The Agentic Pivot

For over two decades, Salesforce (NYSE: CRM) has been one of the biggest players in the Software-as-a-Service (SaaS) space. It pioneered the model of delivering business applications through a web browser, becoming a pioneer in the cloud industry we know today. However, the last year has been a period of intense scrutiny for the company and its investors. As Generative AI (GenAI) models became more capable, a SaaS-pocalypse narrative began to take hold. Why would a company pay for thousands of seat-based licenses if a handful of AI agents could do the work of an entire department?

That’s why many of these companies including CRM have tanked in the last few months and these earnings aren’t doing enough to assuage those fears. Salesforce’s Q4 2026 results represent an attempt to draw a line in the sand, the start of a move away from purely being a system of record to becoming a system of agency.

The Core Business and the “SaaS-pocalypse” Fears

Salesforce operates across several key clouds: Sales, Service, Marketing, Commerce, and Data. Historically, its revenue has been driven by seats, the number of human beings logging into the platform with each person often requiring a license to access the software. If AI makes humans more productive, companies might hire fewer people, leading to a direct hit on Salesforce’s top line. This fear caused significant stock price dislocation throughout fiscal 2025 and 2026. CRM is down almost 25% YTD and down over 37% in the last year with the stock trading down to 2023 prices.

Investors worried that Salesforce was clearly an AI laggard or, worse, a victim of the obviously incoming and destructive AI-driven automation. However, the Q4 results suggest that the company is doing a decent job co-opting this threat. By integrating AI agents into its existing workflow, Salesforce is repositioning itself not just as a tool for humans to use, but as the operating system that hosts the agents who do the work.

Q4 and Full-Year FY26 Performance vs. Expectations

Salesforce delivered a record quarter, proving that despite the noise, enterprise demand for its core platforms remains robust.

Q4 FY26 Results

• Revenue: $11.2 billion, up 12% Y/Y (10% CC) and right on top of expectations. This included a $399 million contribution from the Informatica acquisition.
• Subscription & Support Revenue: $10.7 billion, up 13% Y/Y.
• CRPO (Current Remaining Performance Obligation): $35.1 billion, up 16% Y/Y(13% CC). This is a critical forward-looking indicator, and while it slightly met guidance, the organic growth was 9%, a point of contention for some analysts who wanted to see a larger beat.
• Non-GAAP Operating Margin: 34.2%, showing continued discipline in operational excellence with GAAP operating margins coming in at 20.1%, with most of the difference being driven by the sizable stock-based compensation this company has.
• Non-GAAP EPS and Free Cash Flow: 3.81 versus 3.05 expected. GAAP EPS came in at 2.07 so you can see the vast difference driven by the removal of SBC, restructuring and acquisition related costs and amortization of purchased intangibles. On the FCF side, the company generated $5.3B in free cash flow, most of which went towards share buybacks.

Full-Year FY26 Highlights

For the full year, Salesforce hit a massive milestone: $41.5 billion in total revenue, up 10% Y/Y and right in line with what the market expected. The company also reached $72.4 billion in total RPO, signaling a massive backlog of work. They also generated $14.4B in free cash flow and returned most of it in the form of share buybacks and a small dividend.

When measured against market projections, Salesforce generally outperformed on the bottom line (EPS) and revenue, though the market’s reaction to CRPO numbers shows that investors are still looking for signs of organic re-acceleration beyond acquisitions like Informatica. Last year’s Q4 had 11% growth in constant currency and this year, that’s down to 9% although off a larger base.

Guidance: The Path to $63 Billion

This was the key to the whole report and while management essentially guided towards expectations for the full year but in a market like this one, companies that just meet expectations are often met with a resounding bleh as seen by the market’s negative reaction post earnings release.

• FY27 Revenue Guidance: $45.8 billion to $46.2 billion (approx. 10-11% growth).
• FY30 Long-Term Target: Management raised their FY30 revenue target to $63 billion+. This is a significant statement of confidence, suggesting they expect the AI transition to be additive to growth rather than cannibalistic. That’s a good sign but end of day, even projections one year out are hard to make so who really knows what’s going to happen by 2030 but it’s nice to see confidence from management.

The Financial Engine: Free Cash Flow, Buybacks, and SBC

One of Salesforce’s greatest strengths is its ability to generate remarkable cash flows, as Marc Benioff described them.

