Now valued at $300 million while pre-revenue, according to a source, Tome claims it’s the fastest productivity software maker to ever hit 1 million users since its release in September.
IN In a chat box under a dark, otherwise blank canvas, Tome co-founder Henri Liriani types in a fundraising pitch request for a telescope startup.
In less than a minute, Tome’s software delivers the presentation it asked for: eight slides organized by a table of contents, complete with introductory text, business model and sales plan, even cyberpunk-themed images. , all AI-generated. It’s not a camera-ready pitch deck, but it’s clearly a starting point for one. No need to wade through templates, text boxes, and image sizes in PowerPoint or Google Slides.
“There’s a degree of randomness to the output that I think is actually fun,” Liriani says. “It’s an interesting way to look at this type of tool as a thinking partner.”
Liriani and CEO Keith Peiris, both former Meta managers, tinkered with Tome for two years, before launching a free beta test in September and integrating OpenAI’s AI models two months later. In early February, San Francisco-based Tome surpassed one million users, 134 days after launch. This would make Tome the fastest productivity tool to reach this milestone, according to the company, based on research it reviewed and shared with Forbesahead of Dropbox, Slack and Zoom.
Now Tome is looking to capitalize. The company raised $43 million in a Series B funding round led by Lightspeed Venture Partners and featuring a host of other investors, including Coatue, Greylock, Stability.ai CEO Emad Mostaque, and former Google CEO Eric Schmidt. The round values Tome at $300 million after investment, up from $175 million in its 2021 Series A round, according to a source with knowledge of the increase, all without earning a single dollar in sales.
“We set out to create a company that can help anyone tell a compelling story,” Peiris said. Forbes in an interview. “Storytelling is the lifeblood of productivity for mankind, from cave drawings to fireside stories to PowerPoint.”
Frustration with PowerPoint, the iconic slide presentation program first released by Microsoft in 1987, was part of Tome’s inspiration, which developed from pandemic lockdown brainstorming sessions between Peiris, Liriani, and another former Meta product manager, Seth Rosenberg of Greylock, in 2020. That and Google Slides didn’t work well on mobile, they thought; and including images of prototypes, customer data, or other artifacts often meant sharing and embedding static screenshots. Developing large-language models like OpenAI’s GPT-series, meanwhile, could do for pitch decks what it did for text prompts. Just as many Pixar films follow the journey of a similar hero, an AI tool scanning 10,000 pitch decks could also learn their typical structures.
“You can’t just weld AI outside of a product.”
After other product managers said they would pay for a fuller version of a mockup he had quickly concocted, Peiris and Liriani began work on what was originally called Magical Tome, with billionaire Reid Hoffman, Rosenberg’s fellow Greylock, leading a seed round and joining its board of directors. In March 2022, after raising a Series A funding round led by Coatue, the company announced itself in a TechCrunch article that so far made no mention of AI.
Tome’s software rethinks PowerPoint’s system of uniform, slide-friendly tiles with software that can automatically place and size text and images dropped onto its canvas. Unlike a traditional layout, Tome tiles are modular, with changes to one dynamically updating the others. Since its AI integration, Tome also uses several large language models to perform a number of behind-the-scenes requests to respond to a user’s presentation prompt. First, its software generates a frame for the output, defining its length and style; then, smaller queries to AI models populate the corresponding images and text. Users can then refine the results or have the AI generate more pages through additional prompts.
Tome can generate everything from children’s bedtime stories to 3D prototypes, but it mostly sounded like a PowerPoint and slideshow killer. AI Meeting Notes, startup Supernormal, raised a funding round using a pitch deck built in Tome; others made sales presentations or visual blog posts.
Investor Michael Mignano, the former co-founder of podcast startup Anchor and now a Lightspeed partner, spotted Tome’s announcement video demonstrating his AI tools on Twitter and immediately asked former funder Nakul Mandan, founder of Audacious Ventures, to introduce himself. After a call with Tome’s founders on a Thursday evening, he flew to San Francisco the next morning and came up with a term sheet for his Series B within days. “It amazed me how quickly this high-quality asset could be created, and it was a format that feels so much more modern to me than any other presentation format that came before it,” Mignano said. (Mignano also published a Tome on investing on Wednesday.)
In conversations for a magazine article about the current generative AI boom, investors said Forbes One of the biggest questions for startups looking to build with AI tools will be whether they are creating defensible products and user experiences, or ones that will be easily copied, more or less, by incumbents. .
Tome, which plans to charge around $10 per user for a monthly subscription when rolling out its enterprise version this year, will face such skepticism. More than 30 years after its debut, PowerPoint is still a force, with millions of decks being created every day, and a Microsoft parent company that has invested heavily in OpenAI with a plan to integrate its models across its products. Google, which released Slides more than 15 years ago, squandered an early lead in AI but made shipping new products a priority. Then there are established design and collaboration software unicorns like Canva, Coda, and Notion, all of which have announced AI features and more are expected soon.
Peiris is unfazed. PowerPoint, with its large user base trained over many years of use, won’t be able to revamp its product too drastically, he said. “When people would say you have to make mobile-only or mobile-first products to make mobile, I think the same goes for AI,” agreed Hoffman, its first backer. “You can’t just weld AI outside of a product.”
Tome has also actively sought out allies: Besides Schmidt, investors from the productivity software space include executives from Airtable, Notion, and the venture capital arm of Zoom; in AI, Tome already had early access to GPT-4 through his investor ties to OpenAI, and brought in Mostaque and the CEOs of Adept AI Labs, Copy.ai, and Weights & Biases as investors. personal. After a recent OpenAI outage prevented Tome users, including this reporter, from generating presentations for a while on Tuesday, Peiris said he plans to meet with Stability AI and OpenAI rival Anthropic next week. to discuss using their models in an enterprise solution.
“I don’t see a world where you press a button and your pitch deck is done.”
In the coming months, the founders of Tome also plan to build and maintain their own specially trained models to make its generations more advanced; they describe a future in which teams can live-integrate self-updating Figma files or user metrics directly from a client’s database, while enabling live collaborative feedback to modify outputs by color scheme or tone. Human input will remain crucial, Peiris said, “I don’t see a world where you push a button and your pitch deck is done, and you send it to VCs, and they wire you money.”
However, not everyone will welcome Tome-assisted projects, as at least one early adopter found the path difficult. In January, Liriani tweeted screenshots of emails, with permission, from an anonymous high school student whose teacher discovered he had generated a class presentation in Tome. At the student’s request, Tome provided logs to support their claim that they only used Tome’s AI generations as a starting point. The teacher allowed the student to redo the project, up to 70% credit, and within the required Google Slides.