We are in the midst of the Q2 earnings season for SaaS companies, and so far 16 companies we follow have reported. Another ~55 companies have not yet reported, and we will update this post as they do.
So how is enterprise SaaS doing? Fantastic. Only one of the 16 is struggling (SPS Commerce), whereas all others are beating guidance and setting records. Below are the most salient quotes from the CEO’s about AI. At the end of this article are the key themes we are seeing.
Palantir CEO Alex Karp
General AI presents a data governance and security risk. “People sign up for token self-pleasurings, and that at a real cost like other forms of self-pleasuring, where you are paying for the right for them to migrate your IP, your know-how, your expertise to their model so that they can build a competitive business that doesn’t require your business, your people. Why are they doing it? It’s actually being done for what they believe are moral reasons. They are superior to you. They deserve to colonize your enterprise. You deserve to be colonized….. token maxing is at their own cost, and they certainly understand that token maxing is leading to them transferring their data, their prompts, the way they run their business, their expertise to a third party.”
On being the best. “We are outsiders. Outsiders like you’re an outsider. You come to this country, you better have really good results. Like the same thing for Palantir. We know we need the best results because people aren’t buying our product because we’re swinging the golf club correctly or paying for the steak dinner. They don’t even invite us to steak dinners.”
AI cannot stand alone “AI alone cannot do that. It takes AI in the hands of the American worker, the tribal knowledge earned through success and failure on the line, the insights only they have. The models are commodities, the American worker is not.”
AI needs governance from software. “Our Q2 results are unprecedented but entirely unsurprising, as the abrupt market shift in LLMs that we’ve been warning you about for years is now here. Enterprises that are not using Palantir are seeing their token meters spinning endlessly just to get slop without any correlation to value. This token model may be working for the labs, but it is not working for anyone else. It’s breaking corporate budgets without results to justify the expense.”
Zeta CEO David Steinberg
SaaS incumbents are not standing still. “Investors have historically viewed Zeta as a market technology company. Marketing is where our platform was first applied and where we built our leadership position. That description no longer fully encapsulates who we are today. Zeta has evolved into an intelligent AI infrastructure platform. Marketing is our first application, not our limit.”
History is a moat. “The foundation is our Zeta Data Cloud, built on proprietary data covering more than 535 million individuals globally, trillions of signals, and 20 years of model tuning. To be clear, this data is owned, not rented, because nothing you rent can be a moat around your business.”
Data security and confidentiality is a must among enterprise customers. “OpenAI powers the voice component of Athena. No large language models ever see the data in our Data Cloud or our clients’ data. We keep all of that totally and completely safe. The decisions run on our models, and no large language models ever touch the data in our Data Cloud.”
AI is allowing SaaS companies to move much faster. “In the second quarter, 90% of our new code generated was automated, helping our teams to innovate faster and continuously enhance the platform for our customers. It’s just accelerating it at a pace that I never thought humanly possible. We increased operating margin by 170 basis points. We added engineers in the quarter. We did not eliminate engineers in the quarter.”
Incumbent SaaS is such a beneficiary of AI. “Development cycles that would have taken years are now months, and new products inside of the ZBI, as we productize it, to hours versus what could have been months in the past. We were able to put together a solution in hours that came back and they believe will result in millions of dollars of incremental revenue to them.”
AI products released by incumbent SaaS are driving high use. “The 20% of customers who have comprehensively adopted our AI tools account for roughly 70% of revenue. Among super scaled customers, the 50% who have comprehensively adopted our AI tools drive 75% of super scaled revenue. These AI super users grew four times faster than the 80% of customers still early in their AI adoption journey.”
Freshworks CEO Dennis Woodside
AI is driving SaaS performance. “AI is a tailwind enabling growth in our business. Over 7,000 customers are paying for an AI SKU. Our Copilot attach rate on larger deals exceeds 70%. Our products allow customers to adapt to an agentic world quickly and achieve measurable IT efficiencies faster.”
AI has shaken up sales cycles. “AI actually is more of a motivator for people to think about their vendor. Most of our business is coming from another vendor. They’re faced with a decision: what are we going to do about AI? Are we going to migrate to the incumbent vendor’s platform? Sometimes that requires an upgrade in plans. It certainly requires cost. It often provokes a discussion as to, ‘Well, maybe we should go to market and see what else is out there.’”