Free Cash Flow and Share Repurchases

In FY26, Salesforce generated $15.0 billion in operating cash flow and $14.4 billion in free cash flow (FCF). The company returned 99% of that FCF to shareholders through $12.7 billion in share repurchases and $1.6 billion in dividends.

In a move to capitalize on what they see as a dislocation in the stock price, the board authorized a new $50 billion share repurchase program. For a company with a market cap sitting under 180B, a $50B authorization is massive, representing nearly 28% of the company at current valuations. Of course that won’t all happen tomorrow but if the stock remains low then it could be accretive to long term returns if they can hit their stated goals by 2030.

Stock-Based Compensation (SBC)

A perennial thorn in the side of tech investors is SBC. In FY26, Salesforce reported $3.5 billion in SBC. While high, the company’s aggressive buyback strategy more than offsets the dilution but it does dull the impact of those buybacks. Still, Salesforce’s diluted share count decreased by 2% Y/Y. Management is clearly focused on total shareholder return and using buybacks to neutralize the impact of SBC. While buybacks are great, buying back shares in a company that continually drops isn’t ideal so they must turn that ship around to make those worthwhile.

The AI Strategy: From Seats to Agentic Work Units

The most transformative part of this earnings call was the introduction of a new metric: Agentic Work Units (AWUs).

The Token Era

Salesforce processed 19 trillion tokens for its customers all-time. However, tokens are a cost metric (what they pay model providers like Anthropic or OpenAI). To prove value, Salesforce needs to show work.

The AWU Metric

An AWU represents a discrete task accomplished by an AI agent, a decision made, a record updated, or a workflow triggered. Salesforce delivered 711 million AWUs in Q4 alone, with 2.4 billion delivered all-time so those tasks are clearly growing but the question becomes, how well can they monetize that and at what margins.

Pricing Model Transition

This is the holy grail for Salesforce and its future: transitioning from a seat-based model to a consumption-based model.

To highlight the potential of this shift, Benioff invited Jason Lemkin, the founder of SaaStr and a known figure in the SaaS world, to join the call as a guest. Lemkin provided a startling look into the future, sharing that SaaStr has effectively replaced a traditional human team with 20 AI agents managed by just 1.2 humans. That’s a bummer for their workers.

Lemkin’s testimonial served as a proof-of-concept for the new math:

1. Efficiency Gains: SaaStr used Agentforce to reactivate dead leads, closing $2.7 million in revenue that humans had missed for years.
2. The Valuation Shift: Lemkin noted that for most organizations, the Agentic side of the business could eventually be worth 3x to 4x the software side because the value derived from autonomous work is so much higher than the value of a static software tool.

Salesforce is leaning into this by moving customers to Premium SKUs and selling Consumption Credits (Flex Credits) for customer-facing agents.

The thing to remember is that a lot of these AI tools are very much in the try it before you buy it phase. If these things are as good as advertised then companies will eventually be forced to pay quite a bit more for their use and companies like Salesforce, if they’re still a purveyor of agents within their environment, will be able to make a lot more revenue from that side of the business.

One way to look at is this. Right now, a salesforce seat costs $25-$500 per month. Let’s say you have an employee making $100k using that software and that employee becomes 100% more efficient through the use of agents. Now $200k in salaries becomes $100k in salaries but the reality is that savings is only going to be replaced with higher priced utilization of agents. You still might end up saving money but would it be crazy to think of that company spending $50k on tokens to replace that $100k/year employee in the future? I think that’s the reality we’re moving towards and that’s potentially one where companies like Salesforce can still grow revenue at quite a good clip even if they use a lot of that license revenue.

Right now, Salesforce’s Agentforce segment is a small part of their business and hit $800M in ARR in Q4 2025, growing at 169% Y/Y but that’ll be something that investors will have to see grow to feel better about this transition and Salesforce’s part of it.

Recent Acquisitions: Informatica and the Data Foundation

The acquisition of Informatica was a major theme of the quarter. Why Informatica? Because there is no AI without data.

AI agents are only as good as the data they can access. Informatica provides the data integration and management layer that allows Salesforce’s Data 360 platform to pull in information from Snowflake, Databricks, and even legacy IBM mainframes. In Q4, Informatica’s Cloud ARR hit $1.1 billion. By integrating Informatica, Salesforce ensures that its agents have the context needed to perform real work, creating a moat that pure-play AI startups cannot easily replicate.

