We’ve been reviewing earnings calls of SaaS companies this Q3. So far we’ve gotten through 21 transcripts and and 35 press relesases. Below are the most notable comments from SaaS CEO’s on AI; run them through Claude for key themes. Share what you learn with me.
Platforms over AI point solutions: “While operators outside our platform can buy isolated AI features, a point solution cannot provide the connected foundation, governance, and economic alignment that allows performance to compound month after month, creating an engine for competitive advantage.”
-Appfolio’s Shane Trigg on Q3 earnings call.
AI revenue and ARPU. “RingCentral AI, which we define as customers using at least 1 native paid AI product, now represents approximately 15% of total ARR or over $400 million. Doubling year over year. This cohort of customers also has meaningfully higher ARPU and net retention. These trends demonstrate how AI is expanding wallet share, strengthening retention, and creating a more durable foundation for profitable growth.”
-RingCentral’s Vladimir Shmunis on Q3 earnings call.
Charging for usage and moving to open weight. “Our gross margins on AI products are generally holding in line with the corporate average. On the pricing front, we are being very deliberate — the idea is to charge customers for the value and the usage that they receive. And customers are willing to pay because they are getting tangible ROI benefits both from a revenue standpoint as well as a cost avoidance standpoint. In terms of costs, we are optimizing against model selection, inference efficiency, and infrastructure. Over time, we will benefit from the use of open source models.”
-Ring Central Q3 earnings call, Vaibhav Agarwal
The customer wants a human in the loop. “The big thing for our AI, whether it is AIR, AIR Pro, any of our products, we think about human in the loop. Handle as much as possible by AI. But also make sure that for judgment, sentiment, or more complex questions, those are handed off to a human being. And they work together, preserving that context, which our customers are telling us is very important.”
-Ring Central Q3 earnings call, Kira Makagon
AI needs software governance. “94% of enterprises report gaps in AI visibility. 88% can’t even distinguish personal AI accounts from corporate instances. ServiceNow’s Control Tower discovers and inventories the entire AI estate. As stories of rogue agents persist, ServiceNow controls what agents can do. That’s why we introduced AI Passport to issue every agent an official attested record based on context in our platform. Who owns the agent? What’s it allowed to do? What data can it access? How much risk exposure is it allowed to take? This enables the most capable kill switch available in the enterprise today. When our Control Tower revokes a passport, NVIDIA Century stops honoring it and the agent stops at the infrastructure layer.”
-ServiceNow CEO Bill McDermott on Q3 earnings call.
Incumbent software companies are showing fantastic AI revenue growth. “We began the year with a $1 billion AI target by year-end. We raised that target to $1.5 billion at Financial Analyst Day in May. Today, we are raising that target again this time to $1.7 billion plus by year-end. AI net new ACV was 65% year-over-year growth in Q3. The number of $1 million-plus AI deals grew 4.5x year-over-year. Year-to-date, agentic AI consumption in production is up 13x year-over-year. If you look at the consumption of Action Fabric assists alone, that number is up 500% month-over-month.”
-ServiceNow CEO Bill McDermott on Q3 earnings call.
Traditional SaaS pricing is very much alive, and now supplemented with AI. “The customer likes the predictability of the seat-based pricing model for their budgeting purposes. And also, we’re driving consumption. The idea is simple: get the customer live fast. And once they’re live, the token economics take off, whereby we have a chance to improve our revenue per user by a factor of 5x. And that’s only because the customer derives substantially more value than that.”
-ServiceNow CEO Bill McDermott on Q3 earnings call.
No more AI proof of concepts. “We’re past the proof-of-concept phase. We’re past the experimentation phase. We’re past point solution or vibe coding or do-it-yourself phase. We’re now in serious enterprise phase where it’s all about get the job done, deliver real value. And this platform is one of a few that will do extraordinarily well now and in the future.”
-ServiceNow CEO Bill McDermott on Q3 earnings call.
AI is creating real upsell opportunities for SaaS. “Many of the customers typically start with one or two use cases — incident management or triaging or resolution planning. And once they see that go live, we’ve seen that go up to 3x, 4x usage, but also new use cases emerge because now they’re seeing it in the production environment. We have customers like Nestle, for example — they went live in six weeks with marketing optimization, agentic workflow. And now they’re unleashing that in 90 different countries.”
