thoughts from themes
thoughts from themes
Salesforce: Well-Positioned for the Next Phase of AI?
Date: April 16, 2025
Artificial intelligence (AI) technology is seeing rapid adoption today. ChatGPT, for example, now has 400 million weekly users1 – roughly twice the number of users it had six months ago.
While apps like ChatGPT are impressive, they only offer a glimpse of AI’s potential. Looking ahead, we are likely to see a shift towards AI agents – sophisticated software systems that can operate autonomously and increase business productivity significantly.
One company that appears to be well positioned to benefit from this trend is Salesforce. With its “Agentforce” AI agents – which are designed to handle a range of business interactions autonomously – it could be a powerful force in the next phase of AI.
Transitioning from CRM to Agentic AI
Founded in 1999, Salesforce has historically been a customer relationship management (CRM) software company. Today, it provides CRM solutions to over 150,000 customers2 worldwide, generating annual revenue of more than $30 billion3.
Recently however, the company has been shifting its focus to agentic AI technology with the aim of revolutionizing customer service and sales through intelligent automation. Akin to a digital labor platform, its Agentforce solution is a proactive, autonomous AI application designed to provide specialized, always-on support to customers or employees.
It’s still early days here, but this new technology appears to have a lot of potential. Salesforce’s AI agents can initiate and complete sequences of tasks, handle natural language conversations, and provide relevant answers drawn from business data.
This means that companies can use them to engage with prospects and resolve customer queries 24 hours a day. Leveraging the power of Agentforce, businesses across a range of industries – including retail, financial services, and healthcare – can potentially increase their productivity significantly.
3,000 Paid Customers in Just a Few Months
On Salesforce’s recent Q4 earnings call, CEO Marc Benioff noted that customers are having success with the company’s AI technology and said that the pace of adoption has been rapid. “We’ve never seen products grow at these levels, especially Agentforce,” he told investors.
What’s notable is that just 90 days after the technology went live, Salesforce already had 3,000 paying Agentforce customers. Overall, it has closed 5,0004 Agentforce deals since October.
On the earnings call, Benioff highlighted a few customers who are already using the technology to enhance efficiency. One such customer was OpenTable. Just three weeks after employing Agentforce, it was handling 73% of all restaurant web queries with the technology. This was a 50% improvement over its previous automation tool.
Another customer Benioff highlighted was Goodyear. With Agentforce for Field Service, the tire company can increase efficiency by autonomously scheduling field test appointments and assisting technicians with answers to vehicle-related questions. Other customers using the technology include Singapore Airlines, Gucci, SharkNinja, and Equinox. According to Benioff, customers are experiencing unprecedented levels of productivity, efficiency and cost savings.
It’s worth noting that Salesforce is using Agentforce technology itself. And the results speak for themselves. Since its launch on the company’s help portal in October, Agentforce has autonomously handled 380,0004 service requests, achieving an 84% resolution rate (with only 2% of the requests requiring human escalation). “Our digital labor force is resolving tens of thousands of customer service inquiries, freeing our human employees to focus on the most nuanced issues and customer relationships,” commented Benioff.
The Salesforce Differentiator
Now, there are many companies that are working on agentic AI today. Where Salesforce potentially has an edge, however, is in its unified architecture. Benioff believes that for AI agents to work effectively, there needs to be three key elements that work together – apps, data, and the agents. And Salesforce’s platform has been designed to harmonize these three critical layers.
For a start, Agentforce is deeply integrated with the Salesforce ecosystem where all apps are built on a common core and framework, ensuring consistent data handling. Secondly, its Data Cloud service (which recently surpassed 50 trillion records4) aggregates customer data and allows for seamless data retrieval, potentially enabling agents to access contextual information when interacting with customers.
It’s worth noting that research has shown that implementing Agentforce is far more efficient than trying to implement agentic AI solutions in-house. According to research from the Futurum Group, Agentforce can achieve ROI 5x faster than in-house solutions while lowering costs by 20%. Meanwhile, technology analyst firm Valoir found that Agentforce delivers autonomous AI agents 16x faster than DIY approaches with 75% more accuracy5.
Nonetheless, overall deliveries are clearly showing early signs of decline. The slight misalignment between production and deliveries indicate that factory activities might be turned down going forward: inventory trends indicate a surplus of stock relative to delivery trends.
A $12 Trillion Opportunity?
Of course, the AI agent adoption cycle is still early. As a result, Salesforce is expecting5 a modest contribution to group revenue in fiscal 2026 (with a more meaningful contribution in fiscal 2027).
According to Benioff, the market for digital labor could be worth anywhere between $3 trillion to $12 trillion5, so Salesforce may have a long growth runway ahead of it.
1ZDNET.com, ChatGPT's user base just doubled in 6 months - to more than 400 million weekly users, as of February 24, 2025
2Salesforce.com, as of April 2025
3Salesforce.com, Salesforce Announces Fourth Quarter and Fiscal Year 2025 Results, as of February 26, 2025
4CRM Q4 FY25 Earnings Press Release w financials, as of February 26, 2025
5S&P Global Market Intelligence, Salesforce, Inc. NYSE:CRM FQ4 2025 Earnings Call Transcripts, as of February 26, 2025
