OpenAI launches ChatGPT Enterprise

OpenAI CEO Sam Altman speaks during a keynote announcing the ChatGPT integration of Bing at Microsoft in Redmond, Washington, February. 7, 2023.

Jason Redmond | AFP | Getty Images

On Monday, OpenAI announced its biggest news since ChatGPT’s debut: it’s launching ChatGPT Enterprise, its business class for AI chatbot, available starting Monday.

OpenAI’s COO Brad Lightcap told CNBC that the tool has been in development for “less than a year” and has had help from more than 20 companies of various sizes and industries. ChatGPT Enterprise includes GPT-4 access with no usage limits, up to 2x faster performance than previous versions, and API credits. Lightcap said the pricing will not be publicly announced and will “depend, for us, on the use cases and the size of each company.” Among the beta users are Block, Canva, and The Estée Lauder Cos.

Earlier this year, Microsoft’s expanded investment in OpenAI — an additional $10 billion — made it the biggest investment in AI of the year, according to PitchBook, and in April the startup reportedly closed a project. Sell ​​shares worth $300 million With a valuation of $27 billion to $29 billion, with investments from companies such as Sequoia Capital and Andreessen Horowitz. Two months after ChatGPT launched in November, the number of monthly active users has surpassed 100 million, breaking records for the fastest-growing consumer app in history: “Huge uptake – we’ve honestly never seen anything like it before, and interest has only grown since then,” Brian Burke stated, Vice President of Research at Gartner, to CNBC in May.

More than 80% of Fortune 500 companies have teams that actively use ChatGPT, according to Lightcap and OpenAI.

One key difference between ChatGPT Enterprise and the consumer-facing version: ChatGPT Enterprise will allow customers to enter company data to train and customize ChatGPT for their industries and use cases, though some of those features aren’t yet available in their debut Monday. The company also plans to offer another level of use, called ChatGPT Business, to smaller teams, but hasn’t set a timeline.

Lightcap told CNBC that rolling out the enterprise version first, and waiting on a business level, “gives us a little bit more of a way to approach the teams in a hands-on way and understand what the publishing traffic looks like before we open it up completely.”

“We don’t train on your business data or conversations, and our models don’t learn from your usage,” OpenAI noted in a blog post, adding that customer conversation data will be encrypted in transit and at rest. However, the company does record aggregate data about how the tool is used, including performance metadata and more, as is relatively typical, Lightcap said.

The debut of ChatGPT Enterprise comes as the AI ​​arms race continues to heat up between chatbot leaders such as OpenAI, Microsoft, Google, and Anthropic. In an effort to encourage consumers to embrace generative AI in their daily lives, tech giants are racing to launch not only new chatbot applications, but new features as well. In May, OpenAI launched its iOS app, followed by an Android app in July. Google regularly rolls out updates to its Bard chatbot, and Microsoft does the same with Bing, offering features like visual search. Anthropic, an AI startup founded by former OpenAI executives, launched its new AI chatbot, Claude 2, in July, months after raising $750m over two rounds of funding.

Running ChatGPT, like many large language paradigms, is expensive, with each OpenAI conversation potentially costing “single-digit cents,” according to December tweet by CEO Sam Altman, suggesting that running the service for 100 million people per month could cost millions of dollars.

Lightcap told CNBC that the biggest hurdle to developing ChatGPT Enterprise was figuring out how to prioritize features.

Of all the things that will ship in the next couple of months, he said, “prioritizing how to move these things forward based on how people use the product — what people really want and what enables them — has been the subject of a lot of discussion, I would say, around the team.” .”

One concrete example is Code Interpreter, a ChatGPT Plus feature that has since been renamed Advanced Data Analysis. Lightcap said the team questioned whether the feature was a priority for ChatGPT Enterprise and that it “ranks high on a list of a bunch of other things that we think are equal or kind of more exciting”, but the companies’ feedback caused that. To prioritize submission sooner rather than later.

OpenAI plans to include “as many organizations as possible over the next few weeks,” according to a company blog post.