-ServiceNow CEO Bill McDermott on Q3 earnings call.
Open weight models on domain specific data is the future. “Our research shows that 85% of social impact professionals use AI in their daily work, yet only about 10% of organizations are seeing returns. General purpose models learn from the Internet. Lantern learns from philanthropy. Developed through a strategic collaboration with Databricks, Lantern combines leading open weight models with more than 40 years of Blackbaud’s fundraising experience. It is post trained on synthetic scenarios modeled on our social impact signal graph. It understands giving behavior rather than buying behavior — including retention, upgrade paths, and lifetime relationships — which are central to fundraising but not native to general purpose AI.”
-Blackbaud CEO Mike Gianoni on Q3 earnings call.
The core of AI is software, and must be invested in. “Our AI investment is in addition to, not instead of, our investments in the core products our customers depend on every day. Those core products are the systems of record, the workflows, and the trust that make our AI valuable. And we continue to invest in them aggressively. Every improvement to a core product — such as cleaner data, tighter integrations, and faster workflows — makes the intelligence layer more accurate and more useful. And every agent and model we ship is embedded in a solution customers already trust, which reduces the data gaps and security risks created by bolt-on tools.”
-Blackbaud CEO Mike Gianoni on Q3 earnings call.
No industry can use AI better internally than SaaS can. “Our use of AI in engineering has expanded our capacity and increased our velocity such that we are announcing more new solutions than we ever have. The Raiser’s Edge NXT enhancements were planned for customer use in 2027 but were accelerated and pulled forward by more than a year.”
-Blackbaud CEO Mike Gianoni on Q3 earnings call.
Incumbent SaaS is well positioned to deliver trusted AI. “Trust is critical to new and existing organizations allowing intelligent systems to act on their behalf. Our customers’ trust in us as their partner, in the solutions they are used to, in the data, and in our security are two advantages for both our customers and for Blackbaud. The gap between AI adoption and effective use is exactly what we are working to close — by delivering trusted AI inside the tools our customers already use.”
-Blackbaud CEO Mike Gianoni on Q3 earnings call.
A hybrid including traditional SaaS pricing will prevail. “It is not going to be uniform across products and even potentially across customer types. What seems to be resonating with customers is some form of hybrid model — some base platform fee that protects us and them against overusage, then some variable lever around value. In the case of fraud, the number of businesses or consumers we’re protecting. In the case of Assistant, the number of use cases we’re solving. Pricing has not been a major hurdle for us so far.”
-Q2 Holdings CEO Jonathan Price on Q3 earnings call.
SaaS companies are using AI to accelerate onboarding and implementation. “In Q3, this accelerated onboarding capability was a deciding factor in securing multiple new customer wins. We are prepared to move as quickly as our customers need us to, and delivering faster time to value. To date, we have completed many complex ERP integration and onboarding projects in less than 30 days, and we expect to shorten these cycles even further as we advance our agentic capabilities.”
-SPS Commerce CEO Chad Collins on Q3 earnings call.
AI will work with software, not replace it. “Today, the account holder navigates menus and completes each task. In the AI era, they will describe a goal in plain language, and the software will work out the steps and ask for permission to act.”
-Alkami CEO on Q3 earnings call.
Software has always been the infrastructure on which humans work, and now it will be that for AI agents. “Every institution on it configures roughly 4,000 policy settings of its own across approximately 25 systems of record that are integrated into Alkami. We built that so our clients could each run their own bank on shared infrastructure. It is also what an agent needs. One place to ask, one set of permissions to respect, and a record of every action taken.”
-Alkami CEO on Q3 earnings call.
AI is accelerating every software company’s desire to become a platform. “What used to be five or six separate purchasing decisions becomes one decision, and what used to be a struggle to manage multiple products becomes a single enterprise-level solution. For HealthStream, suites typically mean larger contract values, longer commitments, and a broader adoption across the enterprise. Because customers keep the full suite rather than adjusting individual components at renewal, we believe bundles can also support stronger renewal rates over time.”
-HealthStream CEO on Q3 earnings call.
Exactly the right belief of AI’s impact on software. “The more clinicians, data, and solutions we connect to our platform, the more valuable each of those advantages becomes. AI cannot build an ecosystem like ours, but we believe it can make ours even more valuable.”
