Automation in SaaS Services

We’re watching the rapid transformation - and possible end - of SaaS as we know it. Microsoft CEO Satya Nadella recently pointed out that traditional SaaS is disappearing, and I strongly agree. But I see the timeline accelerating even faster: Phase 1 (Right now): AI as Support AI enhancements like Copilot, Gamma, and Harvey are currently complementing existing SaaS platforms, making them seem more efficient and attractive. Providers feel secure, viewing AI as a feature rather than a threat. Phase 2 (Within 6-12 months): AI Takes Over Operations AI agents will quickly transition from assistants to autonomous operators. Instead of manually using tools like Tableau or Meta’s ad platform, we’ll simply instruct agents to perform analyses or optimize ads directly. The expertise traditionally embedded in SaaS interfaces becomes easily accessible through agents. Phase 3 (Within 1-2 years): Software Becomes Invisible AI agents begin interacting directly via APIs, eliminating the need for human-oriented interfaces like dashboards and menus entirely. This strips away the core value SaaS once provided—human usability. This isn’t standard disruption; it’s a fundamental shift away from human-operated software to agent-operated software. At the same time, the rise of AI-driven coding tools makes custom internal software development dramatically easier and cheaper. Companies no longer need to rely on costly SaaS subscriptions—they can quickly create tailored internal applications that perfectly fit their needs. The winners in this new era won’t simply be those who integrate AI the quickest. Instead, they’ll be companies providing open, agent-friendly APIs, becoming the trusted providers of actionable data and execution within their fields. The real question is whether giants of all industries will swiftly adapt or risk becoming obsolete, much like tech giants of the past. We’re entering an extraordinary period of opportunity for agile startups ready to embrace this change.

The software (SaaS) sector is facing an "Apocalypse". leaders like Salesforce ($CRM), ServiceNow ($NOW), Veeva ($VEEV), and Constellation Software ($CNSWF)are down 40% to 50% from their highs over fears "Agentic AI" will eat their lunch by automating the very tasks these platforms manage. However, history suggests that markets repeatedly declare industries "dead," only to realize later that incumbent leaders are far more adaptive than expected. Before selling in panic, remember times when the market predicted a "Kodak moment" and was wrong: 1) The "Death of Retail" (2015–2016): Investors feared Amazon would make physical stores obsolete. Instead, leaders like Walmart and Target integrated the threat, using stores as "edge warehouses" for same-day pickup (BOPIS). Both eventually hit all-time highs. 2) The Cloud vs. On-Premises Crisis (2010–2012):Markets assumed legacy giants like Microsoft and Adobe were finished when AWS took off. They pivoted to SaaS models and captured the majority of the cloud's value, becoming the best-performing stocks of the decade. Why Agentic AI is a Catalyst, Not a Killer for SaaS Leaders The market currently overestimates the speed of disruption and underestimates incumbent adaptation. SaaS giants are already pivoting away from human seat licenses to Outcome-basedpricing. Software Leaders are Adapting and Integrating Agentic AI into their Revenue Models They are moving away from "per-seat" pricing to capture the value of automated "digital labor". 1. Salesforce ($CRM): Introduced Agentforce with a "Flex Credit" model. Instead of just seats, they charge roughly $0.10 per action. If an AI agent does 5x the work of a human, revenue can actually double per unit of work. 2. ServiceNow ($NOW): Using a "Pro Plus" SKU with a 30%+ premium to unlock "Now Assist". They are betting on "Agentic Fabric," where AI agents talk to each other across departments—a complex workflow startups can't easily replicate. Bottom Line: Markets systematically underestimate switching costs and regulatory friction. While the "Software Apocalypse" narrative is loud, the "SaaS Pivot" to AI agents may actually make these platforms more indispensable and more profitable.

🚨 Your SaaS vendor just added AI. Is your contract ready? That productivity tool your team loves just got "smarter." It's processing requests faster, offering better suggestions, maybe even finishing your sentences. But here's what changed behind the scenes: your vendor is now routing your data through third-party AI APIs from OpenAI, Google, or Anthropic. The problem? This wasn't in your original contract. Once a SaaS tool passes initial security review, we tend to set it and forget it. But vendors are now embedding AI capabilities post-contract, often treating them as routine technical upgrades. They're not routine. They're fundamental changes to: ❓Where your data flows ❓Who has access to it ❓How it's processed and retained Most existing contracts don't address AI usage, creating accountability blind spots for risks you never agreed to. When AI features fall outside your contract scope, you face: → Data processing violations: Your DPA likely doesn't cover sending data to AI APIs → Undisclosed sub-processors: AI providers often aren't on your approved vendor list → Unintended model training: Your data might be used to improve AI systems unless explicitly opted out → Audit gaps: Hard to review what wasn't acknowledged in the first place Let's close the gap: ✅ Create an AI addendum template: Require disclosure of all third-party AI APIs, define acceptable data use, and clarify jurisdictional requirements. ✅Update vendor risk reviews: Add AI-specific questions to your standard process. ✅Set governance triggers: Require notification of any material product changes involving new data flows. ✅Coordinate across teams: This isn't just IT. Get Legal, Risk, and Procurement aligned. Bottom line: When vendors add AI, it's not "just another feature." It's a shift in how your data is handled, often by entities you didn't vet under terms you didn't negotiate. Your governance model needs to evolve as fast as the technology does. What's your experience with AI showing up unexpectedly in your vendor stack? How are you handling contract updates?

