Why PwC is becoming ChatGPT Enterprise’s biggest user

When PwC’s Bret Greenstein is on a work call with someone in, say, Oklahoma, he might tap a custom GPT he built to grab local weather or sports scores to help liven up small talk. He sometimes bats around ideas with a specialized “research assistant” bot, to which he regularly feeds data. And as he spoke with Tech Brew, Microsoft Copilot dutifully took notes for the consultancy’s PR rep.

“These are little day-to-day things, but they make you so fast in what you’re doing,” said Greenstein, a generative AI leader at the firm. “I can take a snapshot of my calendar and just ask me for all conflicts. And it tells me what the meetings were about.”

Few companies outside of Silicon Valley have seemed to embrace the latest generation of AI with quite as much zeal as PwC and other consultancies. Last April, the company pledged to spend $1 billion to grow its generative AI offerings over three years. As of this week, it has widened its deal with OpenAI to make PwC’s US and UK firms the first reseller of ChatGPT Enterprise and the largest user of the product by number of licenses, according to an announcement.

Bot flippers: The partnership will let PwC offer ChatGPT Enterprise along with tailored solutions for industries and tasks ranging from software development to handling purchase orders or summarizing leases, according to Greenstein. Data can then be plugged in to customize them further, he said.

“Our enterprise clients are very focused on outcomes,” Greenstein said. “They want to know, ‘If I do this, how much better will this process be? How much faster will it be? How much more efficient will it be?’ And so the easier we can make that for them, the better.”

So where are these “outcomes” happening most? Greenstein said clients with a “high volume of documents coming in or out” stand to benefit from generative AI.

“For example, contracts, invoices, security logs, software development—these are lots of very heavy document-intensive processes that we can put AI in front of, and it does the analysis, the summarization, the triage, even recommended answers,” Greenstein said. “Up until now, the only way was to have people read all that stuff and summarize it. Now we can have AI do a lot of that work, working with people to give them better insights and better information.”

Leading by example: As Greenstein’s own AI power usage makes clear, though, PwC’s big AI push isn’t just about selling this tech to clients. The firm is also committing to teaching its thousands of employees how to use generative AI and building it into internal operations.

As one might expect, this hasn’t come without some fear and uncertainty among workers, Greenstein said.

“This is the largest change management in our lifetime. Getting people used to the internet was nothing like this—it took a decade to be interesting for people, it didn’t change everything people do,” Greenstein said. “This changes every role.”

Greenstein said PwC was able to push past some of those fears by treating AI as a practical tool—something that’s “productive, but not perfect.” It has held internal hackathons and “enablement sessions” to collect around 3,000 potential use cases, and it has a community of around 9,000 users sharing AI ideas among one another, he said.

“Making this accessible to people where they work is the most interesting and important thing because the best ideas are from people who are just discovering this now,” Greenstein said. “The people closest to work are going to have the best ideas for this.”