-HealthStream CEO on Q3 earnings call.
AI needs software infrastructure to operate on. “AI needs a place, it needs somewhere to act. When we talk about Waystar’s platform being the financial execution layer, as we deliver new SKUs that are AI-powered, we can turn them on our platform and distribute those rapidly.”
-Waystar CEO on Q3 earnings call.
More platforming happening in software. “In a competitive process, a major East Coast health system representing approximately $7 billion in annual net patient revenue selected Waystar’s full platform across patient access, mid-cycle, and backend revenue cycle operations. The organization plans to replace more than 20 separate vendors with Waystar as it simplifies its technology infrastructure and streamlines operations.”
-Waystar CEO on Q3 earnings call.
AI is driving software volumes. “We are no longer on the cusp of the agentic era. We are in the middle of it. AI is writing code, agents are operating software, and the growth of telemetry volumes and complexity is accelerating. Our own research shows that AI workloads drove a 93% increase in log volume in the past 12 months alone. As AI systems become increasingly autonomous, enterprises need greater visibility and control over how their applications operate and how AI agents behave.”
-Dynatrace CEO on Q3 earnings call.
AI driving spend on observability and security. “There’s a huge amount of budget scrutiny right now to compensate for the large token spend that is happening throughout the enterprise space. The positive side of that equation is that we believe very strongly that AI is a tailwind for observability, not a headwind. Ultimately, we expect that spend to have to come through because you actually need more observability to oversee probabilistic workloads. So it’s a double-edged sword in some ways — you have more budget scrutiny, but ultimately, that budget scrutiny should benefit observability.”
-Dynatrace CEO on Q3 earnings call.
AI agents are a fantastic tailwind for software. “On agents, we’re going to monetize those agents directly — Dynatrace agents as of October 1st are monetizable. Prior to that time, we had 1,100 customers in a public preview. Those customers are now being given the opportunity to rotate over, and as they continue to use agents, we’re monetizing by charging them AI units for that usage. Second, agents drive more data — more logs, more traces, more observability telemetry — that drives additional DPS drawdown. Third, with Arize coming onto the platform for AI observability and evaluations, we’ll monetize Arize directly as well.”
-Dynatrace CEO on Q3 earnings call.
Software customers that also use AI are the best customers. “AI customer cohorts consuming more than twice the rate of other customers. Logs consumption continues to grow at well over 100%. Infrastructure monitoring hit a milestone — it is now over $500 million in annualized consumption, growing approximately 30%. More than 1,100 customers leveraged the public preview of Dynatrace agents, up from 800 last quarter.”
-Dynatrace CEO on Q3 earnings call.
Platforms over point solutions, again. “Customers are not buying a point solution. They are standardizing on one platform for service operations across IT, HR, finance, and other teams on one data model. That gives us the power of an enterprise system of record with the agility of modern software. It sets us apart from both legacy incumbents and startup challengers.”
-Freshworks CEO on Q3 earnings call.
Customers using AI are software power users. “Teams using Freddy Copilot handle 50% more tickets. Freddy AI Agent deflects 50% of inquiries on average and up to 80% in mature deployments. In Q3, active Copilot agents doubled year over year. Customers choose us for the platform, the speed, and the AI, and they expand at a higher rate. In fact, EX customers who purchase AI have a net dollar retention rate of greater than 115%.”
-Freshworks CEO on Q3 earnings call.
Agents are their own customer for software companies. “AI center of gravity has moved from people chatting with models to agents doing real work and delivering outcomes. This shift has created two new slipstreams of value. First is token flow. Agents think in tokens, lots of them. Third-party estimates expect agentic AI to lift global token consumption 24x by 2030. The second slipstream is agentic flow. Agents bring together its runtime, tools, memory, and learning loop that turn thinking and intelligence into finished work. They require new primitives and a cloud that is built for agents, not humans. Most providers serve only one of these flows. We are a leader in both.”
-Digital Ocean CEO on Q3 earnings call.
Agents are a new customer for SaaS. “Agents are also now direct customers of DigitalOcean. They do not just run on our cloud. They increasingly select it, sign into it, and deploy on it. Already, roughly half of our unique visitors to digitalocean.com are agents, and more than 37,000 unique customers have deployed on DO through an agent this year. When an agent autonomously chooses where to build, it chooses the cloud it can read, reason, and deploy to without a human in the loop, and agents are increasingly choosing DigitalOcean.”
