We are reviewing Q1 2026 SaaS company earnings calls. So far we’ve gone through 30+ which include Zeta,Freshworks, Procore, Amplitude, Cloudflare, Datadog, Sprout Social, Waystar,Health Catalyst, Klaviyo, Alkami, Certara, DoubleVerify, Dropbox, Figma, JFrog,Kaltura, ZipRecruiter, Zoominfo, Backblaze, Palantir, Fastly, Expensify,Doximity, Dynatrace, Braze, Snowflake,Zscaler, SentinelOne, UIPath, PagerDuty, Salesforce, Crowdstrike, and MongoDB.
Below are the most salient quotes from their respective CEO’s aboutAI. At the end of the blog are the majorthemes we’re seeing. In summary, AI isresulting in wonderful tailwinds for incumbent enterprise SaaS companies. We will continued updating this blog as weget through each Q1 earnings call for all the publicly traded softwarecompanies.
MongoDB CEO - CJ Desai
AImust have infrastructure to be effective at scale. “Adobe's Journey Agent is a composite multimodal AIagent that unifies Adobe's marketing suite and orchestrates end-to-end customerjourneys, with MongoDB as the agent's long-term memory and reasoning layer.Adobe leverages Atlas Search and Atlas Vector Search together to power thesub-100 millisecond hybrid search the agent needs to act in real time.”
-MongoDB CEO CJ Desai on Q1 2026 earnings call
Crowdstrike CEO – George Kurtz
AImay be software’s best opportunity since covid. “As enterprises rapidly adopt AI, CrowdStrike ismission-critical infrastructure — securing AI across every layer from GPU toagent to prompt. The AI revolution is not a threat to CrowdStrike. It is ourlargest generational growth opportunity. The question is not whether AI createsnew attack surface. It does. The question is who governs it.”
Yetanother example of a software company performing well. “FY26 will go down in our history books asCrowdStrike's best year yet. We achieved $5.25 billion in ending ARR — thefastest and only pure-play cybersecurity software company to achieve thismilestone — driven by a record $1.01 billion of net new ARR, our first yearexceeding $1 billion of net new ARR”
AIis software’s opportunity. “Wehave strong conviction to once again raise our FY27 ARR outlook. The AIrevolution represents a new, generational growth opportunity for CrowdStrike,and we are confident in our ability to deliver durable, profitable growth as wescale to our goal of $20 billion ending ARR in FY36.”
AIis expanding software’s TAM. “Endpointbusiness accelerated for the second consecutive quarter amid AI proliferation.The same AI wave that some said would commoditize security is in fact creatingmore endpoints, more applications, more identities, and more attack surfacethan any prior technology cycle. CrowdStrike is the platform that governs allof it.”
Thehyper-scalers need software. “Nearly $1.5 billion in total contract value wastransacted on the AWS and Microsoft partner marketplaces in Q1 — the largestmarketplace quarter in our history. The hyperscaler partnerships are notdistribution channels. They are co-development relationships that embedCrowdStrike into the AI infrastructure stack at the moment enterprises arebuilding it”
-Crowdstrike CEO George Kurtz on Q1 2026 earnings call
Salesforce CEO - Marc Benioff
WhatAI is doing to software. “It isnot the end of software. It is the end of software that makes humans do all thework. Software that listens. Software that understands. Software that canactually do. And once you see it, there is no going back. Human being, agentdoing. This is the Agentic Enterprise….. If you read any industry news, you might have heardabout the SaaSpocalypse. They are coming for the big SaaS. Disruption fear.Stop getting slammed. It is not the end of software. It is the end of softwarethat makes humans do all the work”
Agenticis software’s biggest opportunity, not threat. “Agentic AI is the biggest growth opportunityfor our customers, and for Salesforce. We're the #1 Agentic CRM, withAgentforce now powering every Customer 360 application and helping tens ofthousands of businesses across every industry transform into AgenticEnterprises. With more than $1 billion in Agentforce ARR, $3.4 billion incombined AI and data ARR, and 3.8 billion Agentic Work Units delivered for ourcustomers, Salesforce has never been more essential.”
PagerDuty CEO – Jennifer Tejada
Incumbentsoftware has the edge. “Wedidn't just start building our AI products. We have been building AI into theplatform for several years. We built a distinct advantage through theproprietary context in our own model that comes from more than a decade ofcapturing insights and information from workflows, developers themselves,events that we ingest, and how incidents are resolved”
NewAI native companies are having issues. “Some ofthese native AI companies are scaling really fast, and with that scale comes asignificant increase in operational complexity. They are seeing more incidents,more alert noise, more infrastructure dependencies than they anticipated. Andthey are coming to us because we have the platform that can handle thatcomplexity.”
