Spellbook Went From Rally Legal to a $350M Valuation: Scott Stevenson's Story

October 19, 2026

TL;DR: Spellbook is an AI assistant that drafts and reviews contracts inside Microsoft Word. It was built by Scott Stevenson and his co-founders, who spent years running a legal-templates startup called Rally Legal before they found product-market fit. As Stevenson shared on The Product Market Fit Show, a side project built in about two weeks in 2022 showed "auto complete for lawyers," and customers reacted so strongly that the company moved everything to it. Spellbook raised a $20M Series A in January 2024 and a $50M Series B at a $350M post-money valuation in October 2025 (as of October 2026, the most recent equity round I could find). The episode aired in 2024, so the customer numbers Stevenson gave there, close to 2,000 law firms, were as of that interview.

A note on dates. This is a 2024 interview. Everything attributed to Stevenson below describes the company as it was then. Funding and valuation figures come from public announcements and press coverage and are labeled with their source and date.

What does Spellbook do?

Spellbook is a contract drafting and review tool for lawyers that works as an add-in inside Microsoft Word. A lawyer writes or opens a contract, and the AI suggests clauses, flags risks and proposes edits in the same document. The company is based in Toronto.

The first version was very simple. Stevenson described it on the show as a mirror of what programmers got from GitHub Copilot.

"Ours was auto complete for lawyers basically." — Scott Stevenson, Spellbook

The product has grown well beyond autocomplete since. In its October 2025 Series B announcement, Spellbook said it serves about 4,000 law firms and in-house legal teams in 80 countries and has reviewed more than 10 million contracts (company-reported, as of October 2025, via a Business Wire release and press coverage).

On the show, host Pablo Srugo called it "the ChatGPT for lawyers," a shorthand he admitted could be cleaner. The deeper point is where the product lives. Lawyers already work in Word, and Spellbook went to them instead of asking them to learn a new app.

Who founded Spellbook?

Spellbook's CEO and co-founder is Scott Stevenson. Public reporting (LawNext and Canadian business press, checked October 2026) lists his co-founders as Daniel Di Maria and Matt Mayers. The company started life as Rally Legal, and sources differ slightly on the founding year, with 2018 and 2019 both appearing. On the show, Stevenson placed the first product and the pre-seed round in 2019.

Stevenson has an engineering and product background. He started his first company, a musical instrument business, straight out of university. That is also where the idea for Spellbook began.

"Before I knew it, half of that angel check was gone on legal fees and I was like, what just happened?" — Scott Stevenson, Spellbook

As he shared on the show, that first angel check was $20K. Watching half of it go to basic legal work stuck with him through a stint as a director of engineering at another company. He also spent time in the Ethereum and smart contract world, believing contracts would become cheap code, but found lawyers had little interest in it. That led him to start Rally with two co-founders. He described one, Daniel, as a lawyer, and the other, Matt, as a mechanical engineer who had started a toy company.

His stated mission was to cut the friction in legal transactions. He framed it as asking what would happen if the friction of legal transactions fell by 90%.

How much has Spellbook raised?

All figures below were checked by web search in October 2026, with the source named. Stevenson also mentioned the two most recent rounds in passing on the show, roughly "10 million" in mid-2023 and "20 or so" six months later.

Pre-seed, 2019: about $750,000 as Rally (BetaKit, January 2024).
Seed, May 2023: $10.9 million (The Logic and Pulse 2.0, 2023 and 2024).
Series A, January 2024: $20 million, led by Inovia Capital with Thomson Reuters Ventures and The Legaltech Fund, bringing total funding above $30 million (Business Wire, January 2024). The release said revenue grew 10x from 2022 to 2023 and that customers had grown nearly 300% in the seven months since the seed announcement, to more than 1,700 law firms and legal teams (company-reported).
Series B, October 2025: $50 million, led by Keith Rabois of Khosla Ventures at a $350 million post-money valuation, with Threshold Ventures and returning investors including Inovia Capital, Bling Capital and Moxxie Ventures. Total funding passed $80 million (Business Wire and BetaKit, October 9, 2025).
Debt financing, March 2026: $40 million from RBCx, intended to fund acquisitions in a consolidating legal AI market (ABF Journal and MLex, March 2026).

Key stat: As of October 2025, Spellbook's Series B valued the company at $350 million post-money, according to BetaKit and Business Wire.

As of October 2026, a web search turned up no new equity round after the Series B. Spellbook has not published audited revenue figures, and statements about its growth rate are company claims.

How did Spellbook find product-market fit after years of searching?

The story is not an overnight success. Rally's first product, as Stevenson shared, was a template library for founders, launched in a couple of weeks. People paid about five bucks a month, but the founders saw it would take a long time to become profitable. Buyers still did not understand what the documents said and wanted someone to trust.

Around then, a lawyer at the firm Miller Thomson in Waterloo approached them. The team spent weeks working alongside lawyers and built a product for law firms: send clients a questionnaire, feed the answers into templates, generate the documents. Stevenson said they raised a pre-seed round from East Coast investors after some early revenue and signed a handful of law firms.

Then came what he called the valley of despair. The first customers loved the idea but did not actually use the product. Setting up the templates took heavy configuration, and when a live deal arrived, lawyers often decided their work was too bespoke for the engine. Enterprise implementations could also take a year or two, which meant the team could not tell whether it had product-market fit.