AI is driving productivity in SaaS companies. “We’ve got designers who can work in Figma, create a product in Figma, push it directly to code. The process for doing QA is highly automated with AI, resulting in meaningfully shorter cycle times — about 30% faster. We’re shipping on basically a two-week cycle now, which we were not doing before, for our AI products in particular. We’ve implemented our AI email agent internally to handle questions from our own customers about billing, and we saw about 30% of those questions were completely handled through AI when we turned it on.”
History and experience matters for AI to be effective. “Customers want a system of record that has the kind of control and security that they need, that has AI integrated into it in a way that’s usable, that’s easy for them to get up and running, that’s easier for them to configure. To do AI well, you need to understand the operating environment. You need to understand the workflows that already exist, the controls that already exist in the operating environment, and that’s what we’ve spent over a decade building.”
Klaviyo CEO Andew Bialecki
The data moat of incumbent SaaS is going to be tough to surmount. “Our data infrastructure now stores more than 9 billion consumer profiles and ingests and indexes more than a quarter of a trillion data points every quarter. We built the ability to personalize and power up to 100 million marketing messages and experiences in less than 20 minutes.”
The AI product being put out by incumbent SaaS is compelling. “First, we built the ability to understand the structure of data, the ontology, and semantics into Composer to improve our agents’ reasoning abilities. Second, Composer understands the style of a brand — what we call their taste — because it has the direct context of marketing decisions they’ve made in the past. Third, Composer is available where and how users want to work. Fourth, it has access to aggregated knowledge we’ve curated about what makes marketing and consumer experiences convert.”
Enterprise incumbent SaaS has a significant edge. “I asked the CIO, ‘Why did we win?’ His answer was: ‘You won because of your AI vision and openness, your interoperability — especially with AI players like Anthropic and OpenAI — and because of the robustness of your infrastructure. Those were the reasons you guys won.’”
Atlassian CEO Mike Cannon-Brookes
Context and domain knowledge is the moat. “We believe models will keep improving, organizations will hire that intelligence by the token. Context — their internal knowledge, experience, and memory — is much harder for organizations to build. It cannot be hired. With the Teamwork Graph, we have 25 years of deep data about work. That’s enabled us to build one of the best context graphs that exists for enterprise knowledge, now spanning over 200 billion objects and connections. For agents grounded in the Teamwork Graph, organizations can see up to 44% more accurate answers while consuming 48% fewer tokens. Context is a clear compounding differentiator.”
AI is driving upgrade cycles at software companies. “The Teamwork Graph and AI come up in every single conversation I’ve had. It’s one of the top two reasons quoted by our customers for their reason to upgrade to the cloud and to upgrade to the Teamwork Collection.”
The world needs engineers. “I think there’s going to be more developers in the world in five years’ time than there are today. The cost of building technology is going down. The amount of technology we’re going to build is going to go up. I believe that will continue to be a good trend. We’ve said this for a number of years. Those are the signals we see in our customers.”
AI product from incumbent SaaS is doing very well. “MCP and CLI users passing 1 million now — I think it’s one of the biggest MCP servers that exist. More than doubling in the quarter. This is about people getting to the Teamwork Graph, getting to that context from their agents wherever they’re deployed.”
The world will need more software, not less. “In an AI-driven world, the need for collaboration at the enterprise scale — tracking, planning, managing work — is actually increasing. Atlassian is a mission-critical platform helping customers orchestrate their teams, their agents, and their workflows to unlock value in AI and drive real ROI.”
-Atlassian CEO Mike Cannon-Brookes on Q2 earnings call
Dyantrace CEO Rick McConnell
AI creates new markets for incumbent SaaS. “AI workloads do not simply add volume. They behave differently. They can operate perfectly and still produce incorrect results. That’s a problem observability has never had to solve before, and addressing it represents a significant emerging opportunity.”
AI has the same ROI questions humans have always had. “There are three questions that matter most in an AI-first world. First: Is it working? Are applications, infrastructure, and systems working as intended? Second — and this is new: Is it accurate? Is the AI model delivering output that can be trusted and relied upon with confidence? Third: Are my agentic systems delivering the outcomes they were built for? Code that’s built well, ships safely, and runs reliably.”
