OpenAI hopes ChatGPT Enterprise will answer employers’ data privacy concerns

This story was first published by Digiday sibling WorkLife

Corporate concerns around OpenAI’s tools have consistently revolved around data and privacy. 

But at the end of August, OpenAI announced ChatGPT Enterprise, a new tool to give companies features to protect proprietary data, which some employers say could be a gamechanger for those concerned about privacy protection. It’s now available to companies that are interested in buying a minimum of 150 seats.

ChatGPT Enterprise claims to offer enterprise-grade security and privacy that allows employers to own and control their business data. OpenAI is clear — it will not train on your business data or conversations and its models don’t learn from the company’s usage either. All conversations are encrypted.

Despite privacy concerns, ChatGPT has been adopted by over 80% of Fortune 500 companies since its launch just nine months ago, according to OpenAI. 

“ChatGPT Enterprise is a catalyst that is going to speed up decision-making and innovation for Fortune 500 companies by solving the problem of security first, allowing access to AI tools in a managed way,” said Raphael Ouzan, CEO of A.Team, an AI-enabled team formation platform. “With 80% of Fortune 500 teams already on ChatGPT, the enterprise version erases the previous data risks. It’s not just automating tasks or generating code, it’s a customized AI that can dive into your specific data and offer actionable insights, acting as a force multiplier.”

Ouzan says that ChatGPT Enterprise will be best for companies that are already data-rich and are looking to leverage AI to gain a competitive advantage.

Employers’ stances on ChatGPT (a free tool) and ChatGPT Plus (its souped up, paid tool) are night and day. When plus was released in February, some employers decided to offer the paid subscription as a work benefit, including automation platform Zapier and digital marketing agency Intero Digital.

At Zapier, 200 of their 700 employees comped their ChatGPT Plus subscription as did 100 of the 400 team members at Intero Digital. But both companies told WorkLife that the paid version was difficult to navigate as it led to inconsistency with not everyone on the platform and it didn’t guarantee that employees weren’t sharing sensitive information.

ChatGPT Enterprise establishes and enforces data privacy rules.

“You gave them the policies and then you had to trust them that they were going to do the right thing, like so many things in business. But now with enterprise, we know the data policy that the product has. It’s not up to the individual users.”

Danny Shepherd, CEO of Intero Digital, says the enterprise version has been “game changing.”

“What we couldn’t do in GPT up until this point is we could not put proprietary data, data we didn’t want them to train, or data that we didn’t want to become publicly available,” said Shepherd. “We’d have to make stuff anonymous, which was time consuming. Some things we weren’t even comfortable with making anonymous because what if somebody somehow figured out these relationships. There was a lot of stuff we simply could not put in the platform. I am excited to be able to put everything in now.”

Aside from this, it’s also a more powerful version of ChatGPT. OpenAI says this version removes all usage caps and performs two times faster.

“Even with ChatGPT Plus, you’d go in and prompt it and you’d see a little spinning wheel and be like ‘alright, I need to wait for someone else to finish their query so I can get this data,’” said Shepherd. “We’re not seeing that with enterprise. It’s just like bam, done. And with no more usage caps, it’s going to make for some really meaningful conversations and analysis that otherwise you just couldn’t do in the past.”

And there are shared templates that can be made to collaborate and build common workflows, which is especially helpful for someone who might still be struggling with prompting. Companies made prompt libraries before, but having it all in one place is helpful.

“Having templates is also one of the bigger game changers because people really struggle making prompts for ChatGPT,” said Reid Robinson, lead product manager of AI at Zapier, a company that OpenAI arranged an interview with. “When you see an actual prompt that has strong value to it, it’s pretty mind blowing.”

“Even with ChatGPT Plus, you’d go in and prompt it and you’d see a little spinning wheel and be like ‘alright, I need to wait for someone else to finish their query so I can get this data,. We’re not seeing that with enterprise. It’s just like bam, done.”

On top of that, though, the enterprise version also provides analytics on who in the company is using it and how they are using it. It’s a helpful feature for companies to begin tracking and strategizing how to take advantage of it. But Robinson admits it doesn’t have the best information just yet.

“The analytics are not the greatest,” said Robinson. “It’s certainly an area we hope to see improvements on. What we get to see is how many active users there are, how many messages were sent in the last week, how many were sent in the last 30 days, but there isn’t a lot of breakdown of anything beyond, like who are your most active users and the most-used templates.”

ChatGPT Enterprise is available for companies that will buy a minimum of 150 seats at varying pricing tiers. They have not disclosed the cost as it “varies depending on the needs of each company,” according to OpenAI. Intero Digital pays $60 per user, with a 12 month commitment paid up front. 

“Our team loves it,” said Shepherd. “They are power users and we’re seeing it with client output. The results that clients are getting from the team are better because they’re able to do more, test more and make smarter decisions. It’s been an incredible transformation. This technology is changing so fast … we’re excited to see how it plays out.”

https://digiday.com/?p=518080

More in Media

Earnings from social and search players signal that AI will be a long-play investment

Giants like Google, Meta and Microsoft say investors and advertisers might have to wait longer for AI to generate a better return on investment.

Why some publishers aren’t ready to monetize generative AI chatbots with ads yet

Monetization of generative AI chatbot experiences is slow going. Some publishing execs said they’re not ready to add advertising to these products until they scale or can build a subscription model first.

Media Briefing: Publishers who bet on events and franchises this year are reaping the rewards

Tentpole events and franchises are helping publishers lock in advertising revenue.