Two changes sped things up. They moved toward small, nimble firms that adopted quickly, and they set a hard bar based on CAC payback. As Stevenson put it, if payback was not clearly well under 12 months, they decided they did not have product-market fit.

They also ran experiments constantly. Stevenson said they launched more than a hundred landing pages with different value props, with a goal of one experiment every two weeks, and paired them with ads.

One hit on the sales side was a "menu of services" pitch, a kind of Shopify for law firms that put the templates on a lawyer's website. It cut the cost of getting a lawyer into a demo by about 90% versus earlier tests, as he shared on the show.

"It resonated because lawyers want to make money. They want clients and they want customers." — Scott Stevenson, Spellbook

But it failed the next test. Lawyers could set up a storefront but had no idea how to drive traffic to it.

Never miss a founder's PMF story

Subscribe to The PMF Show

"I think our CAC payback was probably six months on that whole motion or less, but it didn't pass the retention and the usage test." — Scott Stevenson, Spellbook

Key stat: the menu-of-services experiment had a CAC payback of about six months or less and still failed, because retention and usage were weak (as Stevenson shared on the show).

The lesson he drew was that a pitch can sell and still not stick. Most launches, he said, show what his team calls "the wiggles of death," a few spikes of usage and then nothing.

What was the counterintuitive move behind Spellbook?

Two decisions stand out.

First, staying small for years. The team stayed at roughly five to eight people while looking for product-market fit. Stevenson said the board was supportive but skeptical, since the usual pressure after a raise is to hire and scale.

"We decided we're not scaling the company, we're not upping our burn and we're just staying incredibly lean and agile until we have a product that people are pulling out of our hands faster than we can provide it and we're having enormous growing pains and everything's on fire." — Scott Stevenson, Spellbook

Second, building the AI product on the side, and under a new brand. The AI work started as evening and weekend hacking, not part of any plan. The team tried GPT-2 early, but it was not good enough. When GPT-3 and GitHub Copilot arrived, Stevenson saw something he could not unsee. Copilot riffed off his own bespoke code. Lawyers had always said templates were too rigid for their bespoke work. A model that could build on a lawyer's own text finally fit.

They built a first version in about two weeks, in summer 2022, before ChatGPT launched. It was a Word add-in with a draft button and a feature that explained a clause in plain language. The autocomplete took off. Stevenson said the demos got a strong reaction from the first lawyers.

"All the KPIs were off the charts. CAC payback was below one month. There was immediate CAC payback and usage was like we had never ever seen before." — Scott Stevenson, Spellbook

The team launched it on a separate website and brand, Spellbook, instead of bolting it onto Rally. Stevenson said he would put even the earlier landing pages under new domains if he did it again, because a growing feature list overwhelms customers. When they brought the growth chart to their board, he said, it showed the new product would eclipse three years of Rally work within three months. Rally still exists, and the templating engine is meant to fold into Spellbook over time.

Key stat: as he shared on the show, Rally had around 100 law firms, and Spellbook was close to 2,000 after roughly a year to a year and a half, with one to two million contracts a year going through the system.

Key lessons from Scott's playbook

1. Stay lean until usage proves it. Stevenson called staying small the right call even though it was hard to hold with a skeptical board. Pre-product-market-fit is a different game than post-product-market-fit. 2. Pick one test that is hard to fake. Spellbook used CAC payback under 12 months and usage retention cohorts. A pitch that wins on cost per demo can still lose on retention. 3. Choose features by pattern, not by story. Stevenson's team scores ideas against about 16 characteristics, such as zero time to value, a five-second GIF demo, no habit change and living inside the existing workflow. He said they use it to rank ideas, not to generate them.

4. Meet customers where they work. Lawyers rarely adopt new web apps, so the Word add-in was a deliberate bet when others skipped it.

"We knew how hard it was to even get them to go to a web app. That was crucial." — Scott Stevenson, Spellbook

5. Do not bolt everything onto one product. Launch new brands and drop sunk costs while you are still searching.

"Keep launching stuff. Don't worry about the sunk cost. It only takes one landing page and one feature to have that breakthrough moment." — Scott Stevenson, Spellbook

Listen to the full interview: Scott Stevenson on The Product Market Fit Show.

FAQ: Spellbook

Q: What is Spellbook? A: Spellbook is an AI contract drafting and review tool that works inside Microsoft Word. It started as a product of Rally Legal and is based in Toronto. In October 2025 the company said it served about 4,000 law firms and in-house legal teams in 80 countries.

Q: Who is the CEO of Spellbook? A: Scott Stevenson is the CEO and co-founder. Public reporting lists Daniel Di Maria and Matt Mayers as co-founders.

Q: How much funding has Spellbook raised? A: Spellbook announced a $50M Series B in October 2025, bringing total equity funding above $80M, after a $20M Series A in January 2024 and a $10.9M seed in 2023. It also announced $40M of debt financing from RBCx in March 2026.

Q: What is Spellbook's valuation? A: The Series B in October 2025 valued the company at $350 million post-money, according to BetaKit and Business Wire. A web search in October 2026 found no later equity round.

Q: Who invested in Spellbook? A: The Series B was led by Keith Rabois of Khosla Ventures, with Threshold Ventures, Inovia Capital, Bling Capital and Moxxie Ventures. Inovia Capital led the Series A.

Sources: Listen to the Full Founder Story

Scott Stevenson, Spellbook — spent years searching for product-market fit with Rally Legal, then built an AI contract assistant for Word in two weeks and grew it to close to 2,000 law firms by 2024.