Without software, AI isn’t useful. “This is the moment for which Dynatrace was built. With AI agents increasingly acting alongside humans across development and operations, both need a common source of trusted context. Dynatrace provides that through Grail and Smartscape, giving agents and teams a unified understanding of system relationships and behavior. Dynatrace Intelligence turns that understanding into action, combining deterministic and agentic AI to deliver the precise causal insight that lets both people and agents act with confidence.”
AI creates new monetization opportunities for incumbent SaaS. “We see AI contributing in three ways. Increasing consumption across our platform. Creating demand for new AI observability capabilities. And directly monetizing agent usage. Every time a customer uses Dynatrace Intelligence to get answers through AI function calls or MCP integrations, or when one of our agents takes autonomous action to resolve an issue, it drives DPS usage.”
The new markets AI creates for incumbent SaaS are large. “We estimate the AI observability total addressable market will exceed $10 billion by 2030, growing at more than 50% annually. We see AI observability as the next logical evolution of the broader observability market, and that evolution is already underway.”
AI workloads will drive SaaS company performance. “Our focus is ARR acceleration for the year. Consumption is continuing to grow at a robust rate. Logs and telemetry pipelines are building. AI workloads are increasing consumption. Back-half renewals are coming up. There is a large set of drivers that deliver confidence in the overall outlook for the year.”
Datadog CEO Olivier Pomel
AI is expanding software’s TAM. “If we provide more value, we sell more software by helping customers make more money or save money or both. I think if we can automate more and let them do more, we’ll provide more value. When they use Bits AI, they use more of our product. They deploy more of it. They create more dashboards and alerts and everything else. They have more users inside of our product. It’s not a zero sum game.”
Incumbent SaaS has excellent AI product in market. “As of Q2, over 750 AI customers use Datadog to monitor and improve their tech stacks. When we look at the largest companies driving AI, all 10 of the top 10 AI leaders are Datadog customers. Beyond AI natives, we see AI activity growing across our broader customer base. We are also seeing signs of rapid growth in agentic activity with MCP tool calls quadrupling again quarter-over-quarter and growing more than 22x when compared to Q4 2025.”
AI is creating new markets for incumbent SaaS. “The multiplication of models, and open source models in particular, opens the door to customers doing a lot more training on their own. That’s a new market for us. We see some signs that we have a very good role to play there. We’re building towards that as well. Overall, I would say it’s very positive for everyone.”
Incumbent SaaS is performing. “If you backed out our largest customer from our growth, you get pretty much the same growth rate as the rest of the business has been accelerating very steadily. Actually, we’ve seen now five quarters of continuous acceleration from the rest of the business, and we feel very good about what we see in the market.”
Incumbent SaaS has more to offer in an AI world. “There’s only two reasons people buy software. It makes them more money or it saves them money. Anytime we go out in a renewal, we go out in an upsell, or we land a new customer, that’s because we do one of those two things for them. I think if we can automate more and let them do more, we’ll provide more value. That’s as simple as that.”
-Datadog CEO Olivier Pomel on Q2 earnings call
Cloudflare CEO Matthew Prince
AI is driving the growth of incumbent SaaS. “For the first time in human history, in Q2, more than 50% of the traffic flowing across Cloudflare’s network was not human. The number of requests on our network from AI agents continues to grow unabated. With the web shifting from human-driven browsing to AI answer engines and agent-driven commerce, we are witnessing a fundamental rewrite of the internet for machine to machine traffic.”
Incumbent SaaS is well positioned to innovate. “We didn’t actually set out to build that ourselves. We thought maybe we could partner with someone. Unfortunately, a lot of the world that’s building next generation payments networks thinks that they’re competing with Visa, as opposed to trying to figure out how you help build a better internet and ensure that there’s a healthy and successful business model for the internet going forward.”
Software’s opportunity is enormous. “We handle roughly about half a billion requests per second through Cloudflare’s network. Somewhere between 1% and 10% of those you could monetize through some sort of micro transaction. You’d need to be able to support, day one, call it 10 million financial transactions per second, and be able to scale up to call it 100 million financial transactions per second. Visa at peak during the holidays handles about 20,000 transactions per second. You have to build something that’s three orders of magnitude bigger than Visa.”
