Main content

Salesforce World Tour London - reflecting a reality check for the AI hype cycle?

Stuart Lauchlan Profile picture for user slauchlan June 10, 2024
The AI hype cycle hasn't gone away, but there are more pragmatic commentaries and practices coming out of vendors.

reality check

At Salesforce’s World Tour London event in 2023, UKI CEO Zahra Bahrololoumi made the point that CEOs were excited by the potential of generative AI, but only up to a point, saying:

I talk to so many leaders and they're all so excited. They're excited. They can see the benefits. They can see the horizon. They can see the future. But guess what? They're cautious. Very cautious.

There are many reasons for that, not least concerns around trust. Flash forward a year, has much changed? On the face of it, perhaps not much. As Bahrololoumi noted at this year’s World Tour:

Every CEO I meet is realistic about the need to invest in AI to remain competitive. They are excited, they get it - they see the benefit, they see the productivity gains. They see it as the tool to help them drive growth. And yet understandably, they're concerned. They remain concerned about fragmentation, scalability, security, safety, hallucination bias. it's a concern. ‘We're not going to adopt AI unless we can trust it'. Our customers won't adopt AI unless they can trust it.

That prompted thoughts of comments  made by Tom Siebel, founder and CEO of C3 AI, at the end of last month when he said of current AI hype:

Today, many companies are dabbling in trivial AI projects or relying on outside integrators to try to cobble together something that works. These are nothing more than large and expensive experiments. Nobody succeeds.


That said, there are more examples of bleeding edge implementations of generative AI in use - Aston Martin’s successes in this respect was a running theme of this year’s World Tour London - but the same issues and concerns remain, added to by macro-economic pressures extending buying cycles and leaving the C-Suite hesitant to hit go on new projects.

It is also arguable that there is a new pragmatism in the air. It’s essential to keep in mind that the gen AI hype cycle accelerated at an incredibly fast rate. Back in March, Salesforce CFO Amy Weaver made the point that:

If you look back at our earnings call from March of last year, even with [our] background in AI, we only mentioned AI once on the call…if you fast forward 90 days later, we mentioned on our earnings call something related to AI 53 times in one hour.

And bear in mind, Salesforce has been one of the more realistic and pragmatic evangelists for gen AI.

That said, other vendors have been issuing more grounded statements and predictions of late. Meta CEO Mark Zuckerberg recently warned that it will take many years for his firm’s AI business to show a profit, a flash of candor that earned the company’s share price a harsh spanking from Wall Street. But Meta’s not alone here. In September last year, Accenture’s 2023 was revealed to be around $300 million. That’s out of total revenue of $64.1 billion. By March this year, gen AI bookings over the prior six months had risen to over $1 billion, but there’s a very long way to go before gen AI shows on the bottom line.

Another commentary of note came from Vinay Utham, CFA, Associate Professor in Finance and Corporate Governance at Brunel University London, who starkly opined:

AI adoption among enterprise customers is yet to fully take off, as businesses figure out how to make the best of AI and how to reconfigure their existing processes so that they maximize the potential of AI. For instance, based on a recent survey by Microsoft and LinkedIn, three in four white-collar workers currently use AI for their work. On the contrary, another survey by Ramp, according to the Wall Street Journal, found that only about a third of companies pay for at least one AI tool. While this represents a significant jump from a year, when that figure was only 21%, there is a massive discrepancy between the proportion of individuals using AI and the proportion of businesses willing to adopt AI. Put simply, AI adoption among enterprises has simply not taken off yet.

Hype down? 

I asked Salesforce CMO Ariel Kelman if the dial on AI hype had shifted since I last spoke to him at last year’s Dreamforce conference. He told me:

I think we've definitely gone through a cycle. It's always interesting to me how whenever new technology comes out, we go through the same cycle. Everyone acts like it's going to change the world and solve every problem. People get frustrated. They realize it's a lot harder than they thought. Then there's the people who are just practical. They're not trying to make announcements. They're not trying to make themselves heroes. They're just getting to work doing experiments, and then you find out what really works.

Salesforce has taken a very specific approach, he argued:

We've been on the more practical side of AI. We're not making a consumer AI. We're not trying to do a lot of lofty goals. We're working to develop solutions that help employees and our customers be more productive and be more successful with their customers. And so, by narrowing that scope into how can we deliver employee productivity improvements in the world of sales, marketing and support, [we are] giving our customers the ability to experiment with generative AI for a tiny amount of money. With our approach, we're not forcing companies to spend tens and tens of millions of dollars to be able to train custom models...We have an architecture that lets you ground AI in your customer data in a secure way, which really allows these experiments to flourish. 

It's a bit more of a practical approach.  A year ago people were eager to see what can work. Now these these experiments have shown fruition and our customers are seeing benefits and it's accelerating. But it's still very, very early for a lot of these companies who are trying to figure out, 'What do we use AI for?', and then to spin up experiments at all those levels. That's what we're trying to help them with.

My take

I have to tell you, Salesforce was absolutely built for this moment. It feels like everything we've done over the last 25 years has brought us to this moment because Salesforce AI is powered by trusted enterprise data, with safeguards and frameworks in place to ensure responsible AI development.

It’s an assertion from Bahrololoumi that we’re going to hear a lot as we proceed towards this year’s Dreamforce.

Salesforce is, in common with many of its peers, conscious that it is playing a long game here and setting expectations accordingly. The instinctive negative short termism of Wall Street has been made worse by the overheated AI hype cycle and recent months have seen a painful recognition of realities here. Salesforce itself took a beating following its own recent Q1 numbers, as did Workday shortly before on the back of conservative outlook. There will be more examples to come unfortunately as that hype cycle may become less feverish, but it’s not going away.

In the meantime, initiatives such as Salesforce’s new AI Center in London are to be welcomed, as is the emergence into the public domain of more practical use case exemplars. Show, not tell is what’s needed now.

A grey colored placeholder image