Getty Images
Box AI adds Microsoft Azure OpenAI Service integration
Box adds Microsoft Azure OpenAI Service to its lineup of AI tools for document summaries, joining Google's Vertex and OpenAI LLMs for users to choose from.
Box Inc. today rolled out an integration with Microsoft OpenAI Azure Service to its platform. Features include the ability to generate summaries, answers to user queries, blog posts and talk tracks from a given document.
The enterprise document management platform already has integrations with OpenAI and Google Vertex AI, rolled out as part of its Box AI initiative begun last year. Providing access to different large language models (LLMs) is the company's way of keeping its platform open and giving users the ultimate choice, Box CTO Ben Kus said.
"Many of the big AI vendors have technology that resemble each other, many of them are now hosted by a trustworthy company -- Azure OpenAI is a good example -- and then on top of that, they all have these models that are pretty capable," Kus said.
During beta testing and the rollout of Box AI, Kus observed that customers choose LLMs based on performance for a particular job, he said. Some might find that one LLM is slower in returning results for their particular uses. One model might find better results than others in their enterprise's content lake. Another might be better at communicating results in its summaries.
Offering users access to different LLMs is a way to rise above the AI market's competitive fray and let the market decide which is preferred, Deep Analysis founder Alan Pelz-Sharpe said. But in reality, many organizations standardize on or trust the privacy and security of one vendor over another.
![Example screenshot of Box AI's integration with Azure OpenAI Service.](https://cdn.ttgtmedia.com/rms/onlineimages/screenshot-box_ai-f_half_column_mobile.jpg)
In other cases, enterprise users might bring their own model, customized to their particular industry's terminology, especially for customer-facing uses such as service. While bring-your-own is more the exception than the rule now, that could be where things are headed anyway, he said.
"For enterprise software, be it in legal, healthcare or whatever vertical, the future is almost certainly small language models, highly curated, presumably sitting in the enterprises' own repositories," Pelz-Sharpe said. "But that is not here yet, so at the moment, why don't you just plug into everything, and we play with it."
GenAI wish list: Clean dirty data
Box AI, including the integration with Azure OpenAI Service, is generally available today. It is included in all Enterprise Plus plans. Individual users have access to 20 queries per month, with 2,000 additional queries available on a company level.
Seeking generative AI cost relief on behalf of Box customers, the company's co-founder and CEO Aaron Levie openly lobbied his longtime Silicon Valley friend and OpenAI CEO Sam Altman to lower prices on LLM usage at the BoxWorks user conference last fall.
Pelz-Sharpe termed the AI pricing model "clever and realistic," as it shows customers that they shouldn't use generative AI for everything and get hit with surprise bills. AI, he said, is very good for certain jobs, but it's expensive.
End users have been served many generative AI tools, such as Box AI and Microsoft Copilot. Pelz-Sharpe said it's time for AI vendors to focus more on tools for IT staff and content managers.
"All of this GenAI stuff coming out is really focused on the desktop -- how do I collaborate -- and that's cool," Pelz-Sharpe said. "I'm hoping over the next year or so we're going to see a lot more on the back end, such as AI and machine learning to improve data quality and clean it up. That's the problem with any API -- garbage in, garbage out. And that's the state of most people's repositories -- garbage."
Don Fluckinger is a senior news writer for TechTarget Editorial. He covers customer experience, digital experience management and end-user computing. Got a tip? Email him.