Hyperscalers have a really uninteresting business. “Not all revenue is created equal. If you’re selling what is just commodity compute, if you’re basically letting an AI company use your balance sheet and your credit rating in order to buy servers that are the same as everybody else’s servers, then that’s just not attractive business for us. We’re in a very different business than the hyperscalers. We’re selling work getting done, and instead of it being up to the customer to get as much out of the server as possible, we need to do the work to get as much out of the underlying equipment as possible.”
Hubspot CEO Yamini Rangan
Sales cycles are longer. “For prospects, we’re seeing larger buying committees and more of the deals that we are participating in require C-suite approval or board approval, and that means the deal cycles are longer. Having said that, we are seeing a lot of large opportunities within the pipeline, larger than what we’ve seen in the past. They are closing, but they’re just closing a month later or a few weeks later.”
AI elongates sales cycles, but then drives upsell. “We knew that evaluations will slow deal cycles, but we would provide higher confidence for customers as they adopt AI, and that would allow us to seed as many use cases as early on as possible with customers. Once they buy the first agent and the second agent and they begin to see clear outcomes, they are much more progressive about adopting the third, fourth, fifth agent.”
Concerns customers have when adopting AI offerings. “With AI, the real change is that we deliver outcomes. They’ve got to make sure that it works within their environment with their data, and they’ve got to understand the ongoing economics before they commit. It is a different buying motion.”
An example of AI changing pricing models. “Businesses have been hit with unpredictable token costs, and they want pricing that is transparent and tied to value. In response, we introduced outcome-based pricing for several of our HubSpot agents, lowered entry price points, and are providing customers clear visibility and control over usage and spend, including the ability to set thresholds that fit their budget.”
AI adopters are super users, adopting it in many workflows. “Total credits consumption grew in Q2 despite the pricing changes we made in April. Credit usage is now evenly distributed across Data Agent, Prospecting Agent, Customer Agent, and buyer intent. That is an important signal. Customers aren’t adopting a single AI use case. They’re using HubSpot agents across the entire customer journey.”
ServiceNow CEO Bill McDermott
Software is infrastructure, regardless of the model. “Whichever chip wins, whichever lab wins, whichever price per token regime prevails, the enterprise needs one governed layer of record for work. ServiceNow offers needed certainty in an uncertain stack. Our platform is optionality on all AI outcomes, not a bet on any one.”
Software is more critical than ever for what AI is bringing. “There are 2.2 billion agents entering the enterprise globally. That’s 2.2 billion new identities, a quarter of today’s human population. Veza maps access across human, machine, and AI identities. We’ll have 40 billion connected devices in the world in the next four years. Armis already tracks 7 billion of those devices in real time. Every ungoverned asset and identity multiplies the blast radius.”
Build vs Buy. “I’ve yet to meet a customer who would even consider it. The best tech leaders know it will cost 5 to 10x to build an agent versus run one on ServiceNow.”
Customers are done paying for tokens. “In an environment where most enterprises are still searching for AI’s ROI, ServiceNow is the platform delivering it. Customers aren’t paying us for tokens, they’re paying for resolutions. That’s why enterprises are choosing ServiceNow to convert AI ambition into measurable ROI.”
JFROG CEO Shlomi Ben Haim
AI creates software, which requires more software. “As AI accelerates software creation, engineering velocity, and code quality, the challenge is no longer generating source code, but managing the tsunami of binaries compiled. It is now about establishing trust in these software artifacts, models, agents, and packages that AI and human increasingly produce without sacrificing speed. The JFrog Platform is evolving for a future where AI agents become first-class citizens of the software supply chain.”
Yet another use case AI has created for software. “AI-powered software supply chain attacks continued to escalate this quarter, with threat actors increasingly targeting open source package ecosystems. As AI accelerates software creation, it also accelerates the pace and sophistication of software supply chain attacks. Throughout these incidents, customers using JFrog Curation remained protected.”
