Salesforce ramps up agentic AI research with new foundry project
Researchers are already working on new tools for agent-to-agent interaction and “ambient intelligence”
Sign up today and you will receive a free copy of our Future Focus 2025 report - the leading guidance on AI, cybersecurity and other IT challenges as per 700+ senior executives
You are now subscribed
Your newsletter sign-up was successful
Salesforce has announced the launch of a new AI Foundry service aimed at accelerating agentic AI research.
Through the service, the company hopes to ramp up internal AI development and testing, moving from “foundational research to product innovation faster than ever before”.
This, Salesforce noted, will have a downstream benefit for customers, enabling it to roll out more intuitive and powerful AI solutions.
The AI Foundry will see researchers collaborate closely with “strategic customers” and academic partners to develop and test new AI tools and capabilities.
“Many of the old rulebooks simply don’t apply anymore,” said Itai Asseo, VP of Salesforce AI Research.
“AI Foundry connects foundational research to real business problems by collaborating closely with our strategic customers in rapid iteration cycles.”
Targeted AI innovation
The new AI Foundry will focus primarily on areas where the company sees the “greatest opportunity for enterprise impact”.
Sign up today and you will receive a free copy of our Future Focus 2025 report - the leading guidance on AI, cybersecurity and other IT challenges as per 700+ senior executives
This includes simulated environments for testing agents based on real-world best practices and learnings. Indeed, researchers have already developed eVerse, a simulation environment that “exposes agents to thousands of edge cases, handoffs, and judgment calls”.
The aim here is to assess agent performance and consistency, and the tool has been used to stress-test Agentforce Voice using thousands of simulated conversations.
“Ambient intelligence” is another key focus area for the new foundry project, according to Salesforce, helping test agents that “disappear into the environment”.
These are essentially agents that are always on that enterprises can trust to work quietly behind the scenes without the need for repeated commands.
“Ambient intelligence is context-aware, proactive, and timely. It understands the full situation, anticipates needs before they emerge, and surfaces just-in-time insights,” the company explained in a blog post.
The AI Foundry is currently experimenting with ambient intelligence capabilities, with the eventual goal of embedding this within enterprise workflows to help aid workers on the go.
Agent ecosystem building
Another key area of focus for the new AI Foundry lies in agent-to-agent interactions, according to Salesforce, including those that act “across organizational boundaries”.
Researchers are developing rules and protocols aimed at governing cross-boundary agentic ecosystems, for example. This includes agent cards, which are essentially a “business card” for each individual agent, the company noted in a 2025 blog post.
These help enterprises establish whether a third-party agent interacting with their own internal tools is legitimate, and what it's designed for.
“Working with Salesforce’s legal council, as well as with the Office of Ethical Use of Technology, our research team is defining the legal frameworks required for autonomous agent negotiation,” Salesforce said.
FOLLOW US ON SOCIAL MEDIA
Follow ITPro on Google News and add us as a preferred source to keep tabs on all our latest news, analysis, views, and reviews.
You can also follow ITPro on LinkedIn, X, Facebook, and BlueSky.

Ross Kelly is ITPro's News & Analysis Editor, responsible for leading the brand's news output and in-depth reporting on the latest stories from across the business technology landscape. Ross was previously a Staff Writer, during which time he developed a keen interest in cyber security, business leadership, and emerging technologies.
He graduated from Edinburgh Napier University in 2016 with a BA (Hons) in Journalism, and joined ITPro in 2022 after four years working in technology conference research.
For news pitches, you can contact Ross at ross.kelly@futurenet.com, or on Twitter and LinkedIn.
-
Google just revised its ‘Q-Day’ timelineNews Google has warned that “Q-Day”, the point where a quantum computer is powerful enough to crack current encryption techniques, could come as soon as 2029.
-
Red Hat, Google Cloud eye legacy modernization gains with partnership expansionNews Red Hat wants to make it easier to migrate and modernize applications
-
Oracle announces new proactive enterprise agents at AI World Tour LondonNews With a slew of new tools and customization options, Oracle is aiming to ground AI agents directly in enterprise data
-
Meta engineer trusted advice from an AI agent, ended up exposing user dataNews The internal security incident exposed sensitive user data to unauthorized employees
-
Salesforce targets unified customer support automation with Agentforce Contact CenterNews Combining AI agents, telephony, and CRM, Salesforce is making a firm case for automated customer interactions and controlled
-
Salesforce targets telco gains with new agentic AI toolsNews Telecoms operators can draw on an array of pre-built agents to automate and streamline tasks
-
Kyndryl wants to help enterprises keep AI agents in line – and avoid costly compliance blundersNews Controls become machine‑readable policies that AI agents can read and must obey
-
Want to deliver a successful agentic AI project? Stop treating it like traditional softwareAnalysis Designing and building agents is one thing, but testing and governance is crucial to success
-
OpenAI's Codex app is now available on macOS – and it’s free for some ChatGPT users for a limited timeNews OpenAI has rolled out the macOS app to help developers make more use of Codex in their work
-
‘In the model race, it still trails’: Meta’s huge AI spending plans show it’s struggling to keep pace with OpenAI and Google – Mark Zuckerberg thinks the launch of agents that ‘really work’ will be the keyNews Meta CEO Mark Zuckerberg promises new models this year "will be good" as the tech giant looks to catch up in the AI race