TAMfor software has expanded. “Theoperations management market is being fundamentally redefined by AI. Everyenterprise running AI workloads needs a control plane that can manage theoperational complexity those workloads create. PagerDuty is that control plane— and the market we're addressing is significantly larger than the one we'vehistorically competed in”
Theedge incumbent SaaS has. “The PagerDuty Advance SRE Agent represents a stepchange in what operations management can do. It doesn't just alert — ittriages, investigates, and resolves. And because it's built on PagerDuty'sdecade of operational context, it does that with a level of accuracy andjudgment that a generic AI agent cannot match”
Gitlab CEO – Bill Staples
Whatcorporate customers are saying about AI. “In my customer conversations, they are navigatingit like most of the rest of us — which is every day, every week, there seems tobe new innovation, new and exciting opportunities with AI, and they're excitedabout the potential, and they're trying to navigate the many challenges thatcome with it, starting with the privacy and security concerns with theirbusiness, and also the increased costs and expense of investing.”
UIPath CEO – Daniel Dines
AIand SaaS automation co-exist. “AIcreates automation. Sometimes maybe even on the fly. You will run thoseautomations — it is very cheap to run, very deterministic, reliable, auditable— and only when these scripts break you can invoke again AI to fix the scripts”
Softwarecontinues to outperform. “Wedelivered a strong start to fiscal 2027, once again exceeding our guidanceacross all key financial metrics. Before I dive into the results, I want totake a moment to reflect on our progress over the last year. In May of lastyear we launched our agentic and business process orchestration products intogeneral availability.”
Withoutsoftware, AI is rudderless. “I think weall recognize the huge impact of coding agents on the entire ecosystem. AIcreates automation sometimes even on the fly — and UiPath is the platform thatmakes those automations run reliably, at scale, in the enterprise environmentswhere they actually need to work.”
SentinelOne CEO – Tomer Weingarten
Onlysoftware can protect companies from AI attacks. “We are actively pushing the frontier of autonomous,agentic defense across AI, Data, Cloud, and the Endpoint. Enterprises recognizethat securing the AI era requires machine speed defense which only truly moderninfrastructure can deliver, and they are choosing SentinelOne as the foundationto build upon.”
AnotherSaaS company posting a record quarter. “We had a solid start tothe year, highlighted by record net new ARR growth and a landmark milestone asour emerging solutions reached half of our total company ARR. We are activelypushing the frontier of autonomous, agentic defense across AI, Data, Cloud, andthe Endpoint”
Softwarecontinues to innovate for and benefit from AI.“Securingthe AI era requires machine speed defense which only truly moderninfrastructure can deliver. Legacy architectures were not built for autonomousagents operating at scale with elevated permissions. The threat surface haschanged fundamentally, and only platforms built natively for this environmentcan keep pace”
Zscaler CEO – Jay Chaudhry
Agentscreate a new market for software. “Today, users are the weakest link in cybersecurity.But soon, AI agents will be the weakest link. Because they operate at fargreater speed and have far less oversight. Even a single compromised agent canmove from access to data theft in minutes, inflicting catastrophic damage onenterprises”
Agentsreveal new vulnerabilities, that need tobe addressed by software. “Newpowerful frontier AI models like Meta's are finding security vulnerabilities insoftware at machine speed — significantly diminishing the effort, skill, andtime needed to breach enterprises. All enterprises already have thousands ofknown vulnerabilities that they have not been able to patch. Frontier modelsare multiplying these unremediated vulnerabilities by as much as 10x”
Againagents expand the market for software. “We expectit will not be long before millions of AI agents have access to organizations'mission-critical applications and sensitive data. We are now expanding ourexchange to secure AI agents….. We expect AI and frontier models to be one of thestrongest tailwinds our business has ever seen. Our Zero Trust SASE solutionenables us to hide applications and make them invisible to attackers while alsoeliminating lateral movement. With frontier models uncovering vulnerabilitiesat unfathomable speeds and AI agents becoming the weakest link incybersecurity, these differentiators have never been more important than theyare today.”
Snowflake CEO - Sridhar Ramaswamy
Softwareis needed for AI readiness. “AItransformation of workflows and how enterprises interact with their customersis critically dependent on getting their data in a place that's AI-ready. Andthat's where Snowflake comes in. Data that is in Snowflake is increasinglyAI-ready — both for access by cognitive layers like Cortex Analyst or CortexSearch, but also by agentic layers like Snowflake Intelligence.”
Softwaredrives AI readiness. “Datamodernization is just the beginning of the journey. The feedback that I getfrom our customers is that this data modernization journey is even moreimportant than before because they realize that AI transformation is criticallydependent on getting their data in a place that's AI-ready.”
Softwarecompanies have AI product now. “CortexCode compresses migration timelines that once required 18 months and 100-personteams into weeks. We had one partner report 21,000 operations completed in 20days.”
