Salesforce just launched a new catch-all platform to build enterprise AI agents
Businesses will be able to build agents within Slack and manage them with natural language
Salesforce has announced the launch of Agentforce 360, a new combined platform for building, deploying, and managing enterprise AI agents as well as for multimodal and Slack-native agent interaction.
Core to the new platform is a redeveloped Agentforce Builder, through which Salesforce is aiming to make the experience of creating and testing agents as easy as possible.
Workers will now be able to create agents conversationally, using natural language, with Agentforce capable of building agents based on the user’s brief. These agents can then be tested with rapid iterations via a live preview panel and easily exported or connected to existing systems via API or JSON representations.
This pairs with Agent Script, a new portable JSON language to define the behavior of AI agents that was designed with human readability in mind, as well as new configurable controls for Atlas Reasoning Engine, which powers Agentforce.
The announcement was made a day ahead of Dreamforce 2025, its annual conference held at the Moscone Center, San Francisco.
Salesforce eyes a more streamlined Agentforce
Slack is front and center for Agentforce 360, acting as a user interface for interacting directly with Salesforce apps, agents, and more directly within the communication platform. This includes direct interaction with AI agents, as well as connectors for third-party file storage platforms.
Read more in our dedicated write-up on all the new features coming to Slack.
Separately, Salesforce revealed Agentforce Voice, a new service that powers agents that can realistically interact with customers over the phone.
Enterprises can configure Agentforce Voice agents to follow their brand’s tone of voice, as well as the extent to which agents are allowed to interrupt customers and how long they stay on the phone to give customers a pleasant user experience.
Built on a new framework developed by Salesforce AI Research known as ‘eVerse’, Agentforce Voice can transcribe and stream audio with low latency so the conversation flows naturally.
Once built, agents can be used for phone calls, web apps, text messages and more, and natively link to contact center as a service (CCaaS) platforms including Amazon Connect, Five9, Genesys, NiCE, and Vonage.
Salesforce doubles down on Agentforce for AI ROI
The CRM giant first launched its agentic AI offering, Agentforce, last September and has since doubled down on its use of the platform for internal as well as customer use.
Srini Tallapragada, president & chief engineering and customer success officer, told assembled media that Salesforce has been using Agentforce for customer service and that agents have already answered 1.8 million customer conversations.
Salesforce has repeatedly underlined the usefulness of Agentforce at both the bottom and top line, cutting across different enterprise segments.
For example, Tallapragada explained that Salesforce is also using Agentforce for lead generation, as a company that hosts huge events and often has customers express interest in its services.
“But even though we are a big company, we are also limited in the number of people – most often what happens is we are not able to follow up on all the leads,” he explained.
“We have an SDR team which only takes the top leads, top one percent, two percent. We leave a lot of leads without follow-up, it's too costly for us. Now Agentforce has an SDR which is doing thousands of leads, following up.”
Agents rooted in enterprise data
One of the standout features of the new platform is Context Indexing, which uses unstructured data from across an enterprise such as PDFs, images, flowcharts, and data tables as context for Agentforce.
The process is automatic and can be controlled by data teams using low-code inputs to analyze and fine-tune the knowledge bases upon which Agentforce draws within their enterprise environment.
Salesforce has added additional structured-level governance to Data 360 alongside the feature, to tie new context indexes to existing organizational security and privacy controls.
Tallapragada explained that without access to the right context enterprise AI pilots fall into “pilot purgatory”, citing a recent MIT study which found that just 5% of enterprise AI pilots generate ROI.
For example, Heathrow Airport's AI agent “Hallie”, which has hit a 90% chat resolution rate without needing human intervention, is grounded in data from its CRM, customer loyalty program, digital bookings, and Wi-Fi data to encompass relevant customer information alongside relevant live data such as boarding times.
The agent is accessible via WhatsApp and is core to Heathrow’s goal of increasing its digital contact efficiency by as much as 40%.
Individual parts of Agentforce 360 will be released on a rolling basis between now and the start of 2026.
Agentforce Voice and Context Indexing will be available this month, while Agent Script, New Agentforce Builder, and Agent Graph will be released in beta in November.
Make sure to follow ITPro on Google News to keep tabs on all our latest news, analysis, and reviews.
MORE FROM ITPRO
- Salesforce may have had its ‘ChatGPT moment’ with AI agents
- Marc Benioff’s agentic AI gambit is paying dividends
- Salesforce has cut 4,000 customer support staff for AI agents so far
-
“Enterprises are becoming much more rigorous about the economics of AI”: Snowflake wants to help you cut AI costs by choosing the right model for the right taskNews New dynamic model routing capabilities aim to help customers box clever when it comes to their AI model choice
-
Microsoft has joined the growing list of companies cracking down on ‘tokenmaxxing’News The company is updating internal guidance to reduce rising costs
-
‘AI cost management has the same problems that cloud had’: Enterprises are still facing huge AI bills thanks to ‘tokenmaxxing’ – that means FinOps practices are more important than everNews With firms facing surging AI bills, FinOps techniques are more important than ever
-
‘We are now seeing MAI models outperform general-purpose frontier models’: Microsoft CEO Satya Nadella touts in-house models to cut spiralling AI costs – and reduce growing reliance on frontier labsNews The Microsoft chief says pricey frontier models don't have to be used for every task, and its own in-house MAI models could be the key to reducing costs.
-
3 ways to get your data AI-readySponsored Data placement, preparation, and orchestration make or break an AI project – and the Dell AI Data Platform reduces the strain on enterprise IT teams across both storage and the data foundation
-
AI agents could make living off the land attacks ‘much more dangerous’, says CrowdStrike Field CTOAnalysis Living off the land attacks are evolving rapidly as threat actors target agents with deep access to enterprise networks and sensitive data
-
‘A company should be able to use a model without giving up the knowledge that makes it unique’: Microsoft CEO Satya Nadella says enterprises shouldn’t be sharing so much data with AI providersNews The Microsoft chief warned that corporate data could be at risk thanks to AI models
-
‘What we’re seeing right now is just rapid escalation in AI token spend’: Accenture tells staff to stop using AI for unnecessary tasks amid surging costsNews Accenture has told some staff to roll back the use of AI for basic tasks while the consultancy grapples with surging AI token costs.
