Why Salesforce Will Win Autonomous AI

This article first appeared on GuruFocus.

I’ll admit it: Salesforce Inc. (NYSE:CRM) single-digit revenue growth looks pedestrian next to AI darlings posting 40%-plus gains. However, I believe a fundamental shift is underway. Salesforce’s Data Cloud and Agentforce platform which enables autonomous AI agents to handle customer service, sales tasks, and workflows just hit $1.2 billion in annual recurring revenue (ARR) with 120% year-over-year growth. For context, that’s faster than the adoption of Slack, Tableau, or Mulesoft at similar stages. So, while the adoption of autonomous AI agents isn’t as immediately flashy as selling GPUs, Salesforce is monetizing at scale. Simultaneously, the company is expanding operating margins to record highs. That being said, Salesforce’s valuation seems to be at odds with its recent performance and opportunity in agentic AI. Trading at 5.85x sales, near a two-year low versus a 7.73 historical median, its stock appears to be priced for deceleration when the AI revenue inflection is materializing.

Salesforce offers a multi-tenant Software-as-a-Service (SaaS) customer relationship management, or CRM, platform that includes Sales Cloud, Service Cloud, Marketing Cloud, Commerce Cloud, Tableau (analytics), MuleSoft (integration), and Slack (collaboration), among others. The company segments customers by SMB (under $50 million revenue served via $25/user/month Starter Suite), mid-market ($50 million to $1 billion via Professional/Enterprise editions), and enterprise ($1 billion-plus via Enterprise/Unlimited editions). It deploys a sales-led model that aims for annual contracts. In its fiscal second quarter 2026 results, Salesforce reported a current remaining performance obligation, or cRPO, of $29.4 billion (up 11% year-over-year).

Before I get into the numbers, let’s first discuss the backdrop. Salesforce’s growth has moderated in recent quarters into the high single digits. Meanwhile, many of its peers in the Technology industry, such as Nvidia (NVDA), are leveraging artificial intelligence demand to drive immediate growth. Salesforce AI adoption hasn’t been as flashy, but it appears to be gaining momentum.

Salesforce’s FQ2 revenue was $10.24 billion (up 10% year-over-year), accelerating from 8% in prior quarters. In fact, this is the first meaningful reacceleration in over a year. More importantly, Salesforce’s core business (subscription) grew 11% year-over-year. Granted, its professional services declined 3%. Together, this signals that customers are adopting its platform faster and leaning on implementation partners rather than Salesforce’s own consultants.

What’s most impressive about Salesforce’s recent performance is not the reacceleration in revenue growth, but expanding margins. Non-GAAP operating margin reached 34.3% in FQ2, marking the tenth consecutive quarter of expansion. Likewise, Salesforce’s FCF margin expanded to 32.7% in FY25 from ~30% in FY24. The fact that Salesforce is simultaneously growing revenues and margins implies it has a good amount of operating leverage.

Looking ahead, the company raised its full-year FY2026 revenue guidance to $41.1 billion to $41.3 billion from the initial $40.5 billion to $40.9 billion, implying 8.5-9% Y/Y growth.

So, what is driving all of this? Here’s where the autonomous AI agents come in: Data Cloud and Agentforce combined now represent $1.2 billion-plus in ARR, growing an astonishing 120% year-over-year. Salesforce has closed on 6,000 Agentforce deals with an additional 6,500 in the works. So, this isn’t some experiment. Customers are actually putting these AI agents into production.

During a recent conference call, Salesforce described its AI agents as handling millions of conversations while humans are delivering the empathy and expertise. AI agents are also managing scheduling and logistics. It is really more complicated than that, though. They are operating across apps, departments, [and] silos.

Think about it. Instead of hiring three customer service reps at $50,000 each ($150,000 annually), a company can deploy three Agentforce agents at $2 per conversation or $550/month for unlimited usage. At scale, this costs a fraction of one human salary, with the added benefits of 24/7 operation across multiple languages with perfect CRM integration.

Autonomous AI agents are replacing entire workflows, but why Salesforce? I believe that Salesforce has major structural advantages to become the leader.