AIworkloads creating sticker shock. “We areseeing some customers work through the economics of AI consumption — what wesometimes call sticker shock — as they discover that AI workloads consumesignificantly more compute than traditional analytics. We are working withthose customers on demand planning to help them realize the value and managethe cost”
Braze CEO – Bill Magnuson
Softwarecompanies have agentic product now. “Theseare not just road map commitments. They are AI solutions that are in the wildand in our customers' hands, earning their place in how some of the world'smost demanding brands engage with their customers.”
Softwareinfrastructure underlies AI that is actually useful to clients. “Brands are moving quickly to transform theirbusinesses with AI and to further leverage their first-party data anddirect-to-consumer relationships. The legacy replacement cycle remained afertile source of new business this quarter, and demonstrated the market'spreference for AI driven solutions paired with high performance first partydata activation. The combination required to execute modern cross channelcustomer engagement at scale.”
Anotherargument that software underlies useful enterprise AI. “Customers across industries and scale are findingthat our AI delivered what general purpose tools cannot. It operates directlyon their first party data, inside their existing workflows, and against theirspecific goals.”
Thehyperscalers buy software. “We achieveda milestone new business win with a prominent AI lab, further expanding ourpresence across high-scale data intensive workloads.”
Dynatrace CEO - Rick McConnell
Softwareis required to take AI from probabilistic to accurate. “Adoptingautonomous operations requires organizations to trust the accuracy of the datathat fuels agents to take action. Deterministic and causal insights fromDynatrace allow our platform to become the system of record, so development andSRE teams and increasingly AI agents can act with confidence to deliver what werefer to as answers, not guesses.”
Existingsoftware architecture makes AI more valuable. “Our advantage is architectural, not feature-based.Dynatrace has been built as a real-time context engine that operates at massivescale across millions of monitored entities and exabytes of data — allconnected and all in real time. By combining deterministic AI with agenticcapabilities, we deliver faster, more accurate insights than approaches thatrely on agentic AI alone. This level of intelligence, speed and efficiencycannot be achieved with point solutions that offer visibility withoutcausality.”
Agentled environments are a new market for software companies. “We believe the winner in observability will be theprovider that can meet the needs of both human-led and agent-led environmentswith a shared system of truth that spans both modes across AI, cloud native andtraditional workloads. This is the moment for which the Dynatrace platform hasbeen built.”
Agentsopen up a new market for software. “AIadds yet another layer to the software stack, increasing the need for moreobservability. Enterprises are deploying new agents, models, orchestrationlayers and agentic architectures that behave differently than traditionalsystems. Their environments generate dramatically more telemetry, connectdecisions across agents and introduce probabilistic behavior that must stilloperate safely, securely and at enterprise scale”
Onlysoftware systems can take AI from guessing to accurate. “What separates Dynatrace's agents from others isthe deterministic foundation underneath — real root cause analysis, anomalydetection and forecasting grounded in Grail. That's not AI that guesses. It'sAI that reasons from facts.”
Doximity CEO - Jeff Tangney
Niceacceleration of AI adoption. “We'rethrilled to announce that we reached a new engagement record of over 800,000active prescribers using our workflow tools in Q4. Nearly half of thoseproviders used our clinical AI last quarter, while our prompts per user nearlydoubled from January to April alone”
Theguardrails for AI in healthcare are very high. “People should be worried about their AI securityand AI chaos. Especially when you're dealing with a lot of patient information.Each of these has been a full process — AI steering committees, exhaustivereviews — but again, we're at 140 health systems now and we continue to grow.”
AIin healthcare needs a human. “Ourphysician-led AI is different. PeerCheck brings credentialed physicians intothe loop to review AI-generated clinical answers before they're surfaced tousers. This is not a feature — it is a foundational architecture decision abouthow AI should work in clinical settings”
AIcan supplement and drive SaaS. “Ournewsfeed had more than 1 million quarterly active prescribers, our workflowproducts had 720,000, and our nascent AI products had over 300,000. Theaddition of AI features across our platform has made us more useful than ever”
Pubmatic CEO – Rajeev Goel
AIis expanding software TAM. “I seeAI as more than just a technical revolution; it is a value chain revolution. Wecan connect the buyer and the advertiser much closer together and increase ourTAM. That will cut across every ad format — not just the high-growth formats weare participating in like CTV, mobile app, or commerce media. Even withindisplay and online video, as we reshape the value chain with agentic execution,there is an opportunity for us to deliver more value across the ecosystem”
Existingsoftware incumbents are adapting. “Agenticadvertising is transforming the industry. It is the third transformation afterRTB and mobile, and I think it will be bigger than those two. It is creatingautomation and workflow that no longer require a software interface and willmaterially change the value chain of our industry. We are scaling Agenticfaster than our peers, and each additional transaction compounds our dataadvantage and drives superior performance.”
