like sparse autoencoders (SAEs) and mechanistic interpretability (hat tip, Neel Nanda).
So how do these platforms fare? How close can we get AI systems to behavior consistent and reliable enough for meaningful enterprise use cases? How does an enterprise decide between open-source and commercial alternatives? And, from a GTM perspective, how does a startup capture enough value to sustain an engineering team and fund the support and custom development that open-source efforts typically don't provide?
In this post, I'll walk you through my outside-in perspective on the (broadly defined) AI reliability space. We'll cover:
A note on where I'm coming from: I cut my teeth in enterprise design software, first on the GTM team at InVision, later as an Enterprise AE at digital collaboration platform Mural, carrying $1.3M+ annual quotas and closing deals north of $600K ARR. I watched this product category take hold and grow, and I'll draw on those lessons throughout. I look forward to your comments, edits, and pushback, especially from sellers and operators closer to today's sales motions. Let's learn together. Onward.