First, as is often the case, data is king. Agentforce agents aren’t just some general-purpose chatbots you have seen when you’ve gone on a website in the past. They are trained on each customer’s specific CRM data (think service history, product catalog, and business rules). Before the advent of AI and Agentforce, Salesforce had tens of thousands of companies’ customer interaction data. Replicating this data layer will be a major challenge for Salesforce’s competitors.

Second, customers aren’t jumping to hand customer interactions to AI without audit trails, permission controls, and regulatory compliance. Salesforce’s Einstein Trust Layer builds a trust and compliance layer that ensures every agent action is logged, every data access is governed, etc. This was not built overnight.

Last but not least, Agentforce’s embedded distribution across the CRM system simplifies deployment while maximizing data advantage. The agents can run natively inside Salesforce, so they automatically inherit permissions, see real-time pipeline changes, and access the same customer 360-degree view that human reps use. A standalone AI agent (from Salesforce competitors) would need APIs, syncing delays, and constant maintenance to replicate.

How large is this opportunity? There are roughly 17 million customer service roles globally, and, according to Gartner, By 2029, agentic AI will resolve 80% of common customer service issues without human intervention. Even if 30% of these workflows shift to autonomous agents over the next few years, that’s a $50-plus billion revenue opportunity. This is on top of Salesforce’s existing $38 billion (FY25) CRM business. The market’s current valuation of Salesforce seems to imply that Data Cloud and Agentforce are mere features, rather than a second business line inside the first.

My discounted cash flow model isolates what I believe are the two most important drivers of valuation: revenue growth and free cash flow (FCF).

In modeling Salesforce, Year 1 Revenue represents the consensus analyst estimate for 2025. In the trailing twelve months, Salesforce grew revenues 8.3% year-over-year and achieved a FCF margin of 31.6%. Salesforce’s stock beta is ~1.2 per multiple sources. This is also consistent with the software (system & application) industry beta of 1.24 per NYU Stern data. Of course, when it comes to software, Salesforce is more mature and robust than most companies. Everything else is either standard DCF methodology (e.g., terminal value 2.5%) or was derived from Salesforce’s most recent financial statements (e.g., debt, cash).

According to my model, Salesforce is undervalued by nearly 7% on a cash flow basis. However, the model assumes everything stays the same throughout the nine-year projection. What if Salesforce improves its FCF margin or revenue growth by 100 bps?

My sensitivity analysis reveals how small changes to growth or margin impact valuation.

It also tells us that growth is the primary driver of Salesforce’s valuation, which could explain why its stock has been stagnant besides material improvements in margin.

Let’s compare scenarios. What if Salesforce’s growth returns to >10% and margins continue to improve? On the other hand, what if growth continues to contract?

At this moment, Salesforce appears to be priced for my Bear scenario.

Other valuation metrics echo my sentiment that Salesforce is priced conservatively.

Now that we have Salesforce’s Base scenario, let’s explore how the company can leverage its growth and margins.

Per IDC, Salesforce remains the #1 CRM provider for the 12th consecutive year, with 20.7% market share. Microsoft Dynamics comes in second with ~18% market share. Oracle and Adobe are close behind. As of October 2024, Salesforce ranks #1 on the Gartner Magic Quadrant for CRM Customer Engagement Center. Microsoft (MSFT) and Oracle (ORCL) are also pegged as Leaders.

As alluded to, Microsoft is Salesforce’s fiercest competitor. Its Dynamics 365 can integrate easily with Office and includes Copilot AI. That being said, Microsoft lacks Salesforce customization depth and has a smaller ecosystem. Of course, Salesforce addresses Microsoft head-on, claiming that (based on its own internal metrics) 71,000+ Microsoft customers choose Salesforce for CRM.

Certainly, there is room for multiple players. Grand View Research estimates that the CRM market is projected to reach $163.16 billion by 2030 (CAGR of 14.6%). Although competition is a major risk to Salesforce, there’s no reason to think that it is simply going to fall out of its clear and sustained leadership position.