Fastly CEO - Kip Compton
AIcreates new revenue opportunities for SaaS. “AI is not a single workload — it is a newarchitectural pattern that touches every layer of the stack. Inference happensat the edge. Agents call APIs at the edge. Security decisions need to be madeat the edge, in real time, before malicious traffic reaches origininfrastructure. Fastly sits at exactly that intersection, and every AI trend weare watching accelerates demand for what we do”
Anotherrecord quarter for a SaaS company. “Our firstquarter performance demonstrates continued discipline and velocity as wedelivered record revenue, gross margin, and RPO. Driven by an acceleratedinnovation roadmap, we delivered 47% year-over-year security revenue growth”
Anotherexample of AI creating an opportunity for SaaS. “AI agents generate traffic patterns thattraditional security tools were not built to handle. An agent making API callslooks fundamentally different from a human browser session — differentfrequencies, different payloads, different behavioral signatures. Our edgesecurity layer is where those decisions need to be made, in real time, beforethat traffic reaches your origin. That is a new and growing securityrequirement that did not exist two years ago”
Aconcrete example of new AI markets expanding SaaS economics. “Net retention rate improved to 113% from 100% ayear ago. That 13-point improvement in twelve months reflects customersexpanding their use of Fastly as AI workloads grow — not just renewing, butdeepening their platform commitment”
AIis driving SaaS performance. “We raisedfull-year revenue and operating margin guidance. The AI tailwinds we describedare not theoretical — they are showing up in our RPO, in our compute revenue,in our security bookings, and in our NRR. We are investing ahead of demandbecause we believe the demand is real and durable”
Palantir CEO – Alex Karp
Softwareis the backbone of usable AI. “We arefocused on enterprise autonomy, not on dazzling demos. We have, in theontology, the no-slop zone. The ontology is the body to the AI brains. Youcannot actually interact with the enterprise or affect the world; your agentscan go nowhere without ontology. And you are seeing that with our customers.”
Anextraordinary comment. “Almostevery single highlighted example of AI that actually is producing results inthe U.S. is actually Palantir. And one of the ways to pen test what we aresaying is just dig into the examples of AI actually transforming an enterprise.Call the client. Talk to them. They look the same to nontechnical people, butthey do not look the same to practitioners — whether you are on thebattlefield, or whether you are an insurance company, or whether you are ahospital, or whether you are a manufacturer — what they discover is the realityof doing this requires a platform like ontology, currently executed on top ofFoundry with FDEs. And currently, that combination is available from onecompany, and that is us”
Palantirwants you to try and use AI without software guardrails. “The best thing that can happen to thiscompany — and maybe this country — is, of course, they should go out and flirtwith all this slop. Mostly they come home to Palantir. They do not have to allcome home to Palantir. We have limits. But go test it out. Go see how easy itis to make these things work”
Softwareis the picks and shovels of AI. “There seemsto be a rotation amongst AI model companies who engage in an intenselycompetitive race in which we have seen token costs suffer a thousandfolddecline over just a few years and where winners and losers swap places everysix months. Our path has been different, building a juggernaut of a businessthat is delivering results to our partners in the world as it is today.”
Backblaze CEO – Gleb Budman
Thisis going to drive storage demand. “As AImodels move from text to multimodal, incorporating video, audio, and images,the volume of data required to train and run those models grows by orders ofmagnitude.”
Anotherexample of AI expanding TAM for software co’s. “The Neocloud market was $25 billion in 2025and could grow to about $400 billion by 2031. Backblaze's opportunity tosupport Neoclouds is $14 billion by 2030. We are engaged with most of the topNeoclouds, and we characterize our role as providing the data lake layer thatsits behind lower-latency flash storage near the GPUs.”
WhyAI won’t be used to replicate enterprise software. “It is a little bit like the question we werealways asked for almost two decades: what happens if AWS lowers their price tomatch you? We are two decades in, and that has not happened. The challenge isit is not easy to build the type of IP that we have built over the last twodecades. It requires scale and expertise and focus over a long period of time.For the Neoclouds, spending all their resources to try to replicate what wehave built is probably not the best place for them to invest when time to valueis so much faster by using Backblaze.”
Softwarewill benefit from serving AI loads. “AIcustomers are growing about three times faster than our average customer, andthat's just a function of their inherent data growth driven by their AI usecases”
Anotherexample of SaaS having excellent earnings this Q. “Q1 was a proof point. We beat on revenue, beat onEBITDA, B2 is growing 24%. The deal size has more than doubled. The AI customeris up 76% year-over-year. We are not just riding the AI wave, we're buildingthe infrastructure that supports it. We are key for the Neoclouds, key for theAI builders, and we've had nearly two decades of optimizing performance perdollar at scale, which makes us ideal for the needs of AI. We raised guidance.We're on track for our first full year of free cash flow positivity as a publiccompany, and we're picking up steam”
Zoominfo CEO - Henry Schuck
AIhas caused a pause in some software purchasing decisions. “Asmacro conditions worsened at the end of the quarter, we experienced aregression in our downmarket and upmarket growth trajectory. We are seeingcustomer confusion regarding AI and product differentiation, leading to a pausein purchasing decisions — especially among software customers facing aconfusing purchasing landscape compounded by the threat of their own growthdisruption.”
