Enterprises Fully Embrace AI, Whether Prepared or Not

This week marked a pivotal moment for AI companies landing enterprise contracts, highlighted by Zendesk’s launch of AI agents aimed at handling 80% of customer inquiries. In other news, Anthropic and IBM have initiated a strategic partnership, while Deloitte has joined forces with Anthropic as well. Additionally, Google has unveiled a novel platform dedicated to AI in business contexts.

However, the path to AI adoption for large organizations may have hurdles. Deloitte’s announcement came just as the Australia Department of Employment and Workplace Relations indicated that the firm must refund for a report due to alleged AI-generated errors.

In the latest episode of the Equity podcast, Kirsten Korosec, Sean O’Kane, and I examined the latest developments in AI, contrasting them with the previous week’s discussion on the Sora app. Though monetizing consumer-facing social networking applications may take time, enterprise contracts present a quicker route to significant revenue.

Here’s a summarized preview of our conversation, edited for conciseness and clarity.

Anthony: This ties back to our previous discussion about emerging GenAI social networks. We considered them a potential revenue source for AI companies, which I still believe holds true, but the journey ahead is long. While enterprise solutions might lack the excitement of consumer products, they offer genuine profitability.

Sora could represent OpenAI’s revenue model in five years, but enterprise contracts are the current path to monetization.

The situation with Deloitte was particularly enlightening. While it may seem redundant to point out that these models aren’t yet ready for the industry, it’s encouraging to see the Australian government demand accountability.

This isn’t to suggest that AI should never contribute to drafting reports like these; one could argue otherwise. However, if employed, there must be accountability for the outcomes, and thorough verification of the referenced information is essential. You can’t merely input data into a model and consider your job finished, expecting to bill for hours. Anyone doing this should feel ashamed and face consequences.

Kirsten: Absolutely. Sean, Zendesk has also been in the spotlight as they are creating tools to completely automate customer service, thereby reducing the need for human involvement. In your everyday experiences or regarding auto service, are you noticing this trend of automation becoming more widespread?

Sean: Yes, I’ve covered this topic multiple times. Many startups are developing comprehensive customer service systems, voice agents, and large language models for communication through emails and texts with dealerships and service centers. I see huge potential here, as the main issue isn’t a lack of workers, but rather the challenge of reaching someone without being endlessly transferred.

Customers often find themselves shuffled among different departments in service scenarios. If we can make these interactions more efficient and enhance response times, the key question will be how willing businesses are to adopt and sustain such technologies. Many earlier initiatives, like web forms, were launched by dealerships but subsequently fell into disuse, becoming ineffective rather than serving their intended purpose.

I remain optimistic that advancements like these will become the primary point of contact for consumers. We’re on the verge of discovering just how effective they can be.

Equity is TechCrunch’s flagship podcast, produced by Theresa Loconsolo, with episodes available every Wednesday and Friday.

Subscribe to us on Apple Podcasts, Overcast, Spotify, and all the podcast platforms. Follow Equity on X and Threads, at @EquityPod.