-Digital Ocean CEO on Q3 earnings call.
More examples of use based pricing for AI consumption. “We’ve built a clear path from free to paid. Every customer starts with a baseline of Trellis access at no additional cost. When they need more agentic work done, they can buy Trellis Plus per user seat. Based on customer feedback and demand, we’re now introducing a shared credits option, one pool that every user and every automation draws from, which we expect to be appealing to enterprises with large user counts and very different use cases. As Trellis does more work in the background, usage can scale with the value customers get.”
-Sprout Social CEO on Q3 earnings call.
Consolidation is its own moat. “They came to us with a fragmented social stack and consolidated it on Sprout, investing in Trellis+, Premium Analytics, Listening, Employee Advocacy, and Premier Success. Five products in a single new logo is what vendor consolidation looks like at this scale. What stands out to me is the adoption. Their previous Employee Advocacy program had stalled, and with easier onboarding and role-based permissions, it now runs across thousands of employees, with campaign reporting replacing the manual tracking.”
-Sprout Social CEO on Q3 earnings call.
The cost of AI is elongating sales cycles. “It comes down to, for so many of these companies today, certainly they’re thinking about AI spend, and that spend is coming from somewhere, and it’s putting pressure in certain parts of the business. We’re seeing many of these organizations, because of the pressure that they have on their own budgets, that they’re creating new evaluation processes and oftentimes new folks that are involved in the approval process, which can elongate sales cycles.”
-Sprout Social CEO on Q3 earnings call.
Closed models will be the way forward for the enterprise. “Last month, we introduced Scout, our own reasoning model specifically optimized for SRE tasks. Scout was built using our in-house distributed reinforcement learning infrastructure and now powers our Bits Investigation. This is exciting for two reasons. First, it performs on par or even better than frontier models on alert investigations at a fraction of the inference cost. This is valuable as it allows Bits to investigate many more alerts, start earlier, and go deeper on incidents as an always-on first responder for our customers. Second, this is also exciting as a demonstration of what we can achieve when we train models with our own production data, critical user workflows, and internal tooling. Scout is only the beginning.”
-Datadog CEO on Q3 earnings call.
AI’s problems are software’s market. “We continue to see rapid growth in agentic activity, with the number of MCP tool calls up by a factor of 50 so far this year. We are tremendously excited about where this is going. As AI writes and runs more of our customers’ software, they are shipping more code faster across a growing mix of models, agents, and infrastructure. Of course, this generates complexity that is very quickly outpacing what even the best engineering teams can track on their own. That is exactly the problem we are investing to solve.”
-Datadog CEO on Q3 earnings call.
Open weight models will be the answer in the enterprise. “We are big consumers of the top labs, top models, but we also consume open-weight models, a growing amount of opn-weight models, and we train our own and post-train our own. I think it is probably representative of where the industry as a whole is going. There are certain things you can only get from the top-of-the-line models. There is a number of things that are good enough and very cost-efficient with open-weight models you can just pick up and run. And there’s finally a number of other things that you can make way more efficient with your own data, your own workflow, and your own post-training.”
-Datadog CEO on Q3 earnings call.
AI customers are the best software customers. “These customers are growing and they’re deploying much larger workloads, and they’re looking at doubling, tripling, quadrupling their footprint with us over time. The only time we think about it in difficult terms is when we prepare for earnings calls, and we have to figure out, okay, so what part can we communicate, what part do we know, what part do we not know? But overall, these are all great news, great customers, great growth, great adoption, amazing trends that are underpinning the industry as a whole.”
-Datadog CEO on Q3 earnings call.
AI adoption at the enterprise level has not been immediate. “One thing we like to say is that with Composer, we think there is this big, almost like agent overhang. Composer as an agent is so good at doing marketing and customer analysis. The problem is not really ROI, it is more just helping more of our customers know what to prompt. We talked a little bit about setting up recurring workflows, and I think that is the phase that we are entering. The technology is there, our product is there. Really it is about driving large-scale adoption with both current Klaviyo customers and also just other businesses that need to do marketing.”
-Klaviyo CEO on Q3 earnings call.