OpenAI is a JFROG customer. “AI models in today’s world should not be treated as free. They should be treated with zero trust, with the security practices that are required around that. Once this AI model found a vulnerability within Artifactory, they contacted the JFrog team immediately. We remediated fast, worked in great partnership with the security researchers of OpenAI. Great relationship will build a better product. More and more vulnerabilities will be found as models are getting into the pipelines, and I think that what counts is how fast vendors are remediating.”
Software will be required for enterprises to deploy AI safely. “It is becoming increasingly clear that AI will only be adopted at enterprise scale if it is trusted. That trust requires governance, compliance, and auditability to be engineered directly into the software development workflows — not introduced as a separate manual step. This is why we believe DevGovOps represents the next evolution of software supply chain management.”
Security software will be needed regardless if AI’s role. “Whether software is written by a human developer, an AI agent, or both, it ultimately results in more trusted binaries that must be secured, managed, governed, and distributed. As AI reshapes how software is created, we remain focused not only on what is growing, but also on what matters most — the trusted binaries that power production.”
Twilio CEO Khozema Shipchandler
Successful software will require an agnostic approach. “We are going to stick to our positioning as being kind of the Switzerland of it all. Model-agnostic, data-warehouse-agnostic, LLM-agnostic. The world is going to have multiple models. The world is going to have multiple data warehouses. There will be multiple clouds. And all of them are going to want to reach humans. They’re going to need to do that through the Super Network.”
A big market for software is connecting agents with humans. “I think this is durable, not just for one or two years, but for many, many years. The amount of AI-driven voice interactions is going to grow dramatically. Every business is going to want an AI agent. Every AI agent is going to need to reach a human. And they’re going to need to do that through us.”
Another example of AI as a tailwind for SaaS. “Record revenue, record gross profit, record operating income, record free cash flow — all in the same quarter. We raised full-year guidance by 400 to 500 basis points, the highest we have guided to in three years. The business is in excellent shape.”
SPS Commerce CEO Chad Collins
Another way SaaS co’s are using AI. “We completed our first AI-powered customer onboarding, including pre-sale contacts and account provisioning. We are working toward a future where agentic technology can engage a new customer immediately after a deal closes with more of the onboarding processes shifting to AI as we continue to reduce the time it takes for customers to transact with their trading partners.”
You cant just add a generic chat interface and call it AI. “Today, it may take them 20 prompts in the chat to get to the right answer. We are seeing that is something that could be automatically detected and potentially, in some cases, automatically resolved. We are developing those types of agents on top of MAX now. Those agents that can do things more autonomously — identifying anomalies, in many cases resolving them — not only finds the hard ROI in supply chain savings, but will also be a very favorable headcount and efficiency impact for our customers.”
The Trade Desk CEO Jeff Green
Enterprise customers need to trust your AI. What’s really important as we enter the new phase is that you have to get the biggest brands in the world to trust you with their data and then reassure them that you are going to preserve their data so that their insights from buying are put to use for them and exclusively for them.”…. “In an AI world, the premium on trust is going up, not down. People are looking for partners that they can trust. You’re going to see over time more and more of a separation between those that align their interest with their clients and those that don’t.”
Shopify CEO
SaaS companies are putting our AI with real ROI. “Catalog will be one of Shopify’s most important assets for years to come. We’re seeing that AI searches powered by Catalog convert at twice the rate of those using scraped data. That is because with Catalog, merchants’ products show up complete, accurate, and with the right context when someone is ready to buy. Catalog is the discovery engine for the future, and Shopify built it and owns it.”
SaaS is the underlying infrastructure of AI. “Whether commerce is handled by humans or agents, whether stores are built by people or AI, Shopify runs underneath it all. For 20 years, we’ve built a commerce operating system that takes merchants from first sale to full scale by using our partner ecosystem as an extension of our platform. That is our muscle memory. And this model will continue to serve us even better in this new agentic era of commerce.”
AI creates new markets for incumbent SaaS. “75% of AI-attributed orders in Q2 came from outside the top 100 categories. While search engines rank by popularity against a handful of keywords, AI agents make multiple calls into Shopify’s Catalog working with richer structured data to match products with the buyer’s specific intent rather than just keywords. So when a buyer asks an AI assistant for the best car seat that fits 3 across the sedan, traditional search focuses on the keyword car seat. An agent understands the actual need — the dimensions, the vehicle type, and the fact that they need 3. And in this world, relevancy reigns.”
