This week has been a major one for enterprise AI, with several industry giants announcing significant deals and product launches. Zendesk introduced new AI agents that it claims can resolve up to 80% of customer service inquiries. Meanwhile, Anthropic entered a strategic partnership with IBM and also struck a deal with Deloitte. Google joined the mix as well, unveiling an AI-for-business platform aimed at helping companies integrate generative AI into their workflows.
Despite this rapid momentum, not every corporate AI story came with good news. Deloitte’s partnership announcement collided awkwardly with a report that the company must refund Australia’s Department of Employment and Workplace Relations. The reason? A departmental review found that a Deloitte-delivered report contained numerous AI-generated hallucinations, raising questions about quality control and accountability in enterprise AI use.
On the latest episode of the Equity podcast, hosts Kirsten Korosec, Sean O’Kane, and I explored the week’s biggest AI headlines and how they tie into the broader evolution of AI-driven business. While last week’s discussion focused on the new Sora app and the long-term potential of consumer-facing AI platforms, this week’s developments underscore where the real money is right now: the enterprise.
Anthony: Last week, we were talking about GenAI-focused social networks like Sora as a possible long-term revenue stream for AI companies. But when you look at what’s happening now, enterprise deals are the more immediate, lucrative path. Maybe Sora will be how OpenAI makes money five years from now, but partnerships like these are how AI companies are making money today.
The Deloitte situation really stood out to me. It’s a reminder that these models aren’t always ready for prime time, and it’s encouraging to see a government client pushing back. AI can absolutely play a role in report creation, but organizations still need to verify every output and take full responsibility for accuracy. You can’t just feed data into a model and bill for the result. Anyone doing that should be embarrassed—and fined.
Kirsten: Exactly. Sean, Zendesk’s announcement was another big one—its new AI agents are designed to automate most of customer service. Are you seeing that kind of automation starting to show up in your day-to-day experiences?
Sean: Yeah, absolutely. I’ve covered a number of startups building full-service AI customer support systems—voice agents, chatbots, and LLM-powered email responses for everything from dealerships to service centers. The goal isn’t to replace people but to fix a long-standing problem: customers can’t get prompt, consistent responses.
For instance, when you call a dealership, you often get transferred endlessly or told someone will call you back. If AI tools can streamline that experience and make sure people actually get answers, that’s a real improvement. The question is whether businesses will consistently use and maintain these systems. We’ve seen past technologies—like web forms—get rolled out and forgotten. But this new wave of automation could finally become customers’ first reliable touchpoint with a business. And it looks like we’re about to find out.
