After racking up 10 million signups for its AI presentation software in simply 9 months, Tome is releasing new options aimed toward getting these customers to stay round.
Tome can flip textual content prompts into multi-page slideshows, and it credit TikTok chatter as a giant a part of the jumpstart. The brand new model helps longer textual content prompts and provides customers extra management over the content material and formatting of every slide. Future updates will permit customers to import third-party photos and knowledge sources into their shows.
Keith Peiris, CEO and co-founder of Tome, says the replace displays how generative AI can change into greater than only a novelty. Whereas Tome went viral with the promise of producing total slideshows with only one sentence of textual content, the brand new model goals to get folks again on observe.
“I believe this can be a enormous step ahead in making AI extra helpful,” he says.
How Tomé works
Tome offers customers with a easy textual content field describing the kind of presentation they wish to create. The positioning then gives a selection of formatting choices and fills every slide with AI-generated textual content and pictures. Then customers can customise the slides with their very own textual content and pictures or use AI to make additional refinements.
Whereas customers can nonetheless generate total slideshows with only a single immediate, Tome now permits for numerous levels of human intervention. First, a top level view is usually recommended for every web page, permitting customers so as to add, delete, or rearrange their matters earlier than producing their slides. Customers can then select from 4 structure choices per web page.
“The entire level of that is you can repeat step-by-step with the LLM,” says Peiris.
Tome is free for private use, though customers are restricted to 500 AI directions, or “computation credit,” earlier than hitting a paywall. A plan that enables limitless AI use prices $10 per 30 days or $96 per yr.
For who’s it?
Peiris says Tome has gained traction amongst startup founders, distributors and college students. Regardless of AI’s tendency to generate boring, generic textual content, Tome says he is usually a helpful brainstorming software. The place a program like PowerPoint may require folks to first provide you with their concepts after which construct a presentation round them, Tome has them work backwards, with AI offering the inspiration on which to border their concepts.
“Even when you’re an skilled practitioner of storytelling, that is a helpful assemble,” says Peiris.
But the strategy has its limits. To create shows which have one thing distinctive to say, customers want to have the ability to contribute their very own knowledge and content material.
To that finish, Tome already has a helpful ‘Doc to Presentation’ software, which permits customers to stick a number of paragraphs of textual content and summarize them within the type of a slideshow. Peiris says customers will even have the ability to add photos within the coming weeks. The flexibility to import different knowledge sources instantly into Tome will observe in a couple of month.
If all goes based on plan, customers will have the ability to generate much more superior slides, together with charts and graphs, primarily based on person knowledge.
“It is best to have the ability to dump a Google Sheet or connect with your knowledge warehouse and have it populate,” says Peiris.
Keep one step forward
Whereas Tome likes to brag about its 10 million registered customers, the corporate will not say what number of of them will proceed to make use of the service. Peiris acknowledges that Tome was visited by quite a lot of “AI vacationers” in its early days, and that individuals usually had exaggerated concepts about AI that they might use with a easy textual content immediate.
“I believe we have needed to type of reduce trade expectations,” he says. “Even when you’ve got a acutely aware being in entrance of you, it nonetheless cannot predict what’s in your head primarily based on three phrases you say.”
Nonetheless, he says Tome is beginning to see a stronger improve in retention as customers achieve a extra reasonable understanding of what AI can do. The startup says billions of operations occur on its platform each month and Gen Z accounts for a 3rd of its customers.
The massive query is whether or not Tome can stay related as bigger corporations launch their very own AI presentation builders. For instance, Google’s Duet AI for Workspace can now generate Slides shows utilizing knowledge from Sheets and Docs as a part of a free trial for enterprise prospects, and Microsoft is previewing a technique to create and edit PowerPoint slides with its Copilot AI software.
Peiris responds that Tome is constructing his personal AI chops to compete. The startup’s head of AI and machine studying, Ves Stoyanov, was a scientific researcher at Fb’s AI group for 9 years. The corporate’s different AI hires embrace Sarah Sachs, Robinhood’s former head of generative AI and pure language processing, and Archana Sankaranarayanan, a former Googler whose workforce developed the persona for Google Assistant.
Tome trains its machine studying fashions to higher perceive what customers need of their shows and to be extra constant within the outcomes. Peiris argues that constructing Tome round generative AI from the bottom up will give the corporate an edge over incumbents.
“You must construct this format the place you will have a semantic understanding of the whole lot, and also you want the LLM to provide it,” he says. “I simply do not assume you are able to do that when you repeat PowerPoint from the Nineteen Nineties.”
Peiris additionally factors to the generational hole in Tome’s person base as a bonus. These customers haven’t got ingrained habits from a long time of PowerPoint and use Tome to construct issues like web site designs, resumes, and temper boards.
“I believe that is the start of a brand new kind of communication software,” he says.