Muzzy Lane

There is a genre of company that is easy to describe and hard to categorize, and Muzzy Lane is squarely in it. If you ask what it does, the honest answer is: it makes software that lets someone who is not a software developer build a video game that is secretly a test. This is a stranger business than it sounds, and also a more durable one.

Here is the setup. For roughly a century, the way schools have measured whether you learned something is to ask you to recall it - a multiple-choice bubble, a short answer, an essay. This works fine if the thing you want to measure is recall. It works badly if the thing you want to measure is judgment: can this person, dropped into a messy situation with incomplete information, make a reasonable series of decisions and live with the consequences. You cannot really bubble-in judgment. You have to simulate it.

Muzzy Lane's whole premise is that you should simulate it, and that building the simulation should be cheap enough that ordinary instructors can do it. That is the product. It is called Author, it launched in 2015, and it is a no-code platform: you assemble an interactive, branching, scored scenario by pointing and clicking rather than by hiring a game studio. The company reports that this platform and its predecessors have now delivered north of 53 million assessments to more than 6.2 million students, who have collectively made over 531 million individual decisions inside the scenarios.

The company did not start here, which is the fun part. Muzzy Lane was founded in 2002 by David McCool, an electrical-engineering graduate of MIT who had already co-founded a networking-hardware startup, Aptis Communications, that Nortel bought in 1998. McCool did not arrive in education from an education-reform background. He arrived from software, with the specific intuition that education was an under-served market for genuinely complex interactive software.

The first product was not a platform at all. It was a game - Making History, a World War II grand-strategy title that shipped beginning in 2007 and ended up in hundreds of high school and college classrooms. Students played nations, managed economies, negotiated alliances, and, in doing so, absorbed the causal texture of the interwar period in a way a textbook chapter does not deliver. It was a real game that people played on purpose, which is a low bar that most educational games fail to clear.

Making History taught McCool the thesis, but it also taught him the constraint: custom games are expensive to build and do not scale. You cannot make a hand-crafted title for every subject. So the interesting asset was never the game - it was the engine and the authoring workflow underneath it. In 2016 the company made the tell: it sold the Making History franchise to Factus Games and went all-in on the tools. It gave away the thing that had made it famous to keep the thing that could make it big.

In between, Muzzy Lane did what a lot of quietly sensible edtech companies do: it became infrastructure for people with distribution. It built the Practice line of business simulations with McGraw Hill - Practice Marketing, Operations, Government - the kind of interactive activity that gets embedded inside a college course's digital materials. It built experiences for National Geographic, including an Underground Railroad title. It licensed tools to Pearson and Cengage. If you took a business course in the 2010s and clicked through a marketing simulation, there is a reasonable chance Muzzy Lane's engine was running underneath it and you never knew its name.

For most of its life, Muzzy Lane was a good, small, unglamorous company doing sensible work. Then generative AI arrived and, more or less overnight, made a large category of assessment obsolete. If a chatbot can write your essay, the essay stops measuring you. Every dean in America started asking the same panicked question: how do we assess anything now?

Muzzy Lane had, essentially by accident, spent two decades building the answer. A roleplay simulation where the “answer” is not a paragraph but a sequence of judgment calls is much harder for a language model to fake, because there is no text to generate - there is a situation to navigate. The company started describing its assessments as “AI-resilient,” which is marketing, but marketing that happens to be true. It also, sensibly, integrated generative AI into the Author suite itself - as a tool for authors building scenarios and as an in-scenario coach - so it is using the technology rather than merely defending against it.

The second act of that pivot is SkillBuild, launched in 2023 with the nonprofit Education Design Lab. SkillBuild takes “durable skills” - the soft, hard-to-measure competencies employers say they want and cannot verify - and wraps them in simulation practice plus AI coaching, ending in a microcredential an employer is supposed to trust. Whether the market fully believes those credentials yet is an open question. But the logic is clean: a skill only counts if you can prove someone has it, and a simulation is a defensible way to generate that proof.

A note on scale, because it reframes everything. Muzzy Lane is a company of roughly 13 people. It has reached more than 6 million students. That is on the order of 475,000 learners per employee, which is only possible because the company does not teach students - it arms the people who do. The customer list runs through higher education (Arizona State, Western Governors, George Mason, the University of Kansas, CSU Global, Herzing, Excelsior) and through the publishers who bundle its work. Revenue is modest - an estimated few million dollars a year - and the outside funding was always small: a seed round of about $450,000 around 2015, plus non-equity assistance from Massachusetts economic-development programs. This is not a venture rocket. It is a patient, engineering-led company that found a real problem and stayed on it long enough for the world to come around.

What can you actually do with it? If you are an instructional designer, you can build an interactive case study in an afternoon instead of filing a six-month engineering request. If you are a publisher, you can drop scored simulations into course materials without standing up your own game team. If you are a university, you can assess critical thinking in a way a chatbot cannot ghost-write. And if you are a learner, you get the rarer thing: an assignment that looks like the job, not the answer key.

It is worth being precise about what that last claim buys. A well-built roleplay scenario does not just check whether you picked the right option; it records the path you took to get there - the order of your decisions, the tradeoffs you accepted, the information you ignored. That path is data, and 531 million recorded decisions is a large corpus about how people reason under pressure. The open-response feature that instructors keep singling out matters for the same reason: it lets a student explain a decision in their own words, which is both harder to game and richer to grade. The through-line, from a WWII strategy title in 2007 to an AI-coached credential in 2024, is that Muzzy Lane keeps building the same object - a situation with consequences - and keeps getting better at measuring what people do inside it.

Muzzy Lane is not alone in believing that simulation is the future of assessment. There is a small, energetic field of companies making adjacent bets - Mursion and Bodyswaps on avatar-and-VR roleplay, Forio on data-driven business sims, various publisher-built suites doing pieces of the same thing. What distinguishes Muzzy Lane is less any single feature and more the authoring layer: the wager that the winning company is not the one with the flashiest simulation but the one that makes simulations cheapest to build. If you own the tool that a thousand instructional designers use to make ten thousand scenarios, you do not need to build the ten thousand scenarios yourself. That is a platform posture, and it is a patient one.

The culture reflects the posture. This is a long-tenured, engineering-and-pedagogy team, not a churn-and-burn growth org. It pairs software people with instructional designers, which is the correct combination for a product whose hardest problems are half technical and half about how humans actually learn. The company is remote-friendly and small enough that everyone can see the whole board. McCool has been at it for two decades in an industry famous for exhausting its founders, which is its own kind of signal: the people who win in education tend to be the ones stubborn enough to still be there when the market finally agrees with them.

The neat irony is that Muzzy Lane spent years looking like a company that had wandered from games into the less exciting business of assessment, and it turned out the assessment business was the exciting one all along. The name, incidentally, comes from a New England street - a small, specific, slightly obscure thing, which is about right for a company that has been quietly specific about one idea for twenty years: people learn by doing, so measure the doing.