AI allows the customer to use SaaS the way they want. “During a merchant’s first 30 days, roughly half of their conversations with Sidekick are about store setup, design and theme configuration. For merchants 5 years in, that drops to about 8%, while analytics and reporting claims past 40% as they use Sidekick as their intelligence layer to interrogate their own data and make better decisions. Same product, different job.”
AI agents are a new source of revenue for SaaS. “When it comes to monetization, the focus is simple. We unlock more places for our merchants to sell, and we earn on those sales the way we always have. Agentic transactions carry the exact economics as an online store transaction. There’s no new fees, no separate pricing. But more agentic GMV means more Shopify revenue, and that’s the model. It’s been working really well for almost 2 decades.”
Yet another example of AI benefitting SaaS co’s. “Buyer shopping journeys are being compressed as half of all AI-referred sessions are landing directly on a product description page. That is 2.5x more than what we see with traditional search. All of this is a serious tailwind for our merchants and in turn for us at Shopify.”
FIVN CEO Amit Mathrada
For Enterprise AI to be valuable, it needs to sit on top of and work with software infrastructure. “In complex enterprise environments, being early with a feature is not the same as being trusted as an operating platform. These customers need AI embedded into the platform they already depend on. That is why our focus is not AI in isolation. It is AI agents and human agents working together across voice and digital channels inside one trusted platform to deliver Humanic CX.”
Humans and AI together are way better than AI alone. “Our belief is that humans and AI are going to come together to really start delivering new economics in the contact center, as well as improved experiences, and new ways of doing business. Over time, AI agents will handle a larger share of customer interactions, including many routine and multi-step service requests. Human agents will remain essential for complexity, judgment, empathy, escalation, and oversight. The value comes from orchestrating both together so the customer experience is seamless and the platform learns from every interaction.”
Underlying software infrastructure makes AI a good investment. “Today, you come in as a customer with a specific billing issue. You’re infuriated. What happens as a point solution — when the AI agent identifies the issue, it sends you to the billing queue. What a company like Five9 can do is, because we have run agentic quality management on all your agents, have already identified which agents are best of breed to handle that question, which agents have a high empathy score — now with my agentic routing, I can send that call specifically to that one agent that has high empathy and high ability to answer that question.”
Below are the key themes from the calls we have reviewed so far.
Theme 1. SaaS product has AI embedded. AI is a core requirement in nearly all large enterprise deal discussions. Freshworks, Klaviyo, and Atlassian’s all reported that some form of their AI infrastructure, vision, or interoperability was a factor in deals. From Atlassian’s CEO: “The Teamwork Graph and AI come up in every single conversation I’ve had. It’s one of the top two reasons quoted by our customers for their reason to upgrade to the cloud and to upgrade to the Teamwork Collection.” At FIVN, every million-dollar deal that goes out the door has 100% attach of AI.
Theme 2. The model is a commodity. The context and data are the moat. At Shopify, Catalog-powered AI searches convert at 2x the rate of scraped data from generic AI. Atlassian’s CEO was direct on this: “Intelligence can be hired by the token. Context cannot. With the Teamwork Graph, we have 25 years of deep data about work spanning over 200 billion objects and connections.” For agents grounded in the Teamwork Graph, customers realize up to 44% more accurate answers while consuming 48% fewer tokens.” It’s thanks to context, not the model. Zeta’s CEO: “Nothing you rent can be a moat around your business.” Palantir’s Karp: “An organization’s data is its treasure. In its richness is the alpha. We are fully aligned with our customers, building a stack that enables the compounding of their alpha.” From Kalviyo: “Delivering meaningful customer experiences at scale requires AI grounded in real data. We built the ability to understand the structure of data, the ontology, and semantics into Composer to improve our agents’ reasoning abilities.” In summary the model is available to everyone; the proprietary context built over years is available to no one else.
