Salesforce and Google partner on agentic cross-platform collaboration

Salesforce Inc. said today it’s trying to break down the siloes that separate enterprise’s customer relationship management data from their productivity tools, specifically for artificial intelligence agents.

It’s doing so by expanding its partnership with Google Cloud, unveiling a deep set of integrations that will enable autonomous agents to work seamlessly across both company’s ecosystems. The focus of the expanded partnership is all about “agentic interoperability,” Salesforce said. The goal is to enable AI agents deployed in the Salesforce platform to reach into Google applications such as Docs, Slide and Drive, so they can retrieve the context and data from those platforms to perform complex work.

The partnership will work both ways. Just as Salesforce’s agents can leverage Google’s tools, Gemini Enterprise will be able to trigger actions and retrieve data from its partner’s flagship CRM platform. The promise is one of a “bidirectional flow” of intelligence that aims to eliminate the friction of switching between different applications.

Salesforce describes this friction as a kind of “hidden toggling tax,” and says it’s a major limiting factor on worker’s productivity. For instance, someone might see a security alert pop up in Salesforce’s Slack platform, but the context behind that alert will be buried in Google Docs, with the final approval process living inside Salesforce itself. To deal with such an event, a human would have to keep flitting between various apps, eating up valuable time that could be better spent on more productive work.

A second challenge Salesforce wants to take on is the problem of data gravity. By this, it means the need to constantly move or copy data from one platform to another. This has always been necessary for agents to obtain the necessary context they need to perform more complicated tasks on behalf of humans, and it’s another major productivity killer. The company’s idea is to create a unified system that shares context across both platforms, so agents can access the data they need effortlessly, no matter where it lives.

Deep, cross-platform integration

Salesforce and Google have developed a number of new capabilities to fix these problems, integrating the former’s Agentforce layer with Gemini and BigQuery, which is Google’s cloud-based data warehouse.

For instance, the zero copy integration with Google Lakehouse will allow Agentforce agents natively to read and act upon data stored in Google Cloud without needing to move it first. The data will stay where it’s housed, remaining secure, while enabling agents to access it in real-time to aid in their decision-making.

Meanwhile, Agentforce now supports Gemini models natively through the Atlas Reasoning Engine. According to Salesforce, this will enable Agentforce to see across text, image and video formats and draw upon years of customer history to solve complex problems more accurately.

Elsewhere, the companies are integrating Gemini Enterprise with Slack. Gemini will be able to act as a cross-platform assistant that can, for example, summarize a Google Meet transcript while comparing it with a specific Slack thread that discusses action items resulting from that meeting. In addition, the Agentforce Sales agent has been given the freedom to operate entirely within the Gemini interface, enabling salespeople to engage leads, update their CRM records and create briefs without leaving whatever platform they’re working in.

Today’s announcements should help Salesforce and Google get closer to realizing their vision of truly autonomous business operations and an “agentic enterprise” that allows humans to drop mundane data entry tasks and focus on higher-level strategy. Under this vision, AI agents will automate the grunt work of cross-platform task execution.

The companies said they plan to double down on their partnership in the coming months and years with deeper integrations, including more vertical-specific agents and more sophisticated reasoning capabilities. The end goal is to blur the distinction between CRM and productivity tools and replace this with a single “intelligent fabric” that supports autonomous workers at large scale.

Salesforce President and Chief Engineering Officer Srini Tallapragada said enterprises are ready to go all-in on agentic AI, but they still need the infrastructure to support that by allowing AI models to operate across platforms. “Our deepened partnership with Google Cloud gives joint customers exactly that, so they can deploy Agentforce across every part of their business and accelerate their Agentic Enterprise transformation,” he said.

Image: SiliconANGLE/Gemini