Morecustomer confusion. “We saw someheadwinds in Q1 particularly from our software vertical customers who areseeing some confusion in the marketplace about what to buy and what not to buygiven AI”
Ziprecruiter CEO – Ian Sigel
AIshould be used to accelerate the product roadmap, not cuts. “AI is permeating really every departmentwithin our company. It is driving extraordinary efficiencies across the board,which we have not looked at as a cost-saving opportunity as much as we havelooked at it as a mechanism by which we can realize and increase our ambition.So if anything, it has accelerated our roadmap as opposed to saved us money.”
LLM Marketing is the next SEO. “Thenew app went live on ChatGPT and on a percentage basis, the growth of LLMs ingeneral has been impressive as a new traffic source. Overall, LLMs stillrepresent a tiny contributor in the overall mix of where traffic comes from toZipRecruiter. But it is good to be there at the beginning and to enjoy the rideup with them as they become an ever more popular way for job seekers to lookfor work.”
JFrog CEO – Ben Haim
AIneeds software. “As AIgenerates more code, more models, and more artifacts at a speed no human teamcan manually govern, the need for a universal platform that tracks provenance,enforces policy, and ensures that only trusted software reaches productionbecomes existential”
AILabs can’t replicate many core enterprise needs. “AI labs may offer binary reverse engineering, butthey do not replace the need for a universal governance tool and system ofrecord. No single vendor's tools can govern artifacts across every language,every framework, every cloud, and every AI model provider simultaneously. Thatuniversality is what JFrog provides.” Acompany running Python, Java, and Go microservices across AWS and Azure,pulling from three different AI model providers, needs one governance layerthat spans all of it. No AI lab can provide that.
MoreAI code means the need for more SaaS tools. “AI coding tools are dramatically increasing therate at which developers pull open-source dependencies. AI-generated code tendsto rely heavily on external libraries. As that dependency rate increases, theattack surface for supply chain attacks grows proportionally — and the value ofJFrog Curation grows with it.”
Figma CEO – Dylan Fields
AIneeds humans. “More designtools are launching, more people are creating, more software is now being builtthan ever before. And in this world where bits are abundant, what's scarce ishuman creativity, actual point of view, care and craft and judgment. This iswhat makes a product, a company, a campaign cut through the noise.”
Thehyper-scalers need software. “Google is along-time Figma customer. Many of their most iconic products have been designedand built on our platform. As they build the next generation of AI-nativeproducts, they're doubling down on Figma. The team designing agentic Geminiexperiences for millions of enterprise customers uses Figma end to end, astheir single source of truth, from the earliest concept work all the waythrough to shipping.”
Institutional knowledge is edge. “We expect more tools to emerge, somecomplementary and others competitive. Figma focuses on deep understanding ofdesigner workflows and quality, shipping faster than ever. We aim to integratewith complementary tools and maintain strong partnerships”
Dorpbox CEO – Drew Houston
Nosoftware company is standing still. “We areevolving from file storage to AI-powered content management. Dash bringscontent from Dropbox and other cloud applications into a single experience —and we've built a context engine that gathers context across content and appsand connects it to AI models for faster and more accurate results”
Existingcustomer base gives incumbent SaaS a huge headstart on AI product rollouts. “18 million subscribers is our home-fieldadvantage. We are investing heavily in deeply integrating Dash into the coreDropbox experience — and that's where we have the most leverage.”
Butsome AI rollouts are not going great. “Morethan 30% of weekly engaged users used Dash's AI features again the followingweek, and more than 50% of monthly engaged users used them again the followingmonth. Retention patterns have been stable as we expand access to new cohorts.”
Openinga new market for old school SaaS. “Protectpositions Dropbox in the security and governance conversation for the firsttime. In the AI era, where employees are sharing sensitive content withexternal AI tools in ways that IT cannot see or control, content governancebecomes a security imperative — and that's a new budget category for us.”
DoubleVerify CEO - Mark Zagorski
AIopening new markets for incumbent SaaS. “Slop Stopper is now being applied on themeasurement side to about 40% of all of our impressions — one of the fastestscaling attach rates we have seen.”
