Get all your news in one place.
100’s of premium titles.
One app.
Start reading
Fortune
Fortune
John Kell

Autodesk's 'discovery mentality' gives managers permission to roll out AI experiments

(Credit: Autodesk)

Prakash Kota has worked at Autodesk for two decades, long enough to remember the days when the design software company sold DVDs to customers.

Back then, it took several weeks for a customer to reach out to a reseller, place an order via the phone or a fax machine, and then wait for the software to be shipped and installed. But today, orders are booked online, instantaneously of course, and Autodesk's business now greatly relies on subscriptions. Recurring revenue represents 98% of the company's $5.5 billion in annual revenue for the latest fiscal year.

A shorter sales cycle demands far speedier customer service. “It’s all about instant gratification,” says Kota, SVP and chief information officer at Autodesk since 2018. “And that's where my role came in.” 

Kota oversees traditional IT functions, including overseeing all companywide applications, vendor management, and investments in innovative technologies like artificial intelligence. But he also leads Autodesk's global customer operations support team, given the high priority of keeping the company's vast subscription-based clients satisfied.

As Autodesk’s subscription base rises, the company isn’t keen to proportionally add more customer service agents. That's to control costs, to be sure, but Autodesk says there's been an evolution in how the company thinks about providing the best customer service today. There are far fewer inbound phone calls and a lot more inbound customer inquiries have migrated to live chats. Customers also want to solve their own problems through self-help services whenever possible.

These shifts in behavior have spurred Autodesk to invest more in automation and generative AI capabilities, which the company uses to make its customer service agents more efficient. Earlier this year for example, Autodesk rolled out an AI assistant agent that relies on large language models from vendors like Microsoft and Salesforce. The AI system combs through Autodesk’s knowledge library and generates a quick response, which is then approved by a human agent before being shared with a customer. There's also the customer-facing Autodesk Assistant, a self-service interactive chatbot.

Another project spearheaded by Kota was on the backend, pulling together more than 20 different software applications that agents relied on and combining them into one master application. These changes have also helped speed up onboarding. With 40 years of history, it could take up to six months for Autodesk to train newly hired support agents. It now takes less than six weeks.

In total, Autodesk has 15 generative AI use cases in full production and 23 use cases that are in pilot internally. Around 50 more use cases are being explored today. “We want to lean in and experiment,” says Kota.

Other internal generative AI use cases that have been deployed include the rollout of Microsoft 365 Copilot for around 5,000 Autodesk employees, a copilot tool that's being leveraged by the sales team, and an internal chatbot called Lumi that leverages OpenAI’s large language models. 

While Autodesk does weigh the spending required to support new technologies like generative AI, Kota says he centers his thinking on an outcome-based approach. Some questions he asks himself include: What problem is Autodesk trying to solve? How can technology help? What vendor should Autodesk work with? How many employees can be helped if a new tool is deployed? And how much will the technology actually be adopted? 

“We’re a big believer that AI capabilities are going to augment and support humans,” says Kota.

He also has found AI "champions" across the organization, including in sales, marketing, and finance, to better understand what those leaders want to get out of AI.

There is strong buy-in from Autodesk's C-suite leaders, who are encouraging the workforce to lean into AI, but also not get too attached to any new solution. If a generative AI tool is deployed and it doesn't work or employees aren't using it, Autodesk will simply turn it off and move on. 

“They want to take some risks,” says Kota. “Because they know that people won’t get blamed if something doesn’t work. That discovery mentality is a huge thing.”

John Kell

Send thoughts or suggestions to CIO Intelligence here.

Sign up to read this article
Read news from 100’s of titles, curated specifically for you.
Already a member? Sign in here
Related Stories
Top stories on inkl right now
One subscription that gives you access to news from hundreds of sites
Already a member? Sign in here
Our Picks
Fourteen days free
Download the app
One app. One membership.
100+ trusted global sources.