Is There Value In A Curated Enterprise AI Experience?

Forbes

22-04-2025

Google Agentspace Ideation Agent
Google
Over the past couple of years, the frantic pace of AI innovation has had the big three cloud players vying to keep up with each other when it comes to AI capabilities. At this point, it's pretty easy to say that if one cloud vendor has a new AI capability, your preferred vendor either already has it or will have it within the next few weeks. One would think that AI presents new opportunities for a vendor to differentiate itself and take share from its competitors. But what's really happening is more a game of defense where the race for parity is about keeping existing customers.
Heading into the Google Cloud Next 25 conference a couple of weeks ago, I was interested to see how Google would differentiate its AI offerings. I had some hope based upon its recent announcements about Agentspace (which I covered here) and the Customer Experience Suite (which I covered here). Both of those were notable in that the messaging was less about technology and more about creating business value and changing how people work. There were two key takeaways from the event. First, market execution is as important as technology in an ultracompetitive market like AI. Second, Google is using some proven but uncommon methods to differentiate its innovations.
(Note: Google is an advisory client of my firm, Moor Insights & Strategy.)
Google Cloud has held the number-3 revenue position in cloud services for a while now, but over the past year it's been taking market share from its competitors. It's even feasible that Google Cloud could eventually take the number-2 position away from Microsoft Azure. Certainly some of that has to do with the technology, but I think that hiring a new go-to-market leadership team and investing more in training and certifications — a trusted blueprint borrowed from many other established tech players — also has a lot to do with it. The timing for these initiatives is good, given the business-value-driven product messaging.
Google has also made some smart decisions when it comes to building customer confidence, such as 30-day commits on spot pricing, along with making investments in customer education and services. Google is proving that how well you educate and take care of your customers is a major ingredient for tech sales and retention in times of disruption. This may not have always been the case for Google Cloud, but at Next 25, I became convinced that my perception of the company needed to be updated.
In terms of technology, Google, like all other AI vendors, delivered a full plate of new innovations at its marquee event. And there were a lot of aha moments, including the reverse engineering and re-release of The Wizard of Oz. But it wasn't the technology itself that was amazing; it was how the technology is designed, how users are engaged and how solutions are deployed.
That's not to say the technology isn't good. I would suggest that Google's core AI tech is competitive but only marginally differentiated. Rather, what I saw from Google was a deeper degree of business thinking and user-centric design than its AI competition. This is something I would call a curated experience. I believe this is deliberate, because a curated experience is a critical complement to Google's investments in improved market execution — not to mention how Google can gain new customers while retaining the ones it already has. To break this down a bit further, let's consider three big developer-related announcements.
Agentspace represented a well-thought-out user-centric design. Google has had some success when it comes to AI user experience with other technologies such as NotebookLM. But Agentspace is a new type of work interface. For starters, it's personalized based upon the user's individual profile or other contextual inputs. For example, maybe you are running agents for a specific task. In that case, Agentspace will have the ability to present other relevant agents and downplay those that are irrelevant. Also, the UI looks more like a consumer product than something you'd typically see in an enterprise. The Agentspace product management team shared that this was a deliberate choice, and that they collaborated with Google's consumer UI teams to do it. The rationale was that AI adopters tend to get their initial introduction to AI from consumer-oriented projects. Therefore, give the user something they can learn more naturally — based upon experience rather than technical standards.
To help drive further user adoption and engagement, Google also announced the Agent Developer Kit, which is an open source set of methods to foster collaboration between agents and other remote services. This makes a lot of sense because Google has leveraged open source to great effect in the past. The most notable example of this is Kubernetes, which is now the de facto standard in container management. Google's biggest contribution in the ADK was the Agent2Agent protocol, which is provided only with an AI platform like Google's own Vertex or Salesforce's Agentforce. The ADK also supports the emerging MCP standard. By open-sourcing ADK, Google will be able to attract developers to code collaborative agents without a lot of extra software and cost. (Why am I so sure? Look at the whole history of open source adoption in the enterprise.) It's a great way to get people to try out agents and, if they like it, to then consider Google's higher-end agent capabilities in Vertex and other products.
Finally, in terms of solution deployment, Google's roadmap and packaging is quite clear. For example, in tooling there's Vertex and Firebase for professional developers and Agentspace for the no-code development environment. Another example is Model Garden, which, with 200 models available, is not too constrained but also avoids the chaos of more than one million models on Hugging Face. Google's simplicity here is quite refreshing, especially compared to the other cloud providers, which have more complex and entitlement-driven solutions.
A curated experience is not a new idea. In fact, Red Hat Enterprise Linux was the poster child for the whole concept 20 years ago. By providing enterprises with gated access to the best of standards and open source and a world-class support team, Red Hat gave customers a feeling that they were tapping into both meaningful innovation and sensible risk mitigation. But we also have to be honest that RHEL was a lot cheaper than the Unix systems it was initially displacing. Eventually, as cloud came to the fore, open Linux also became more mainstream, and new competitors to Red Hat emerged (including cloud-specific variants from AWS and Google).
Fortunately for Red Hat, it was able to transfer the RHEL thought process to other areas like virtualization and DevOps. (In this context, it's worth noting that Kubernetes is core to Red Hat's OpenShift cloud management platform.) Based upon this example, one can assume that curation can be a meaningful value proposition — but it may not be sustainable.
Given the level of fragmentation, low standardization and user confusion about AI in the marketplace today, now is a good time for users to consider a curated experience. But how long it will last remains an open question. Additionally, Google's timeline may be different than Red Hat's, because Google is also tapping into something of a different positioning that may achieve a better result. Instead of providing 'leading-edge but not bleeding-edge' AI, the company is conveying a sense of 'Let's get going.' And for many mainstream companies and users, that may be the right type of encouragement to choose Google — particularly given how early we are in the age of AI.
So, is Google's new TPU 10x faster than two years ago? Is Gemini better at a given benchmark this week? Yes to both — and those tech milestones certainly have their importance. But Google's real bet is that your experience in learning about and using AI is more important than those types of headlines.