What Google Cloud Next Revealed About the Next Phase of Enterprise AI

By Kevin Meeks, Field President, LivePerson

Google Cloud Next has always been a place to see where enterprise technology is heading. This year, the signal was clear: AI is no longer being discussed as an experiment. It is being evaluated as an operating model.

Across keynotes, customer sessions, booth conversations, and meetings with our partners at Google, the discussion was less about whether AI can transform customer experience – that’s already established – and more about how enterprises can move from ambition to scaled execution.

That shift matters. For large organizations, AI success is not defined by the number of pilots launched or models tested. It is defined by whether AI can be deployed safely, integrated into existing operations, measured against real business outcomes, and trusted by customers, employees, legal teams, compliance leaders, and executives.

Here are a few takeaways that rose above the noise over a very noisy few days in Vegas: 

1. The enterprise AI conversation is moving from capability to execution

For the last several years, much of the AI conversation has focused on what the technology can do. At Google Cloud Next, the more urgent question was how organizations can operationalize it.

That theme came through clearly for Nathalia Saliba Dias, PhD , LivePerson’s CX Strategy Lead, who leads a team of experts advising large enterprises on customer experience strategies that connect business goals, operating models, technology investments, and measurable outcomes. A recognized CX strategist, Nathalia has a front-row view into what separates AI experimentation from enterprise-scale impact.

Reflecting on the event, Nathalia pointed to a consistent set of themes across keynotes and customer conversations: executive alignment, human-in-the-loop models, stronger integrations, reusable multi-agent infrastructure, and a sharper focus on defined use cases over generic AI capabilities.

One observation stood out in particular:

“Operating models now require builders, not just managers.”

That line captures an important shift. The organizations making progress are not simply experimenting with AI tools. They are building the connective tissue around AI: governance, integrations, operating models, measurement frameworks, and cross-functional alignment.

2. RCS is becoming a strategic channel for customer engagement

One of the most important conversations at Next centered on the continued momentum behind Google RCS.

For enterprise brands, RCS represents more than a richer messaging format. It creates an opportunity to rethink how businesses engage customers in the channels they already use, with more interactive, branded, and trusted experiences.

That matters in a world where customer engagement is increasingly fragmented. Brands need ways to create more useful, contextual, and secure digital conversations without forcing customers into disconnected experiences.

Conversations with our partners in Google’s RCS leadership reinforced how quickly this channel is advancing globally, and how much opportunity exists for brands that are ready to pair RCS with AI-powered orchestration, automation, and human support.

For LivePerson, this is an important area of focus because messaging has always been central to our view of customer engagement. As RCS matures, the opportunity is not just to send better messages. It is to create more intelligent, end-to-end customer journeys.

The results we are seeing with our brands are exciting, and the conversations at Next reinforced how much opportunity there is for our customers in this channel. 

3. AI trust is now an operational requirement

Another major theme was trust — not as a brand concept, but as an operational requirement.

David Serna Perez , a LivePerson account leader who works with large enterprise brands on AI and customer engagement strategy, summed up the message he heard from financial services leaders at Next:

“AI success in FSI isn’t about more models — it’s about secure foundations, clean data, clear ROI metrics, and controlled agent orchestration. The winners are not just building AI — they’re operationalizing it.”

That reflects what we hear from customers every day: enterprises want to move faster, but they need confidence that AI will behave as intended across real customer interactions, complex policies, regulated workflows, and edge cases.

This is where AI assurance becomes critical.

At LivePerson’s booth, leaders came to see Syntrix, our AI assurance platform, and talk through how to test, evaluate, and improve AI systems before they impact customers. We were glad to see that the interest came not only from CX leaders, but also from technical, product, and innovation teams focused on how AI is governed, monitored, and scaled.

This further showed us that assurance is moving from a niche concern to a core enterprise capability.

4. Human-in-the-loop is not a temporary bridge

There is a tendency to frame human involvement in AI as something that will eventually disappear. The conversations at Next suggested the opposite.

In the most mature enterprise environments, humans are not simply fallback support. They are part of how organizations accelerate learning, manage exceptions, improve outcomes, and build confidence in AI systems.

Human-in-the-loop models are especially important in CX, where intent, context, emotion, compliance, and brand trust all matter. The goal is not to choose between automation and people. The goal is to design systems where AI and human expertise work together intelligently.

That is also why the future of CX will not be defined by automation alone. It will be defined by orchestration: knowing when AI should act, when a human should step in, how context should transfer, and how every interaction should be measured and improved over time.

Turning the learnings from Next into an AI execution plan

For customer experience leaders, the path forward is becoming clearer: AI outcomes do not happen simply because a model is deployed.

They happen when AI is connected to the right channels, grounded in the right data, governed by the right controls, and continuously evaluated against real business and customer outcomes.

That is why the conversations at Google Cloud Next were so meaningful. Across discussions with Google Cloud teams, enterprise customers, prospects, developers, and CX leaders, the same operational questions kept coming up: 

• How do brands know when an AI agent is ready for production? 
• How do they validate performance before customers are exposed? 
• How do they satisfy legal, security, and governance stakeholders without slowing innovation? 
• And how do they monitor agentic systems once they are live?

These are the questions enterprise leaders are now working through as they move from pilots to production.

I look forward to tackling these big questions with our customers and partners as they move from AI pilots to trusted, production-ready customer engagement.