Agentic AI Has Proven Its ROI. The Next Frontier Is Beyond the Contact Center.
Reading the latest Google Cloud × National Research Group report (3,466 executives, including 179 in France) through the eyes of an AR & AI agency.
The debate is settled. Executives are no longer asking whether AI improves customer experience — they are asking how to scale it. The latest Google Cloud report, conducted with National Research Group across 3,466 senior executives worldwide between April and June 2025, confirms this with numbers that are hard to dismiss: 88% of early agentic AI adopters now report positive ROI. On the ground, 37% of executives already claim ROI on their generative AI use cases for customer experience — a figure that climbs to 55% among first movers.
At ARGO, we read this report carefully. But what strikes us most is what almost everyone will draw the same conclusion from it.
The Trap: Everyone Is Looking in the Same Place
The reflexive conclusion is: “we need an AI agent on our customer service.” And it is not wrong. The report shows that conversational AI in the contact centre works, sometimes spectacularly: German bank Commerzbank handles more than two million conversations with its Bene agent and resolves 70% of them autonomously; Mr. Cooper gained 28,000 hours per year by improving average handling time by 3.5% across 500,000 monthly calls; Mercari is targeting 500% ROI by reducing its advisors’ workload by at least 20%.
The problem is not that it does not work. The problem is that it works for everyone. AI in the contact centre is now a standard component of the customer experience stack: 63% of companies already use it. In other words, it is becoming the entry ticket — not the competitive advantage. The ROI is real, but increasingly expected — and therefore decreasingly differentiated.
The report says it in its own way: first movers do not win because they deployed a chatbot somewhere. They win because they have a deliberate strategy — 82% of them have deployed more than 10 agents (versus 39% of the overall sample), and they deploy AI well beyond the after-sales function.
The Real Frontier: Beyond the Contact Centre
This is the section of the report that few people will comment on — and it is the most strategically important one. Under the heading “An enhanced customer experience beyond the contact centre”, Google Cloud observes that 54% of executives already use generative AI for digital commerce and optimised experiences — on mobile, on kiosks, on the web. Companies are meeting their customers where they actually are: in-store, in their vehicle, on their smartphone.
And the mode of interaction is changing in nature. We are no longer talking about text typed into a chat window. The report describes agents capable of understanding text, voice, image, and video — grounded in company data — that guide the customer from initial inspiration through to final purchase. Keyword search gives way to complex natural language questions, which the agent answers by surfacing the right product, the right information, the right action.
This is where, in our view, the next wave of ROI lies. Not in yet another contact centre improvement, but at the point where the digital meets the physical world.
Our Conviction: The Next AI Agent Is One That Sees
This is precisely what ARGO does. We design experiences where computer vision and image recognition meet agentic AI and augmented reality. Concretely, this means a customer can:
- point their phone at a product on a shelf and receive personalised conversational assistance — availability, alternatives, usage advice — grounded in your data;
- scan a location, an artwork, a storefront, or a piece of equipment and trigger a contextual information layer, in AR, powered by an agent;
- navigate a complex discovery journey (a retail space, a tourist site, an industrial environment) guided by a multimodal assistant rather than a static map.
This is exactly the report’s promise — multimodality, data grounding, fluidity from inspiration to action — but moved from the virtual counter to the real point of contact. And that is where an experience becomes memorable, therefore differentiating, therefore loyalty-building.
One important clarification: the report insists on a “human + machine” strategy. The best deployments do not replace humans — they create augmented “super-agents”. We share this conviction without reservation. AI surfaces the right content at the right moment; the human retains ownership of the relationship.
Good News: You Don’t Need Google’s Scale
This may be the most liberating message in the entire study. ROI does not depend on budget size. It depends on three things the report documents precisely.
A well-chosen use case. High-performing companies focus on the single point of contact that has the most impact for their customer base — not ten initiatives at once.
Executive buy-in. 78% of companies fully aligned at leadership level report ROI — a marked gap versus those that are not.
A clean, secure data foundation. When executives choose a model provider, security and data privacy top their criteria (37%), ahead of integration and cost. An agent is only useful if it has secure, controlled access to your systems.
Nothing here requires a multinational. It requires a partner who masters both the technology (vision, models, agents, integration) and experience design. That is our positioning.
In Practice
The report closes with a checklist. Ours fits in one sentence: identify the point of contact, outside your contact centre, where an augmented experience would change the game for your customers — and let’s prototype it.
At ARGO, we do not simply advise on agentic AI: we build these systems, and we deploy them internally ourselves. If reading this report has made you want to move beyond the chatbot, let’s talk.
Source: The Future of Customer Experience: AI and Agents Transforming Interactions, Google Cloud × National Research Group, 2025 — 3,466 executives surveyed across 8 countries, April–June 2025.