Neural Notes: Community-driven AI is the secret to Leonardo.Ai’s success

To say that it’s been a whirlwind six months for Leonardo.Ai is an atrocious understatement. During a period that is usually quiet in the startup space, Leonardo came in clutch with a mammoth $47 million raise in December.

If that was the shot, the chaser was certainly the announcement of an enterprise-grade product – Leonardo for Teams – just last month.

Leonardo is moving fast, and isn’t afraid to pivot. In fact, its entire business model has changed since it first launched in 2022.

Gaming assets to AI startup powerhouse

Originally, its hyperrealistic AI image generation was aimed firmly at the gaming industry.

And despite quickly expanding into other industries, this legacy is still present through its 1.8 million strong Discord server =- the third largest in the world.

This sense of community continues to permeate Leonardo’s direction and products – an advantage that the company’s co-founder and COO, Jachin Bhasme, says is unique in the space.

“The users are the ones that drive how the tech is used, which is something we’re really keen to lean into,” Bhasme said to SmartCompany.

Today, the platform serves multiple industries, including marketing, fashion, interior design, and architecture.

“Staying on the cutting edge and putting out the latest tools and being able to leverage them to define what they’re actually used for is super important.”

“Being able to lean on our community to define what these workflows look like gives us a unique place to stand in terms of still staying right on the cutting edge, while making it useful in a commercial context,” Bhasme remarked.

A more cautious approach for enterprise products

While Leonardo.Ai began in a section of the generative AI space that is fast, loose and always changing – something it is celebrated for – it had to appreciate the more cautious approach required for your platform to be adopted by the brand-safe suits.

“If something comes out in the week, we’re finding ways to integrate it into the platform that same week or the next day. But there’s just so much more required on the enterprise side, so many more stakeholders,” Bhasme said.

“We quickly realised that for this enterprise stream, we couldn’t just operate in the same way we have been with our core product.”

Leonardo.Ai’s enterprise solutions include collaborative tools and dedicated infrastructure aimed at ensuring reliability and scalability. These features are designed to meet the stringent requirements of businesses while also providing integration into existing workflows.

“One of the things we’re really excited about with Teams is that we’ll be rolling out a fully licensed model with attribution that’s collectively safe to use. So I think that’s going to be a major point of difference between our enterprise offering and consumer offering,” Bhasme said.

Safety is, of course, crucial for business use of AI, particularly as platforms are still finding their feet. The past 18 months have been awash with generative AI horror stories from the majority of big tech companies.

Leonardo hasn’t escaped unscathed, with the platform being used to create deepfake celebrity porn earlier this year.

“This has been an ongoing challenge from the beginning for any company in the AI space,” Bhasme said.

“From the start, we’ve had an ongoing project in perpetuity for constantly improving and evolving prevention detection methods and moderation in general.”

According to Bhasme, this has been strengthened further in the wake of Leonardo’s own user misuse issues.

It has also been stringent about prioritising ethical data sourcing and licensed model training for enterprise use. While unpredictable AI models can be fun and even helpful in some cases (OpenAI CEO Sam Altman even claimed that AI hallucinations are valuable), no company wants to take that gamble when their reputations are on the line.

“These large general models that apply to anything and everything were really useful early on, but we see a lot of value in customising those models – seeing what they look like for specific use cases,” Bhasme stated.

This approach not only mitigates legal risks but also sets Leonardo.Ai apart from competitors who may rely on less transparent data sources.

“It’s all about sort of having more predictability in what the outcomes can be, I think that’s super important for enterprise. Lower predictability is useful and interesting when you’re really in the weeds and understand the tech but we’re quickly realizing that with teams that will be using these sorts of products – not everyone has been on this two-year journey. A lot of people are starting today.”

The other and perhaps most significant point of difference for Leonardo.Ai is that it’s a completely new platform.

It’s been built as a generative AI-focused company from the ground up. Comparatively, many competitors in the market are tech, cloud or SaaS giants retrofitting generative AI into their product offerings.

“A differentiator for us compared to some others is that we’re not building our tools into legacy tech. We have this opportunity to really define what these workflows can look like,” Bhasme said.

Never miss a story: sign up to SmartCompany’s free daily newsletter and find our best stories on LinkedIn.