Google Cloud went large on its shift to agentic in its synthetic intelligence (AI) technique, the place it set out its imaginative and prescient to maneuver to agentic enterprise automation at its London Summit this week. That transition comes with a raft of instruments – largely introduced earlier this 12 months at Google Cloud Subsequent – designed to permit organisations to maneuver to multi-agent programs that may motive and execute complicated enterprise workflows autonomously.
Talking on the occasion at Tobacco Docks, Maureen Costello, vice-president for the UK, Eire, and Sub-Saharan Africa at Google Cloud (pictured above), advised attendees the period of experimentation is being supplanted by a drive for operational deployment.
In line with Costello, AI brokers obtained however one point out on the earlier 12 months’s occasion, however have now grow to be the central pillar of the supplier’s roadmap because it seeks to scale multi-agent enterprise programs.
“The agentic enterprise is going on proper right here, proper now,” she stated. “Main UK companies are transferring previous fundamental pilots. They’re deploying multi-agent programs – brokers that may suppose, motive, collaborate and execute complicated enterprise workflows.”
The slew of instruments fell into three core product pillars: frontier fashions, agent platforms and private productiveness instruments.
Firstly, there’s a refresh of the Gemini frontier mannequin household. Google confirmed that Gemini 3.5 Flash – a mannequin performance-optimised for brokers and coding duties – will likely be out there to UK clients by the tip of June. This will likely be adopted by Gemini 3.5 Professional later within the 12 months and Gemini Omni, a multi-modal mannequin designed for excessive velocity and low latency throughout textual content, audio and video inputs.
The supplier additionally launched a collection of agentic platforms geared toward builders and enterprise customers. The Gemini Enterprise Agent Platform is focused as a centralised mission management for managing autonomous workflows. It’s complemented by Agent Designer v2, a low-code device that makes use of pure language to construct brokers throughout disparate functions. For engineering groups, Antigravity 2.0 brings an agent-first code surroundings out there by way of desktop or command-line interface.
The agentic enterprise is going on proper right here, proper now,. Main UK companies are transferring previous fundamental pilots. They’re deploying multi-agent programs – brokers that may suppose, motive, collaborate and execute complicated enterprise workflows Maureen Costello, Google Cloud
The growth into the private productiveness house was marked by Gemini Spark, a 24/7 private assistant meant for enterprise laptops and cellular units. These instruments are underpinned by the Data Catalogue, which Google describes as a common context engine designed to index and categorise enterprise knowledge to make it accessible for AI reasoning.
In the meantime, on the {hardware} and safety entrance, the supplier unveiled its eighth-generation Tensor Processing Models (TPUs), customized AI {hardware} designed to enhance effectivity for coaching and inference. This was paired with the launch of Google AI Menace Defence, a cyber safety resolution that makes use of autonomous brokers to watch for and mitigate AI-powered threats in actual time.
How Unilever is utilizing the Gemini Enterprise Agent Platform
Through the keynote, Google showcased its strategic partnership with Unilever, signed earlier this 12 months. The patron items big is utilizing the Gemini Enterprise Agent Platform to drive model discovery and advertising. The deal is notable because it marks a big growth of Google’s footprint inside Unilever, and successfully supplanted legacy Microsoft-centric ecosystems for superior AI workloads.
We’re working with Google to go deeper on agentic commerce, constructing our guardrails and our observability layer as we embed AI throughout our world enterprise Sam Kini, Unilever
Sam Kini, chief digital and expertise officer at Unilever, defined that the corporate is systematically unpacking its core enterprise processes via an agentic lens. Kini famous that the trail to buy is basically altering, and requires organisations to grow to be “AI native” to remain related in rising commerce channels.
“We consider that having the suitable steadiness of tempo and management is vital,” stated Kini. “You need to be courageous and you must be ready to throw out the playbook. We’re working with Google to go deeper on agentic commerce, constructing our guardrails and our observability layer as we embed AI throughout our world enterprise.”
How Google Cloud is investing within the British financial system
The supplier additionally showcased its ongoing dedication to UK infrastructure, centred on the scheduled opening of a datacentre in Waltham Cross, Hertfordshire, later this 12 months. This venture varieties a part of a broader £5bn funding programme within the UK’s AI financial system, which additionally features a partnership to upskill 100,000 civil servants in AI applied sciences by 2030.
“Sovereignty isn’t only a coverage requirement. It requires bodily metal within the floor. Google Cloud is investing closely within the British financial system … guaranteeing that each British organisation has the instruments, expertise and expertise to relentlessly innovate,” stated Costello.
The ‘wait and see’ period of AI is over. If there was one phrase I may call to mind that basically appears at how the UK is behaving proper now, it’s acceleration. We’re shifting from writing code to having brokers construct and deploy code for us at scale Alex Rutter, Google Cloud EMEA
To help with the governance of those sprawling agent environments, Google launched Agent Gateway. This device serves as a single management level to control first- and third-party brokers throughout a number of clouds, offering OTEL (Open Telemetry)-compliant telemetry and granular “agent id” monitoring. This enables directors to deal with brokers like a digital workforce, to watch dependencies and actions.
Alex Rutter, managing director for AI at Google Cloud EMEA, argued that the “wait and see” period of AI is over for the UK market, including that the main target is now on “AI fluency” and the power of organisations to deploy industry-specific brokers that resolve issues on the edge.
“The ‘wait and see’ period of AI is completely over,” stated Rutter. “And if there was one phrase I may call to mind that basically appears at how the UK is behaving proper now, it’s acceleration. We’re shifting from writing code to having brokers construct and deploy code for us at scale.”
Rutter additional famous that the addition of Agent Observability gives deterministic instrumentation throughout all the agent move, giving companies absolutely the visibility they want into why an agent is behaving a sure means earlier than it impacts manufacturing.
As UK companies transition from proofs of idea to production-ready agentic options, the competitors between hyperscalers is intensifying. By positioning its “unified stack” – from customized TPUs to high-level agent platforms – Google is trying to seize the subsequent wave of enterprise spending by providing a extra built-in, “agent-first” different to conventional productiveness suites.
