Like it or not, corporations are wagering on AI.

AI companies have had a significant week, with Zendesk introducing new AI agents designed to handle 80% of customer service inquiries, Anthropic and IBM declaring a strategic alliance, and Deloitte revealing an agreement with Anthropic. Furthermore, Google unveiled a new AI platform for businesses.

However, large organizations may not find AI implementation to be entirely seamless. The timing of the Deloitte announcement was somewhat unfortunate, as it coincided with the Australia Department of Employment and Workplace Relations stating that the professional services and consulting firm would need to issue a refund for a report submitted to the department that contained what seemed to be several AI-generated fabrications.

On the most recent episode of the Equity podcast, Kirsten Korosec, Sean O’Kane, and I talked about the latest AI news, comparing it to the prior week’s coverage of the new Sora app. While AI firms may eventually profit handsomely from consumer social networking apps, enterprise deals offer a more direct route to substantial earnings.

A preview of our conversation, modified for brevity and clarity, is available below.

Anthony: I believe this connects to our previous week’s discussion regarding GenAI social networks. We presented that as a possible route for AI businesses to eventually generate revenue, which I certainly believe is possible, but it will take time. And while the enterprise may not always be seen as as interesting or appealing as consumer [applications], it is truly where the money is. 

Sora may be how OpenAI generates revenue five years from now, but this is how these companies will generate revenue now.

And the Deloitte [news] was particularly noteworthy. It may seem repetitive to emphasize that these models [are not always] completely ready for widespread usage, but I am encouraged that the Australian government challenged this and stated, no, you cannot do this.

It is not necessarily that AI should never be used in the creation of these types of reports, though I believe a case could be made for that. However, if you plan to use it, you must be responsible for the results. You must carefully review and ensure that the information cited is accurate. You cannot simply feed data into a model and declare, “My work is done, the billable hours are complete.” Anyone who does so should be ashamed and penalized.

Kirsten: Exactly. Sean, Zendesk also made an announcement this week, and they are developing tools that will essentially handle all of customer service, thereby eliminating human involvement in the process. In your daily [life], whether it’s how you navigate the world or how automakers handle service, for example, are you seeing this type of [automation] becoming more prevalent?

Sean: Yes, I’ve written about it several times. There are numerous startups creating comprehensive customer service suites, voice agents, and LLMs for emails and texts [from] dealerships and service centers. I believe it’s a worthwhile concept because the issue isn’t that we don’t have enough people to perform these jobs, which would result in job losses. The problem is that it’s often impossible to reach someone by phone or you’re constantly transferred. 

Especially when seeking service, you are routed to the service department, where everyone is occupied. Therefore, if you can accurately capture the information and make it easier for people to receive a response, the question for me is how many businesses will embrace and maintain it. Over the years, there have been various technologies, such as web forms, that dealerships have implemented but subsequently neglected. They remain on their website, giving the impression that they work, but they don’t because the dealership prefers you to call them. 

So, I am cautiously optimistic that solutions like this will become people’s initial point of contact with [a business]. And it seems we are about to find out.

Equity is TechCrunch’s premier podcast, produced by Theresa Loconsolo, which releases episodes every Wednesday and Friday.

Subscribe to us on Apple Podcasts, Overcast, Spotify, and all podcast platforms. You can also follow Equity on X and Threads, at @EquityPod.