Salesforce has achieved overwhelmingly positive results with its AI agent platform Agentforce, setting a record for its fastest-growing platform to date.
In its latest quarterly earnings report, the CRM giant announced that the platform it released just over a year ago had reached an annual recurring revenue (ARR) growth of 330% year-over-year.
Customers who adopted products like Agentforce and Data 360 saw success in implementation, with many now transitioning from experimentation to real-world use.
Agentforce – Salesforce’s Fastest Growing Product
Agentforce has proved to be Salesforce’s most popular product, having driven strong customer deployment rates and positive feedback.
Mark Benioff, Chair and CEO of Salesforce, highlighted how the platform has well surpassed the expected benchmarks.
“We’ve delivered incredible results with Agentforce. It’s really exceeding our expectations,” He explained.
“This is our fastest growing product ever.”
Its deployment foundation, Data 360, also showed substantial growth numbers in Q3,
The AI agent platform has excelled in both adoption growth and profitability, with Agentforce alone achieving an ARR of $500MN, and with the help of Data 360, earned Salesforce an ARR of $1.4BN, growing 114% year-over-year.
Furthermore, active customer accounts of Agentforce have grown 70% quarter-over-quarter, with high potential for more enterprises to begin live deployment, having moved beyond pilot testing phases.
Since the platform launched, Salesforce has closed a total of 18,500 Agentforce deals, up 50% quarter-over-quarter, while also handling over 1.5 million support requests, many of which managed without human agent intervention.
Benioff further emphasized the positive effects of the platform, describing Agentforce and Data 360 as “the momentum drivers.
“We now have over 9,500 paid Agentforce deals and 3.2 trillion tokens processed, underscoring our leadership in building the Agentic Enterprise and driving real outcomes.”
Early Adopters and Success Stories
These results are primarily an effect of early customer adoption, with existing customers expanding their systems with Salesforce’s products to achieve positive outcomes.
Customer investments into Salesforce have improved, with more than half of Q3 Agentforce and Data 360 bookings being completed by existing customers, as users continue to see great potential in the platform.
It also shows that early adopters are likely experiencing positive results from Agentforce adoption, indicating enterprises are already seeing ROI if Salesforce investments are deepening.
This highlights a broader trend of customers becoming more comfortable with AI adoption, with deployment willingness on the rise, as more enterprises move out of the experimentation stage.
In Q3, Salesforce had reported a 119% upsurge in AI adoption for customer service during the first half of 2025, from both consumers and customer enterprises, noting a 65% increase in AI agent interactions.
This real-world deployment can be traced back to Salesforce’s large customer base, which has been transforming its customer service functions with the help of Agentforce.
In fact, Benioff points out that “Nearly 90% now of all of the Forbes top 50 AI companies are using Salesforce.”
In one example, the kitchen and homeware business, Williams-Sonoma, has deployed Agentforce 360 across the brand and internally renamed it ‘Olive’, functioning as the company’s ‘sous-chef agent’ to handle customer-facing tasks on the website such as interactions and providing answers to queries.
This tool has proved beneficial to Williams-Sonoma, as the Olive agent currently handles approximately 60% of conversations via the website, delivering success in customer personality and delivering valuable interactions.
Benioff highlights his admiration for the company’s spin on the tool:
“I think the quality is what I’m most impressed with, that it’s really very, very good. You don’t see hallucinations.”
Agentforce has also seen deployments in government services in both the US and the UK to automate and handle routine activities.
The IRS has utilized Agentforce agents to provide automation capabilities to various departments for time-saving efficiency.
The organization reported that the Office of the Chief Counsel was saving up to 98% of manual activities with Agentforce’s automation tool, speeding up tax court case openings from 10 days to 30 minutes, allowing for higher levels of productivity and court processing.
Furthermore, it reported that another division had been saving approximately 50,000 minutes every year after deciding to retire their traditional systems and switch to Agentforce.
“Now Agentforce with IRS is going to be able to further optimize, automate, accelerate business processes across the entire agency,” Benioff explained.
The UK police department has also utilized Agentforce, having deployed the Service agent “Bobby” earlier this week to handle customer-facing calls in non-emergency situations.