Theme 3. Data must be protected from models. Palantir’s CEO was combative (again): “People sign up for token self-pleasurings, paying for the right for them to migrate your IP, your know-how, your expertise to their model so that they can build a competitive business that doesn’t require your business, your people. They deserve to colonize your enterprise. You deserve to be colonized.” Zeta’s CEO was less dramatic but clear: “OpenAI powers the voice component of Athena. No large language models ever see the data in our Data Cloud. The decisions run on our models.” Data security and confidentiality is becoming a customer requirement.
Theme 4. Agents are a new customer for software companies. Datadog’s CEO sited numbers: “MCP tool calls quadrupled again quarter-over-quarter and grew more than 22x when compared to Q4 2025.” Every MCP call is an AI agent executing a tool that requires monitoring, governance, and security. Dynatrace’s CEO: “AI workloads do not simply add volume. They behave differently. They can operate perfectly and still produce incorrect results. That’s a problem observability has never had to solve before.” Agents have also dramatically increased the attack surface; ServiceNow says there are 2.2bln agents entering the workforce globally, and every ungoverned agent increases the blast radius for cyber attacks. AI is generating its own demand for the governance, observability, and security layers that enterprise software provides.
Theme 5. Agentic product built by software companies is very popular. Atlassian’s “Rovo-assisted actions grew 50% QoQ, MCP calls grew 400% QoQ. MCP and CLI users passing 1 million now, I think it’s one of the biggest MCP servers that exist. More than doubling in the quarter.” Klaviyo: 95,000 Composer users in the first month of GA, credit consumption growing 30% week-over-week, nearly a quarter becoming recurring weekly users. Customers that are using Composer are the ones that end up being more successful with Klaviyo literally in the first few weeks.” Datadog: $115 million sequential revenue add, a record by a wide margin, with 58% of customers now using four or more products.” Zeta: “AI super users growing four times faster than other customers, accounting for 70% of revenue, with NRR 400 basis points above the company average.” Shopify’s Sidekick handled 34 million conversations in Q2, growing 4.8x in daily sessions year-over-year. Five9’s AI revenue is now 15% of total subscription revenue. AI product from software companies is driving measurable financial acceleration.
Theme 6. AI is really accelerating product velocity. At Zeta, “90% of our new code generated was automated. Development cycles that would have taken years are now months, and new products inside ZBI to hours. We added engineers. We did not eliminate engineers. We increased operating margin by 170 basis points.” Freshworks is experiencing “30% faster cycle times. We are shipping on a two-week cycle now. We implemented our AI email agent internally and saw about 30% of billing inquiries completely handled through AI when we turned it on.” The best SaaS companies are not using AI to cut costs, they’re using it to ship more product faster.
Theme 7. The financial results this quarter are exceptional, a trend which started in Q4 2025. Palantir: revenue +93% YoY, Rule of 40 of 155, $1.22 billion in adjusted free cash flow, largest guidance raise in company history. Cloudflare: revenue +36% YoY, 986 large customer net additions in 12 months — a record by a wide margin. Datadog: first $1 billion revenue quarter, $115 million sequential revenue add, fastest sequential growth since Q2 2022, 13% of customers now using 10 or more products up from 7% a year ago. Atlassian: EPS 24% above consensus, RPO +44%, $3M+ ARR customers growing 50% YoY. Zeta: 20th consecutive beat-and-raise. Freshworks: first ever GAAP profit. Klaviyo: largest deal in company history. Of the 16 companies we’ve reviewed so far, only one is struggling (SPS Commerce). The rest are beating guidance and setting records.
Theme 8. AI is expanding the addressable market for software companies. This builds on theme 4, but goes beyond just the fact that an AI agent is a user that needs software. For instance, at Shopify, AI searches are opening the long-tail merchant discovery market that traditional search never reached; long tail makes up a lot of Shopify stores. According to Twilio, less than 5–6% of voice interactions are AI-driven, which means more business for Twilio as that figure climbs. Five9 says AI is converting contact center spend away from labor and toward mission-critical software. New markets are opening up for SaaS, but in ways that go beyond just adding users that are agents.
Thank you for your readership. See more blogs and SaaS data at blossomstreetventures.com. Other resources we’ve built for founders include: SoftwareMultiples.com; softwareMRRcalculator.com; FounderInvited.com, and TwoFoundersTalk.com. Founders are always welcome to reach out to sammy@blossomstreetventures.com as well.