Salesforce (NYSE: CRM) reported $11.35 billion in Q2 FY2027 revenue and $5.90 in adjusted EPS, beating Wall Street estimates of $11.32 billion and $5.25, respectively.
I’ve been watching Salesforce through the same lens since Q1, when revenue reached $11.13 billion, and EPS came in at $3.88, both ahead of expectations, yet CRM finished the session down 0.88%.
This time, investors reacted very differently, sending the stock up 22.6% after earnings as the market finally began to pay attention to what customers are doing on the platform.
That reaction gives us a much better starting point than another earnings beat, because while Salesforce spent Q1 trying to convince investors that AI would strengthen its position rather than destroy it. Q2 gives us a much clearer look at whether customers are actually behaving that way.
The AI Fears Are Running Into Customer Behavior
Salesforce has spent months fighting the idea that better AI models would make traditional enterprise software less necessary, but the customer numbers coming out of Q2 make that argument harder to sustain.
Agentforce ARR crossed $1.5 billion, up more than 240% year over year, while Agentforce and Data 360 reached nearly $3.9 billion in combined ARR, up more than 210%.
Then the usage numbers become more interesting: customers generated 3.2 billion Agentic Work Units during Q2, up 97% from the prior quarter. Another 2,000 customers put agents into production, up 70% sequentially. Agentforce One Edition and Agentforce for Apps bookings more than doubled quarter over quarter.
But one number stood out among others – half of Agentforce bookings came from customers replenishing consumed credits. That sounds simple, but it gets right into the heart of the AI debate. A customer testing an AI product once tells me very little. But a customer burns through its credits and comes back for more? That’s the biggest tell that the product has entered the workflow. Which is exactly where Salesforce’s existing position becomes useful.
The Moat Looks Different When AI Needs What Salesforce Already Owns
Large companies have spent years stuffing their customer data, workflows, permissions and business rules into Salesforce and ripping that machinery out remains a much bigger job than replacing a chatbot.
The Q2 transcript gives us a real-world example through Uber Technologies (NYSE: UBER) for Business. The company pointed Agentforce at its existing Salesforce data and workflows, launched within six weeks and generated 60% more leads within two weeks.
Salesforce also says nine of the world’s top 10 AI companies use Salesforce and Slack, with their spending up 435% year over year. That creates an interesting reversal of the SaaS apocalypse argument, because the companies building the new AI economy still need enterprise data, permissions, workflows and business context to make those models useful within a real company.
Salesforce already runs much of that infrastructure, and the company is pushing the AI layer directly into it rather than asking customers to start over elsewhere. This is why the result could look less like AI replacing Salesforce and more like AI making the Salesforce installation customers already paid for more valuable.
Numbers Are Starting To Back Up The Reacceleration
Salesforce now has something it lacked when investors punished the stock after Q1, because management can point to stronger new-business activity while customer attrition sits near record lows.
Management said first-half net new annual order value growth significantly outpaced annual order value growth, which it believes puts the company on track for second-half organic revenue reacceleration. Q2 cRPO reached $33.5 billion, up 14% year over year and 14% in constant currency.
Salesforce also raised FY27 revenue guidance to $46.1 billion to $46.4 billion, with the $200 million increase consisting of $100 million from organic growth and $200 million from the pending Contentful and Fin acquisitions, partly offset by a $100 million foreign-exchange headwind.
Management expects Q3 revenue of $11.42 billion to $11.5 billion and cRPO growth of about 14%. Meanwhile, Salesforce produced $1.1 billion in free cash flow during Q2, up 81% year over year, while operating cash flow reached $1.3 billion, up 71%.
None of this says the risk has disappeared. In fact, Salesforce still needs to turn AI adoption into durable organic growth, and the company carries plenty of expectations after this move. If that reacceleration fails to materialize, the stock can give some of this rally back.
CRM Finally Looks Like The Market Believes It
One look at the chart and it tells the story better than another paragraph about management confidence ever could, because CRM spent months trapped below its major moving averages while investors waited for the business to prove itself.
After Q1, the stock sat around $177.51, below the declining 50-day and 200-day averages, while repeatedly defending the mid-$150s.
This quarter, CRM has pushed above the 20-day moving average near $203.58, the 50-day near $180.89, and the 200-day near $200.32, reaching roughly $256.48 after the earnings move. Volume also exploded to roughly 34.36 million shares on the earnings session. That is a completely different market.
I Would Rather Own The Evidence Than The Story
I like CRM here because the market finally has something tangible to argue about beyond whether AI will kill enterprise software.
The bull case now rests on customers using Agentforce, expanding their spending and leaning harder into a platform they already know. The bear still has a path if organic growth refuses to accelerate or the AI spending fails to translate into enough revenue.
But after watching this stock get punished while the business kept building, I would rather follow what customers are doing than what the loudest AI skeptics think should happen.
