Thomas Kurian On Google’s Key Agentic AI Pillars, Microsoft Competition

Google Cloud CEO Thomas Kurian explains to CRN his company’s bold agentic AI strategy and Google’s market differentiation versus Microsoft.

CEO Thomas Kurian is bullish about Google Cloud’s agentic AI future as the tech giant’s innovation engine roars around Gemini, Workspace and AI agent platforms.

The $48 billion Mountain View, Calif.-based company has launched a slew of agentic offerings for partners to create new offers and next-level AI services on top of them, including its new Agentspace platform for building custom agents, its flagship Vertex AI solution, as well as pre-packaged AI agents that partners can sell.

“We have state-of-the-art models and platforms for people to build their own agents on top as we see this as a huge opportunity for partners,” said Kurian. “We’re not a services company. We want partners who deliver value. We’ll share the pie with you. So it’s not about taking a fixed-size pie and cutting it up. It’s about giving partners the opportunity to grow with you.”

In an interview with CRN, Kurian talks about Google Cloud’s AI market differentiation versus Microsoft Copilot as well as his company’s go-to-market strategy with channel partners.

How is Google Workspace with Gemini different from Microsoft with Copilot?

The big differences are, No. 1, every one of our customers and all users have access to Gemini so it makes it much simpler for partners to build value-add on top.

No. 2, the quality of the technology is materially different. So when you go to different clients and get them to use it, the quality of our software is materially better.

In what we do in Google Meet versus what others do with their AI agents in videoconferencing, for example, there are material differences both in quality and capability. For example, a partner says, ‘Hey, we’re now going to introduce videoconferencing in call centers because a lot of call centers want to use it.’ Don’t you think simple things like how good is the fidelity with which the recording happens are important? How good is the translation capability? Because many customers may be talking in Spanish or Tagalog or other languages, and the agent is listening to the translation coming in over the videoconference. Is it a good, real-time translation? What’s the quality of the fidelity of the translation?

Lastly, No. 3, is AI going to lead to more security breaches and cybersecurity problems? I don’t need to say anything. Just look at the companies’ track record on security between [Microsoft] Office and [Google] Workspace. The data speaks for itself. I have not had to write a memo to my organization saying security is a priority because we’ve had all these breaches.

What are Google Cloud’s agentic AI key pillars and go-to-market strategy?

An AI agent is a software system that uses a sophisticated AI model that both understands different kinds of data and information, can reason on it, and then can use tools to automate tasks and automate workflows in companies. That’s what it is. We have four things today.

One, we offer platforms for people to use a broad range of models—certainly Google Gemini, but also Anthropic and other models to build AI agents. If you’re a professional developer, you can build on Vertex. If you’re a line-of-business user like the head of marketing and you want to build a campaign agent, you can use Agentspace.

Two, we work with partners like systems integrators such as Accenture, Deloitte, 66Degrees and Pythian to build agents for different industries and for different domains within a company. Different industries, for example, health care—I want a digital nurse who I can ask questions to after I leave the hospital because I need to get answers on a lot of after-care questions. There are many hospitals building after-care digital systems that partners can build on top of.

Also, much of the use of these agents is not just in the IT department of companies, but in the business lines like, ‘I’m working with Puma to build a marketing campaign for their shoes using our models.’ Or ‘I’m working with Verizon with Accenture to build a customer service agent for them.’ The key person there is the head of customer service. So that’s one big opportunity for the partners.

Third, we offer packaged agents of our own. So with our packaged agents in Google Workspace, we’ve integrated AI tools to help people write better, create slides for them, assist taking notes in meetings, etc. Think of an agent as a graphics artist or a person who’s a meeting note-taker. That helps take Workspace into markets that we may not have reached before and it expands the addressable market.

We also offer packaged agents in our Google Cloud Platform, like data agents to help people do analysis better. We built a cybersecurity agent that allows you to identify what threats exist externally and then compare that to your current operational posture to detect if you have security vulnerabilities and what’s the attack path somebody’s going to take.

Then the last key thing we’ve done is maintain an open platform, which allows us to interoperate. Because we enable an open platform and interoperability, it allows the addressable market for our solutions to be as wide as possible, which in turn, is creating a lot of opportunities for our partners.

What’s your message to Google partners?

We as an organization are 100 percent committed to giving you the best products and solutions to help you grow your business. As the pie gets bigger, we’ll offer more solutions to you that expand our opportunities together.

My message to all our channel partners is really simple: Thank you for all that you’ve done with us these past several years. For every partner who has not done business with us, give us a shot. We’ve helped so many grow and succeed.