Answer in 30 seconds
The recent earnings reactions across Braze, Klaviyo, Snowflake and Salesforce suggest that markets are not simply rewarding companies for adding AI or maintaining SaaS growth. Braze and Klaviyo were punished despite 26% revenue growth as investors focused on forward profitability, margins and evidence that AI investment is translating into economics. By contrast, Snowflake surged after 37% product revenue growth, raised guidance and explicit AI-driven acceleration, while Salesforce rallied after record results and rapid Agentforce and Data 360 ARR growth. Fuse's view: as AI makes execution cheaper, durable value moves towards proprietary customer context, governed data, decisioning, orchestration and closed-loop learning.
Braze reported a strong quarter. Revenue reached $227.2 million, up 26% year on year. Adjusted earnings per share came in at $0.19, ahead of expectations. Revenue guidance for the next quarter was also above consensus. The stock still fell roughly 19% during the following session.
A month earlier, Klaviyo reported almost the same top-line growth rate: $370.6 million of Q2 revenue, up 26%, ahead of expectations, and it raised full-year revenue guidance. Its shares also sold off sharply.
At first glance this looks absurd. Growth is strong. Both companies are investing aggressively in AI. Customer engagement is hardly becoming less important. So why is the market reacting as if something is wrong?
The answer is more interesting than a bad earnings day.
The immediate explanation for Braze is straightforward. Investors looked past the Q2 beat and focused on Q3 adjusted EPS guidance of $0.13 to $0.14, below the $0.16 consensus. They also focused on slower remaining performance obligation growth, gross-margin pressure and the cost of continued investment. Braze's AI adoption may be promising, but investors want to see it become visible in the financial model.
Klaviyo produced a similar tension. Demand remained resilient and AI-agent adoption was part of the growth story, but higher messaging costs and product investment put pressure on margins and the operating-income outlook. Again, the market was not saying there was no growth. It was asking what that growth will cost and what the next increment of software value looks like.
But Snowflake breaks the easy version of this story
It is tempting to say that markets are simply punishing SaaS companies while they work out the AI transition. Snowflake shows why that is too simplistic.
Snowflake's latest quarter was rewarded, not punished. Q2 FY27 product revenue grew 37% to $1.49 billion. The company lifted full-year product revenue guidance to $6.07 billion and raised its operating-margin outlook. Its shares jumped more than 20% in extended trading. Management said AI products accounted for roughly half of the recent growth acceleration.
That is an important counterexample. The market appears perfectly willing to reward AI investment when it can see AI translating into accelerating consumption, stronger guidance and improving economics.
Salesforce is another useful counterexample
The legacy incumbent is not standing still. Salesforce is reorganising its product architecture around Agentforce, Data 360 and agentic applications. In Q2 FY27, total revenue grew 11%, non-GAAP operating margin reached 34.1%, and Agentforce plus Data 360 ARR reached nearly $3.9 billion, up more than 210% year on year. Salesforce shares rallied strongly after the results.
That matters for anyone tempted to frame this as modern platforms versus legacy platforms. Salesforce is trying to use its installed base, customer data, workflow footprint and distribution as advantages in the AI transition. Marketing Cloud is increasingly being pulled into a broader agentic architecture rather than treated as a standalone campaign machine.
The market is not asking: do you have AI?
Almost every serious software company will have AI. Agents will build segments, generate creative, configure journeys, analyse results, write queries and recommend next actions. The existence of those capabilities will rapidly stop being differentiating.
The harder question is: does AI improve the economics and strategic position of the software business? Does it accelerate growth? Increase consumption? Deepen the customer relationship? Improve margins over time? Create a new monetisable product surface? Or does it simply make capabilities that customers already pay for cheaper and easier to reproduce?
That is why the contrast between these earnings reactions matters. Braze and Klaviyo are growing quickly, but investors are scrutinising the cost of the next phase and waiting for clearer AI monetisation. Snowflake showed AI contributing to acceleration. Salesforce showed material AI and data ARR alongside very strong profitability. The reactions were different because the evidence was different.
What happens to customer engagement software when execution gets cheap?
The first generation of marketing technology was largely a system of execution: segment, campaign, journey, channel, report. A huge amount of value sat in the interface and the skilled human labour required to operate it.
Modern platforms such as Braze and Klaviyo moved the category towards real-time behaviour, cross-channel orchestration and much richer customer context. AI now pushes the model again.
If an agent can build the audience, write the message, configure the journey and analyse the result, the scarce asset is no longer the ability to perform each of those tasks. The scarce assets become the context the agent can reason over, the governed customer state it can trust, the decisioning logic, the ability to act across channels, and the feedback loop that tells the system whether the action worked.
The customer technology value shift
We think the emerging architecture looks less like a stack of campaign tools and more like a learning system: customer state → context → intelligence → decision → orchestration → experience → customer response → updated customer state.
Databricks and Snowflake are fighting to own more of the governed data, context and intelligence layer. Salesforce is trying to connect data, applications and agents across an enormous installed base. Braze and Klaviyo are pushing intelligence deeper into customer orchestration. The boundaries between those categories are becoming less comfortable.
That does not mean orchestration becomes unimportant. It means orchestration has to become part of a system that learns.
The question we would ask every customer technology vendor
What compounds?
If every new customer interaction leaves the system with better context, better decisions, better models or a more defensible position in the customer's architecture, AI can expand the value of the platform. If AI mainly automates clicks inside an existing interface, the customer benefits, but the vendor may discover that some of what it used to monetise is becoming a commodity.
That is the tension sitting underneath these earnings reactions.
What this means for customer leaders
Do not choose your next customer platform because it has the best AI demo. Those demos will converge.
Ask where customer truth lives. Ask how identity and consent are governed. Ask what context an agent can actually use. Ask who owns decisioning. Ask how an AI-generated decision reaches the customer. Ask what comes back after the interaction. And ask whether the system gets smarter as a result.
The winners in customer technology will not necessarily be the companies with the most AI features. They will be the companies that occupy a difficult-to-remove position in the customer learning loop.
Fuse view
Braze's sell-off does not convince us that customer engagement software is becoming less important. If anything, AI makes the ability to understand and act on customer context more important.
But it may change where the value sits. Sending the message gets cheaper. Building the segment gets cheaper. Producing the creative gets cheaper. Knowing the customer, deciding what should happen, executing safely across the customer experience and learning from the outcome become more valuable.
For brands, that means the strategic question is no longer simply which platform should we buy? It is: what customer learning system are we building, and which parts of it will compound as AI improves?
