Tech giants Microsoft and Google have already introduced generative AI tools into their enterprise products, with Co-Pilot in Office and Teams, and Gemini in Google Workspaces. In both cases the AI can securely access work data, making them more specific than the general consumer models.

Bhasme said that, with Leonardo for Teams, the company had an opportunity to outmanoeuvre some of the bigger companies in the space.

“We’re a start-up, so we’re not necessarily trying to slap AI into legacy tech. This is completely revolutionary technology, and we have a chance to redefine what some of these workflows can potentially look like,” he said.

“It’s not one large general model that’s applicable to everyone, but we really see the value in personalised experiences and being able to fine-tune your models for specific purposes and curate workflows for particular verticals.”

Another thing Leonardo.Ai plans to leverage is the size and dedication of its fan base. The company’s Discord server, where users discuss AI image generation and how they use the tools, is the third most populated Discord server in the world.

The company runs regular competitions to drive engagement, and paying customers who create custom AI models with the tools can choose to share them with the entire community. Bhasme said this is a unique asset as while Leonardo.Ai supplies the cutting-edge technology, its users define the new use cases and curate hyper-specific solutions.

While the most common use for AI image generation in the workplace has been for creative firms like advertising and marketing, it’s also been gaining traction in other industries. Real estate can use it to populate a photo of a blank room with realistic furniture, for example, cutting down on staging costs. Moviemakers can use it for story boarding, eliminating the need to collect visual references from other sources.

Bhasme said that taking this work into an online collaborative space was the logical next step. Leonardo for Teams will let distant workers create, generate, review and edit artwork side-by-side, as well as share fine-tuned models among entire teams alongside organisation-level asset management.

“When you’re fine-tuning models on your own IP, and being able to share those models within a collaborative environment, having access to that and visibility over what goes into that, it’s super powerful,” he said.

Loading

Yet while the space is clearly growing, there’s also been a question of risk in enterprise AI – especially as regulation of the sector is largely yet to catch up to the tech. Some firms purely use image generation for internal or pre-production use, for example, as some AI-generated art has been subject to accusations of plagiarism if it has (or is perceived to have) turned out an image too similar to one it was trained on.

Bhasme was confident enterprise AI would weather future regulatory and public perception challenges.

“Obviously, it’s an evolving space and we’re constantly evolving with it,” he said. “As we look to the future, we think about things like fully licensed models with attribution. And that’s something that’s in the works for us at the moment.”

Get news and reviews on technology, gadgets and gaming in our Technology newsletter every Friday. Sign up here.