Management Q&A: Addressing the Hard Questions

The Q&A portion of the call was particularly revealing. Here are three highlights:

1. Can Salesforce Do Both? (Core Growth vs. AI)

The Question: Keith Weiss (Morgan Stanley) asked if Salesforce can grow a big Agentic business while sustaining momentum in the broader portfolio, noting that organic CRPO was slightly disappointing(meeting expectations seems to be disappointing these days).

The Answer: Marc Benioff and Robin Washington emphasized that AI is incremental to the software. They noted that seat counts are still growing Y/Y, suggesting that AI is currently a productivity layer on top of humans, not a replacement for them. They are using a hybrid model where seats provide the baseline and consumption provides the upside.

2. M&A vs. Buybacks

The Question: Brent Thill (Jefferies) asked why they aren’t leaning harder into M&A given the lower multiples in the market, rather than just buying back stock.

The Answer: Benioff was emphatic: At these prices, the most prudent investment we can make is in Salesforce. He noted that while they will continue to do accretive M&A (like Informatica), the current stock price offers a rare opportunity to retire a massive amount of shares at an attractive valuation.

3. Competition with Model Providers

The Question: Kirk Materne (Evercore) asked about the lines of demarcation between Salesforce and partners like Anthropic or OpenAI who might eventually compete as platforms. The Answer: Benioff argued that while models provide raw intelligence, Salesforce provides the system of work. A model can’t update a customer’s specific sales pipeline or trigger a refund in a compliant, secure, and scalable way without the Salesforce metadata layer. “Our job is to take what’s available today and make it successful for the enterprise,” Benioff said.

Valuation and the Investor’s 10-Year Outlook

For a long-term investor (10+ years), Salesforce presents a fascinating play.

Is it Cheap?

Salesforce is currently trading at some of its most attractive multiples in recent history on a Price-to-FCF basis. With $14.4B in FCF and a market cap around $180B, the company is trading at roughly 12.5x FCF. If you adjust it for the SBC in those numbers, it’s a 16.5X. For a dominant software company with 10%+ growth and a achievable path to $63B in revenue, this is historically cheap but is it cheap enough for investors? Companies like Adobe saw punishment post earnings and kept falling closer to a 10x FCF multiple so it’s possible CRM sees further pressure since this earnings still doesn’t clearly answer whether this company will be succeeding and thriving 10 years from now.

The Pain Ahead

That means there will likely be more volatility. The transition from seat-based revenue to consumption-based revenue is never linear. There will be quarters where the seats might soften before the agents fully take over. Furthermore, the market remains skeptical that GenAI won’t eventually commoditize the basic functions of CRM. All it takes is one news release from Anthropic about their new easily codable and modifiable CRM software and CRM could drop another 10%.

The Verdict

In my mind, the Agentic vision laid out in this call is compelling. Salesforce has:

1. Distribution: 150,000 core customers and 15,000 sales reps.
2. Data: The Informatica acquisition and Data 360 ensure they own the context.
3. Capital: $15B in annual FCF to fund R&D, M&A and buy back a sizable chunk of the company.

For a 10-year investor, who can look past any short term volatility, you get the massive buyback and the critical nature of the core data while you wait for the strategy to play out. The upside is a generational pivot where Salesforce stops being a database and starts being a bit part of the digital workforce. If Marc Benioff is right, and the Agentic business is 3x more valuable than the software business, then today’s fears are simply a gift for those willing to look past the next few quarters.

It all comes down to whether you think these AI tools are going to be as simple to use and deploy, so much so that they just completely replace CRM. Personally, I’m not so sure it’ll be that easy. CRM has a big time lead in the CRM space, tons of data and the underlying relationship to make their product sticky and make it easier for customers to just use agentic offerings through them than try to switch and/or design their own. Plus it all comes down to cost. Right now, AI tools are the new shiny thing and seem amazing because no one is actually charging what they’re truly worth. Once those are fully monetized with an eye towards profit, these seat based licenses and actual employees may actually starting looking good again.

Still, that period of time until we find out what actually happens may mean these SaaS companies are just going to wobble around and be dead money for a while with volatility to both sides. As such, I’d likely wade slowly into a position here and dollar cost average when the stock is attractively priced as it’s starting to get right now.

Disclosure : I may be long stocks discussed in this article in the near future. I am not a financial advisor and do stock analysis as a hobby. Opinions are my own. This is not financial advice. Consult with a financial advisor before making any investments as stock investing comes with risk of loss.