Againon more markets. “AI isdriving demand for our solutions in three distinct ways: it is generating morecontent that requires verification, it is powering more sophisticated fraudthat requires detection, and it is enabling more efficient advertising thatrequires measurement. Every AI development that creates risk for our customerscreates opportunity for DoubleVerify”
AIis driving more fraud, fast. “AI-poweredbot scheme variants increased 140% in the first quarter of 2026 compared withthe first quarter of 2025”
Certara CEO - Jon Resnick
TheAI use case is velocity. “AIdoesn't replace biosimulation — it supercharges it. Our models are built on 30years of validated pharmacokinetic and pharmacodynamic science. An LLM trainedon general scientific literature cannot replicate that. What AI can do is makeour scientists dramatically more productive, reduce the manual steps betweensimulation runs, and allow us to explore more of the design space in a fractionof the time. The result is better drug development decisions, faster — and thatis what our customers are paying for.”
AIopens a new market. “We want todemocratize biosimulation. Today, the most sophisticated quantitativepharmacology is accessible only to large pharmaceutical companies with deepscientific teams. AI-assisted biosimulation can make those capabilitiesavailable to biotech startups and smaller organizations that currently cannotafford or staff a full modeling and simulation function”
AIalso is forcing a restructuring. “Regulatorywriting and medical writing are areas where AI-assisted tools are advancingrapidly. By divesting those businesses, we are reallocating resources to ourcore biosimulation platform — the part of our business that AI cannot replicateand where our validated scientific models create the most durable competitiveadvantage.”
Alkami CEO – Alex Shootman
Incumbentsare trusted. “Onceon the Alkami platform, our investments in service and reliability, themission-critical nature of our platform, and high switching costs drive grossretention rates 8 to 10 points above typical SaaS companies”
Thedata moat. We havehad 39 client meetings since February where AI has been on the agenda, andevery one of those clients has a specific, concrete use case they want topursue. This is not exploratory conversation anymore. These are financialinstitutions that have identified workflows they want to automate, complianceprocesses they want to accelerate, and member experiences they want to improve— and they want to do it on our platform because their data is already here.”
Softwareside is still wonderful. “Every fiveyears, our clients grow by more than 100% of their original platforminvestment, with our 2021 through 2023 cohorts spending above 2x their landingARR and clients 2016 and older spending close to 4x their landing ARR”
Klaviyo CEO – Andrew Bialecki
Data-lessAI is useless. “Deliveringmeaningful customer experiences at scale requires AI grounded in real data.Agents are only as good as the systems beneath them, and we've spent 14 yearsbuilding exactly that foundation”
WhatAI means for software. “Wehave entered the era of agents and infrastructure, at least so far as softwareis concerned.”
Softwarethat is revenue facing always wins. “Unlike mostenterprise software, we are focused on revenue generation. If you can help growtop line and profits, there is insatiable appetite to spend more, and we seethat constantly.”
AImakes the software customer smarter. “Whenwe think about Composer, ultimately what we are providing is intelligence,delivered in the form factor of tokens. Customers are using it to review whotheir customers are, the effectiveness of their marketing, and then figure outwhat to do next. Those sessions are incredibly valuable — generating thousands,even hundreds of thousands of dollars in incremental revenue. We can go muchdeeper because our agents have access to internal benchmarks and best practicesthat are not publicly available. That intelligence is an entirely new revenuestream.”
Serviceproviders have a place. “Whatyou can do with Klaviyo and the infrastructure that we built is actually a lotmore than our customers and businesses are taking advantage of. Our agencieshave always helped close that gap. Now, with our agents — Composer and CustomerAgent — agencies are accelerating the adoption of both”
Agenttiming. “Thecompany anticipates that AI agents will become standard and ubiquitous withinthe year, creating a positive feedback loop where agent usage generates datathat further improves the underlying infrastructure.”
HealthCatalyst CEO – Ben Albert
Moredata moat arguments. “Wehave 18 years of proprietary healthcare improvement data — the largest suchrepository in the world — and we are building AI models on top of thatfoundation that no competitor can replicate. AgenTeq is not a generic AI layer.It is intelligence built from the specific clinical, operational, and financialpatterns of healthcare delivery at scale.”
Internaldev performance is doubling. “Ourinitial pilots using new development pods and proprietary AI agents resulted inup to 100% more story points delivered per developer. We are applying the sameintelligence we are building for clients to our own engineering operations.”
GenericAI won’t suffice. “Generic AIcan tell you what the literature says about reducing readmissions. Our modelscan tell a specific hospital which interventions have worked for patients withtheir specific case mix, in their specific market, given their specificoperational constraints — because we have seen those patterns play out acrosshundreds of health systems over 18 years”
AImakes incumbent software even more powerful. “Health Catalyst exists to be theintelligence layer of the American health system. AI makes that mission moreachievable than ever — and more urgent than ever.”
Waystar CEO – Matt Hawkins
Somesoftware companies see an expanding TAM with AI. “That shift unlocks a much larger opportunity— the approximately $100 billion in annual revenue cycle labor servicesperformed across the industry today. We believe we are well-positioned toautomate a meaningful portion of this labor pool through new AI-poweredcapability launches like denials, prior authorization, and recoupment. We'rebuilding toward what we believe is the future of this industry: the autonomousrevenue cycle platform.”