This agent can provide instant responses to customers and cover more than 90 topics, acting as the public’s first point of police contact in several police departments.
These departments have already reportedly seen a 20% reduction demand in human agents receiving non-emergency calls.
“We work with multiple police departments to roll out Bobby. Everybody loves Bobby,” Benioff said.
These results have prompted Salesforce to expand Agentforce’s portfolio further, launching products and capabilities such as Agentforce 3 to transform AI adoption, and Agentforce Service to tackle rising customer demands.
This service cloud now offers Omnichannel Supervisor, allowing enterprises to switch between LLM and a human agent immediately if an LLM runs into its conversation limit, ensuring that human-agent collaboration remains vital even beyond AI experimentation.
Additionally, Agentforce has created and expanded partnerships with other giants such as Google to deploy AI agents on its communication products, and launched fraud protection tools with Vonage to improve its security governance capabilities.
Salesforce has also made strides to support global outreach by enhancing language capabilities and expanding deployment efforts to widen its market base.
Regarding tools such as Agentforce Service and Slack, Benioff highlighted: “We’re delivering this capability to a global customer base, more than 150,000 Salesforce customers, and one million companies are now on Slack, now have the immediate opportunity to work side by side with agents and Agentforce – that’s why we’re uniquely positioned for this new era.
“We have the strategy, the platform, the global scale.”
Shifting From AI Experimentation to Agentic Enterprise
When it comes to making the initial step from experimentation to agentic enterprise, many companies begin to struggle as they don’t yet have the tools and information required to do so successfully.
In fact, attempting to scale AI for real-world use correctly has left some companies in experimenting limbo for years, seeing neither deployment nor ROI.
However, whilst customer frustration is a challenge for Salesforce when it comes to introducing products that no-one can implement, there are still ways to combat this barrier.
Miguel Milano, President, and Chief Revenue Officer at Salesforce, explains how companies can get out of their AI rut and start deployment.
“The last mile is hard. The last mile is hard because companies need the context,” he said.
“For enterprise AI to be successful and accurate in the enterprise, you need the context. You need the data. You need the metadata. You need the deterministic workflows.”
This highlights a core issue for many enterprises: LLMs aren’t enough to be integrated on their own, with deployments needing workflows, apps, and data to function at a basic working level.
Miguel continued: “You want the agents to execute in a deterministic way the same workflows that that company had already codified in the apps for the years that humans are already using.
“They need AI that is embedded where the humans are. That’s why it’s so important to have the data with the context, to have the apps, the deterministic workflows, to have the AI where the humans are.”
Salesforce is already solving this challenge by positioning Agentforce as the key to deployment success.
By utilizing Agentforce’s Data Cloud tool for data and context unification, enterprises can give AI agents organized, messy, and meta data to output contextual answers and results for teams.
Deploying Agentforce 3 allows enterprises to revamp their current architecture and receive improved performance and resiliency with enterprise readiness, allowing large scale deployments to be integrated without fear.
This tool also includes MuleSoft, which allows AI agents to access various enterprise systems, services, and databases for speedy context retrieval.
Other tools, such as Agent Builder and Prompt Builder, let customers create personalized AI agents for testing and deployment for customized usage.
The observability and management layer, Agentforce Command Center, offers complex agent management, allowing enterprises to track, monitor, measure activity, and manage agent rules in line with company policies.
In fact, Salesforce has redesigned all of its products to include agentic enterprise abilities, allowing customers to utilize the ability across all workstations once experimentation has been completed.
Benioff explained: “It’s that customer data, the service data, the sales data, the marketing data, and then we’re able to deliver it in a tremendously friendly way.
“We’ve rebuilt all of our products to deliver this agentic enterprise, really just getting going.”
Salesforce Key Earnings Results
Salesforce has achieved strong results in its third quarter, having achieved solid performances in AI solutions.
- Salesforce reached a total revenue of $10.26BN, achieving a 9% increase year-over-year
- It’s largest ever-growing platform, Agentforce, achieved an ARR of $500MN, resulting in a 330% increase year-over-year
- Salesforce saw EPS increase to $3.25BN, exceeding expectations by 13.64%
- It achieved a free cash flow profit of $2.2BN, up 22% year-over-year