Beyond product features, Salesforce benefits from switching costs that create significant customer lock-in. Think about it. Once a company’s operations are deeply embedded in Salesforce’s ecosystem, moving away becomes quite expensive and risky. Years of customer interaction history, custom fields, workflows, and business logic accumulate, and migrating all the data is costly, time-consuming, and risky.

Many of its customers have built thousands of custom objects, fields, and Apex code (Salesforce’s proprietary programming language) tailored to specific processes. Redoing all of this on another platform essentially requires building from scratch.

Another one is ecosystem lock-in. Salesforce’s AppExchange marketplace hosts over 5,000 third-party applications that integrate natively with the platform. Many companies rely on these apps for specialized functions like marketing automation and contract management. So, switching CRM providers is not merely switching CRM providers; it is replacing the entire app stack.

Interestingly, I think the data network effect will compound with Agentforce. What do I mean by this? Every interaction processed through Salesforce becomes a breeding ground for its AI agents to become more effective. This creates an elusive flywheel, wherein better AI agents ? more customer value ? more usage ? more training data ? even better agents.

Lastly, don’t underestimate organizational inertia. Changing the CRM requires CEO-level sponsorship, board approval, and acceptance of significant business risk. This is not something that companies undertake with ease or certainty.

That being said, this is all hypothetical. How does this play out in real life? So far, so good for Salesforce, judging by its market share leadership (IDC) and repeated Gartner Leader placements. You can also combine this with double-digit cRPO growth and low revenue attrition (<10%). GuruFocus GF Score of 92/100 with Profitability Rank 8/10 and Financial Strength 7/10 further supports durability despite slower headline growth. To sum it up, Salesforce has a durable competitive position in CRM, and Agentforce could not only expand its TAM but also expand its leadership position within CRM.

While Agentforce is off to a solid start, the true test is renewal rates and expansion. If customers find that AI agents require excessive human oversight or don’t justify ROI versus human labor, adoption could stall. Agentforce is still very much in the early innings. The $1.2 billion ARR figure is merely ~3% of Salesforce’s total revenue.

Second, while Salesforce has a first-mover advantage in CRM, it’s apparent that competition is intensifying. Microsoft poses a significant threat, particularly as it bundles Copilot across its entire ecosystem, and it has the advantage of deep Office 365 integration. Oracle and Adobe are also making heavy investments in AI-powered CRM. All three of these companies have deep enough pockets to undercut prices to achieve more market share.

As always, macro headwinds, such as a recession or tightening credit, can pressure enterprise spending. Salesforce’s average deal size of over $1 million means that decisions require executive sign-off, board approval, and extended procurement cycles.

Operating margins have expanded for ten consecutive quarters, but Salesforce is nearing best-in-class levels for enterprise SaaS (~35%). This begs the question: how much more can Salesforce actually improve its margins? It could cut back on R&D or sales investment, but this could handicap its competitive positioning.

While FQ2 revenue growth accelerated to 10%, full-year guidance implies 8.5-9% growth. If Data Cloud and Agentforce fail to offset core CRM saturation, Salesforce may slip back into mid-single-digit growth. Should this happen, Salesforce’s P/E ratio of 35 would look fair rather than cheap.

A broader industry risk is regulatory and trust concerns. AI agents handling customer data and making decisions face increasing regulatory scrutiny. This is already taking place in Europe (AI Act) and in healthcare/financial services.

Recall that I opened by acknowledging that Salesforce’s single-digit revenue growth looks unexciting compared to AI infrastructure plays posting 40%-plus gains. That being said, I believe the market is mispricing the inflection point that is happening right now.

The company is monetizing AI at scale. Data Cloud and Agentforce eclipsing $1 billion in ARR and growing over 100% Y/Y is nothing to scoff at.

As implied by the millions of customer services roles globally that Gartner projects can be almost entirely replaced by agentic AI, the opportunity is substantial for the CRM leader.

While my DCF analysis suggests 7% upside even in the base case, the real opportunity lies in what happens if revenue growth is >10% (as FQ2 demonstrated) while margins continue expanding. In that scenario, Salesforce could be undervalued by over 25%.

To wrap up, the risk/reward appears asymmetric. The market is pricing in deceleration while the AI revenue inflection is materializing.