AIis increasing bookings velocity. “AItraction is accelerating. AI-powered capabilities drove roughly 40% of newbookings in Q1. Our clients leaned into the platform for prevention,automation, and visibility rather than downstream rework.”
AIwill lead to tool consolidation. “I waswith a CIO of a very large system not long ago — very impressive lady. Shesaid, 'Matt, can you help us? I currently use more than 12 point solutions justto manage our revenue cycle process.' And of course, that's where the platformapproach really comes to play.”
SproutSocial – CEO Justyn Howard
Onwhy Sprout Social is doing a $50mm share repurchase: “Wehave a meaningful disconnect between current valuation levels and the long-termvalue we expect to create....We are seeing customers making longer-termcommitments to Sprout with multiyear contracts now representing nearly half ofour contract mix, up from about a third two years ago. This reflects growingconfidence in Sprout as a strategic platform.”
Cloudflare – CEO Matt Prince
Theydid a 20% RIF while achieving 34% YOYG. “Justbecause you are fit does not mean you cannot get fitter. Over the last sixmonths especially, the productivity gains from the people directly talking tocustomers and directly creating code have been incredible, and a lot of thesupport roles behind them are not going to be the roles that drive companiesgoing forward.”
Oninternal use of AI: “internal AI usage increased more than 600% in asingle quarter….97% of Cloudflare's R&D employees use AI coding tools….withproductivity gains of "two, ten, even 100 times"
Softwareis the backbone of AI: "Cloudflare is now processing hundreds ofbillions of agentic requests per month" according to their latest earningscall. Every AI agent calling a tool,querying an API, or taking an action on the internet is generating a requestthat may pass through Cloudflare's network.
Datadog – CEO Olivier Pomel
AItraining is becoming production infrastructure. “AI training was very new a couple of years ago. Itwas something that was only done by very few companies, and it was, in a way,very artisanal — it was not a production workload. It was something thatresearchers were building, and it was very one-off. And now it's turning intoproduction. It's turning into something that many more companies are doing.It's scaling by orders of magnitude, and it's becoming something that has to beon all the time, reliable — and every minute you lose, or every failure you havein your training round, is a week you give away to the competition.”
Thedual strategy. “We'retalking about our AI efforts in two buckets: AI for Datadog and Datadog for AI.AI for Datadog — these are AI products and capabilities that make the Datadogplatform better and more useful for our customers. Datadog for AI — thisincludes Datadog capabilities that deliver end-to-end observability andsecurity across the AI stack.”
Hoursto seconds. “We'veseen Bits AI Security Analyst reduce investigations that could take hours to aslittle as 30 seconds.”
Thecloud still matters. “Weshowed broad-based acceleration of revenue growth across cohorts, includingboth our AI and non-AI customers. Non-AI customer revenue growth acceleratedagain this quarter — we think this is a sign of strong continued cloudmigration, greater adoption of our products, and customers of all kindsaccelerating their use of AI.”
Amplitude
Themost honest confession from any CEO we’ve seen. “It is much more important to get there with a lotof speed for a lot of different reasons than it is to say, 'Hey, let's try toprotect some existing thing we have.' The existing thing we have, frankly,isn't valued much.”
Thegoal of Amplitude’s AI. “We'reentering a new era of analytics — one where AI can monitor your product aroundthe clock, and free up your team to focus on improving the experience. The realadvantage is how quickly a team can learn, iterate, improve, and automate.Agentic analytics is the key.”
AIgenerated code means you need more software. “AI has dramatically lowered the barrier to buildingand shipping software, boosting productivity for experienced engineers andenabling non-traditional roles to become AI builders. While teams can generatemore code than ever before, the software development life cycle remainsbottlenecked in many other places. AI builders are generating code faster thanthey can understand its impact. The challenge is now evaluating code beforeit's released, tracking what's working after release, knowing when you need toroll things back, and turning behavioral signals into what to build next.”
Ontheir own coding. “Over90% of the code our team ships today is written by AI.”
Onthe internal adoption of AI. “Wepaused normal work across the entire company so that every function could buildand ship AI-powered workflows to reimagine their daily jobs and functions. Theteam shipped hundreds of amazing demos, including automatically creating customdemo websites per customer, automating part of the quarter close process, andautomating how we create new creative assets in marketing.”
Grossmargin is compressing, but that means…“The inference cost growth is a leading indicator ofhow much our customers are actually using our AI tools. That is what we want.”
Zeta
AI’simpact on the market now. “AI isno longer a feature. It is driving a replacement cycle where enterprises aredemanding fewer systems, measurable results, and applied intelligence thatworks today.”
Wecan’t just talk about AI anymore. “Customerswant to invest in applied AI, not road map AI.”
