Google Cloud Summit: Unifying data, data residency and agent-driven AI
Google Cloud’s London summit underpinned a focus on artificial intelligence with some significant infrastructure updates
Google Cloud continues to gain traction as businesses roll out artificial intelligence (AI). Its data management and analytics tools are also receiving significant improvements. And Google is improving connections between its AI tools and the rest of its technology stack.
Google Cloud’s president of EMEA, Tara Brady, told attendees at the London edition of the Google Cloud Summit that the majority of AI businesses in Europe and the UK are running on Google Cloud Platform (GCP).
As many as 90% of AI “unicorns” are in the cloud and over 60% of funded AI startups are Google Cloud customers. Google Cloud itself would rank as the world’s fourth-largest software company if it were a business in its own right, he said.
Google is increasing its investment in the UK startup scene, with a London Startup Hub forming a new “digital-native community space”, and the company has showcased a range of AI and cloud-based ventures.
These range from OXA, an Oxford University spin-off that retrofits autonomous driving kit to existing vehicles, to VEED, a video tool that uses AI to create ultra-realistic video avatars that individuals can then use to read out presentations, in their own voice, from a script.
More prosaically, DIY retailer Kingfisher showed how it was using image recognition and Google’s generative AI (GenAI) tools to allow its Screwfix trade customers to find replacement parts from the firm’s online catalogue using their phone camera.
Intelligent agents
Alongside its reference customers, Google Cloud announced several developments in its AI technology, especially for connecting GenAI tools to enterprise data.
Lloyds Banking Group is using Google’s AI tools across back office and engineering functions. One of its early applications is code translation for developers working on application modernisation. According to Ranil Boteju, the bank’s chief data and analytics officer, AI has brought efficiency gains of between 30% and 40%. The bank has also made significant savings in origination processes, which have long been a labour-intensive part of financial services.
Lloyds Banking Group is just one organisation making use of AI to improve software development efficiency.
Changes to Google’s AI toolset announced in London include Gemini Code Assist Enterprise, which claims to offer developers “code that is more accurate and relevant to their applications”, and moves beyond AI-powered coding assistance.
Google is also promoting the idea of “agentic AI”, where AI-based agents will be able to create workflows and interrogate data – as well as interact with external systems – without the direct input of humans.
Agentic systems hold the promise of allowing enterprises to solve problems directly with AI tools, rather than using AI as an assistant with a “human in the loop” to craft the prompts that large language models (LLMs) currently need to be effective.
With agentic AI, chains of agents can pass actions or data from one to another, rather as humans do in a current workflow, but also have the intelligence to improve the process as they go.
These chains of autonomous agents might still be some way off, according to Google Cloud’s Yasmeen Ahmad, product executive for data, analytics and AI, but she expects them to become increasingly important as enterprises try to extract value from ever-growing volumes of data.
To build agentic AI, Google has around 130 models in its Vertex AI platform, with the promise of more to come, especially in the data analytics space. “We are at the precipice of a revolution, driven by data and algorithms,” said Ahmad.
Conversational analytics
As Ahmad suggests, progress in AI will be driven by – and could also be limited by – access to relevant, high-quality data. Google Cloud announced investments both in business analytics tools and its underlying infrastructure.
Enterprises need additional ways to interrogate data. One area where GenAI has been successful is in allowing firms to use more conversational approaches, rather than relying on structured queries written by data experts.
Gemini will now be available in Looker Conversational Analytics so that organisations can ensure they have “a single source of truth” for their data.
And Gemini in BigQuery is now in general availability, allowing “entire teams to chat with their data” to gain insights, according to Google. These are part of Google Cloud’s strategy to improve connections between businesses’ proprietary – and quite often sensitive and confidential – data and AI.
In analytics, Google Cloud has added a number of enhancements to BigQuery, including a managed experience for Iceberg, Hudi and Delta file formats, BigQuery’s Unified Catalogue, now generally available, and semantic search for BigQuery, now in preview.
There will also be additional governance tools in BigQuery and support for open data formats, including Flink and Kafka.
Underlying this, for UK customers, Google Cloud has expanded its data residency. This allows customers to carry out all Gemini 1.5 Flash machine learning processing in the UK. They will be able to store their data at rest, as well as ML processing, “entirely in the UK”.
Google Cloud now has over 40 cloud regions and 121 zones, according to UK sales director Anne-Marie Lamb, with another 10 under construction, as well as 32 undersea cables. One of these, Grace Hopper, was laid in 2021 and has a Kevlar construction that even resists sharks. Google is spending $1bn on a new UK datacentre at Waltham Cross. The site will run on carbon-free energy by 2030.
Competitive edge
Google Cloud believes its infrastructure investments, AI developments and work on its data analytics tools give it a strong value proposition, compared with other hyperscalers.
It is also continuing to invest in its productivity toolset, Workspace. Future developments for this include Vids, a tool that allows non-specialists to create video content, linking both into media sourced via Google and corporate assets.
At the same time, Google is pursuing a complaint against Microsoft through the European Commission, accusing its rival of anti-competitive practices, including punitive pricing for customers that want to use Microsoft’s server in non-Azure cloud environments.
But Google Cloud argues its integration of AI and business intelligence gives it an edge.
“Our data and analytics capabilities enable organisations to leverage our cutting-edge ML and AI capabilities such as Gemini, so enterprises can transform their business operations, unlocking significant business value,” said Adrian Poole, director, digital natives, at Google Cloud UK and Ireland.
“Our Vertex Platform also gives clients the choice to use many other [AI] models such as Anthropic or Mistral. At Google Cloud, we have a strong competitive edge in key areas such as ML/AI, data analytics and multicloud flexibility, to name a few.”
Poole added that 75% of leading independent software suppliers are already using agentic AI technology within Google Cloud to carry out business functions.
Although there is increasing talk about cloud repatriation among CIOs, Google Cloud believes cloud growth will continue, even as some enterprises look for hybrid solutions.
“We recognise that these types of conversations are important,” said Poole. “However, the broader trend continues to be around the adoption of hybrid cloud and multicloud environments. Google Cloud’s hybrid and multicloud approach enables solutions like Vertex AI and BigQuery to be deployed in hybrid setups, which allow businesses the flexibility to place their workload where it makes most sense for their needs.”
Read more about enterprise AI
- The industry is ramping up efforts to help enterprise IT deliver artificial intelligence-based business applications.
- The successful execution of an enterprise’s AI strategy lives or dies on the quality of the data underpinning it, so how can companies ensure they are on the right path?