6969
Domain specific agents from software companies have a higher ROI. “On our internal marketing benchmarks, we find that Composer outperforms frontier models at common marketing tasks. This translates to better performing marketing and ultimately more growth for our customers and for us. This is the advantage of domain-specific agents that are trained to excel at tasks that are less verifiable and require data specific to that domain.”
-Klaviyo CEO on Q3 earnings call.
Sales cycle is elongating, and free trials are in. “Customers are evaluating AI differently. They want to see AI work with their data in their environment. They want proof before they buy. We saw the shift early and leaned into it in Q2. We introduced free trials and pricing tied to value and outcomes. We knew this would extend buying cycles and create near-term headwinds, but we believe it’s the right long-term trade-off. We’re seeing encouraging early results. Agent trials grew 64% quarter-over-quarter in Q3. But we’re also learning that customers need hands-on support to turn AI trials into real outcomes.”
-Hubspot CEO on Q3 earnings call.
Multi-threading with the C-suite matters again. “With upmarket customers, they’re not doing a technology purchase. They’re actually making a strategic choice for the AI platform that’s ahead, which means they’re bringing in CEOs, CTOs, sometimes even the board for approval. When we do multi-thread and talk to multiple stakeholders, we see win rates improve to nearly three times higher in deals with five or more buying committee members.”
-Hubspot CEO on Q3 earnings call.
MCP’s with software platforms serving as the infrastructure is the way. “We were the first CRM to connect with all four leading frontier models, and today, we are the most popular CRM connector on both ChatGPT and Claude. What is important is what happens when customers connect HubSpot to their preferred LLMs. They’re not just pulling data out of HubSpot. They’re doing more work in HubSpot. They’re analyzing reports, managing pipelines, creating CRM records, and taking action, all through conversations. In Q3, customers created over 22 million CRM objects through connectors within HubSpot, up 157% quarter-over-quarter.”
-Hubspot CEO on Q3 earnings call.
AI is creating discovery and marketing challenges. “Inbound demand remains under pressure as AI changes how buyers discover software. We have been diversifying through the top of the funnel through intent-based marketing. We are now doing LLM connectors as a source for signups as well as leads, as well as AEO. These new channels are nascent, and they don’t really meet the changes in the inbound demand.”
-Hubspot CEO on Q3 earnings call.
AI agents need critical software, just like humans. “As AI use cases continue to accelerate, one of the world’s leading AI companies signed a seven-figure commitment with Fastly to secure and scale its agentic traffic. Agents are central to its strategy. They chose Fastly to protect that mission-critical work.”
-Fastly CEO on Q3 earnings call.
AI is making software more valuable, and creating new markets. “Our hypothesis for a while has been that the emergence of AI bringing accelerated vulnerability discovery and exploitation would actually increase the value of security products, particularly like our Next-Gen WAF at the edge. That was our reaction earlier this year when Mythos came out, and we’ve seen that a number of times, and I think PHP2Shell is just the most recent example of that. Customers can use the Next-Gen WAF to virtually patch all of their components because we can detect that exploit and block it before it reaches their components.”
-Fastly CEO on Q3 earnings call.
R&D efficiency at SaaS companies has exploded. “In the last year, we released more solutions than any previous 12-month period. We have enough product in testing right now or already in use by select clients to have confidence that 2027 won’t be any different… it’s the result that the R&D that we’re spending is delivering right now.”
-Paycom CEO on Q3 earnings call.
AI has exploded the volume of work, and that needs to be coordinated by software and humans. “The value of Jira and Confluence for coordination in the AI era is even stronger than it has been in the past… The human role is incredibly important in intent, in applying intuition and understanding, in guiding and handling exceptions, in initiating work… There’s no doubt the amount of work happening is going to go up, precipitously so. The coordination tax of that then becomes even more so. If you are trying to understand 10x as much work, you need to understand more coordination, more connections than you have been beforehand.”
-Atlassian CEO on Q3 earnings call.
AI is accelerating the growth of enterprise accounts. “Customers north of 3 million growing at over 50% year-on-year rate, and those at the 5 million level growing at 70% year-on-year rate. Certainly, at the enterprise end, I think you can see very strong, healthy customer growth resulting in monetization as they bring more and more workflows on.”
-Atlassian CEO on Q3 earnings call.
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.