Thedata and software flywheel. “Asadoption increases, Athena learns from more data, outcomes improve and usagedeepens, driving ARPU expansion and ultimately reinforcing the same flywheelthat has powered our growth. That flywheel is powered by more than just Zeta'sAI models. It's driven by the data and infrastructure behind them.”
Thedata moat. “ZetaSuperGraph, our proprietary identity and intelligence graph, unifies dataacross the enterprise and enables a complete deterministic view of the consumerthat we believe is difficult to replicate at scale.”
Freshworks
TheAI conversation is about impact. "AIis not causing sales delays — it has become a core requirement in nearly alllarge enterprise deal discussions. Customers are not asking whether we have AI.They are asking which AI, how it works, and what outcomes it delivers.Freshworks is winning those conversations because Freddy AI is embedded in theplatform, not bolted on."
AIis the last thing listed. "We had our biggest deal ever — a largenutrition company that was a 10-year customer of one of our competitors — thatis moving over to us for all the reasons we have discussed: enterprise-gradescale, much faster time to value, easier to manage the platform, and AIcapabilities."
Procore
Abig advantage of the incumbents. “Procore AI is a reasoning engine purpose-built forconstruction. It understands the language and logic of the project — how an RFIconnects to a submittal, how a submittal connects to a drawing, how a changeorder gets approved. It works as a multi-step system that holds context acrossmultiple steps. It doesn't just answer a question — it understands the thread.”
Thedata moat. “Thisdata and context can only be accessed within a system of record andcollaboration like Procore. Building on our flagship system of collaborationwith nearly 3 million active users and a massive proprietary dynamic data set,Procore AI can deliver outcomes simply not possible with traditional software.”
Areal ROI use case. “Byleveraging Procore AI, Crest has taken a manual process that could span weeksof effort down to an automation that can take as little as 20 minutes. Thisisn't just an incremental improvement in speed. It is a fundamental shift intheir competitive advantage.”
Thereis more to come. “It is still early. As we continue to developProcore AI, going deeper into our proprietary data and broader across projecttypes, the reasoning engine will only become more capable. We expect oursolution to continue to improve with every layer we unlock, and we have a longrunway ahead of us.”
Our big take-aways are below.
Theme 1. AI conversations in enterprise software are nowabout actual performance. Zeta's CEO put it well when he said "customerswant to invest in applied AI, not road map AI". At the same time, AI without systems,workflows and data is slop, as Palantir's CEO alluded to. The model is thecommodity while the data and underlying workflow layer from enterprise softwareare what matters. As Klaviyo's CEO putit: "agents are only as good as the systems beneath them" andDynatrace’s CEO refers to customer requirementsfor “answers, not guesses.” Incumbentslike Braze, Snowflake, and JFrog are benefitting from thoughtful architecturewhich allows enterprises to have real use cases for AI.
Theme 2. CEOs are being candid about AI's impact on theirorganizations. For instance Amplitude's CEO said "the existing thing wehave, frankly, isn't valued much" and paused the entire company to rebuildworkflows from scratch. Others, likePalantir's CEO sounded angry on their calls about the disrespect for software. No software company is standing still, andbetting against software’s evolution and value is to bet against technologistslike Karp.
Theme 3. AI is splitting companies into two camps: i) you’reeither using AI as an excuse to cut headcount (looking at you Cloudflare,Freshworks, Health Catalyst, Block); or ii) you’re using it to drive theproduct roadmap faster (ZipRecruiter, Figma, Datadog). Bet on those in the second category. Those in category one sound like they'recoyly still shedding employees from covid over-hiring.
Theme 4. One of the things we didn’t appreciate until now isthe TAM expansion argument. Agents that access software are now users, and arebeing charged as such. Traditionalservices TAMs are also becoming markets software can serve with their ownagents. AI proliferation thattheoretically threatens to commoditize software is creating new markets for it. Theme 5 expands on this below.
Theme 5. AI issimultaneously creating more of the problems that incumbent software solves whichexpands the market for solutions. Forinstance, AI generated content requires more brand safety software (DoubleVerify), AI coding tools need more observability software (Datadog, Dynatrace),AI agents need more permissions and governance that software provides(Sailpoint), more agentic traffic across the internet means more edgeinfrastructure from incumbent software (Fastly, Cloudflare). The same AI wave that allegedly threatensthese businesses is actually driving their product offerings.
Theme 6. We’re seeinglots of repositioning; MongoDB is repositioning as "the memory layer forAI agents." PagerDuty is repositioning as "the control plane for AIoperations." ZoomInfo is repositioning as a data infrastructure layerrather than a seat-based platform. Oneway to make the market understand software’s importance is to repositionit. Financial performance largelysupports the repositioning as most ofthese companies are posting record quarters.
Thank you for your readership. Seemore blogs and SaaS data at blossomstreetventures.com. Email the author atsammy@blossomstreetventures.com.