The SaaS Reckoning Is Coming to the Office of the CFO The biggest threat to legacy SaaS in finance isn’t AI models; it’s architecture. For over 20 years, finance systems were built on the assumption that humans do the work—routing invoices, reviewing expenses, investigating exceptions, and escalating approvals. The software manages workflow while people apply judgment. Now, vendors claim they’ve “added AI everywhere.” However, adding AI to a human-centric workflow doesn’t make it agentic; it merely speeds up the routing of work to people. CFOs are beginning to recognize that when AI Agents can: - Execute invoice processing end-to-end, including exceptions - Audit 100% of expenses and communicate with employees - Enforce policy consistently, without fatigue - Produce a trace for every decision The human reasoning embedded in those old workflows disappears. When that happens, SaaS architectures designed around human action become bottlenecks. You can’t bolt autonomy onto systems meant for manual intervention, nor can you change operating economics with assistive features. This isn’t about UI improvements; it’s about structural change: - 50–90% reduction in operational headcount - Days dropping to minutes - Compliance by design - Audit traceability by default CFOs focused on real transformation are no longer swayed by “AI-powered” labels. They are asking: - Can this remove the work entirely? - Can this change my cost structure permanently? - Can this reduce risk at scale? If the answer is no, it’s merely incremental, and incremental won’t survive the agentic shift. Finance will be the first function redesigned around hybrid teams—humans for judgment, AI Agents for execution—and #AppZen is already making it happen in production. This redesign favors platforms built for autonomy from the ground up. #AgenticAI #CFO #OfficeOfTheCFO #FinanceTransformation #HybridWorkforce #EnterpriseAI #AppZen

AI in SaaS isn’t an add-on anymore.  It’s your foundation for growth. If you’re a founder, embedding AI from day one is how you ignite real growth and create lasting customer value. 𝐇𝐞𝐫𝐞’𝐬 𝐰𝐡𝐲 𝐢𝐭 𝐦𝐚𝐭𝐭𝐞𝐫𝐬: 𝐒𝐦𝐚𝐫𝐭𝐞𝐫 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧: AI adapts and automates complex tasks, saving users time and effort every day. 𝐏𝐞𝐫𝐬𝐨𝐧𝐚𝐥𝐢𝐳𝐞𝐝 𝐄𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐞𝐬: Real-time AI insights deliver tailored experiences that keep customers engaged. 𝐓𝐡𝐞 𝐀𝐈 𝐅𝐥𝐲𝐰𝐡𝐞𝐞𝐥: More users generate more data, which makes your AI smarter and your product stronger. 𝐂𝐨𝐦𝐩𝐞𝐭𝐢𝐭𝐢𝐯𝐞 𝐌𝐨𝐚𝐭: Startups embedding AI deeply win bigger deals, boost upsells, and cut churn. 𝐅𝐮𝐭𝐮𝐫𝐞-𝐏𝐫𝐨𝐨𝐟𝐢𝐧𝐠: Waiting to add AI risks falling behind fast-moving competitors and market leaders. 𝐀𝐜𝐭𝐢𝐨𝐧𝐚𝐛𝐥𝐞 𝐬𝐭𝐞𝐩𝐬: Build your product and data infrastructure with AI baked in from the start. Use AI to free your team for high-value work and improve customer success. Prioritize responsible AI practices focused on privacy and transparency. Embedding AI isn’t just tech,  it’s your growth engine. Start now. Outpace the competition. Lead the SaaS revolution.

One underestimated impact of AI will be the disruption of the SaaS industry. At Essor, we just ran an exercise to optimize our SaaS spend on non-core tools. Our Core Tools — the backbone of our operations (NetSuite, Google, AWS, Snowflake, Zendesk, Rippling, Slack, Klaviyo, Shopify) — account for 60% of our spend. Here, the focus is constant optimization of licenses and resources. But in September, we took a radical stand on the non-core stack. Tool by tool, we either: 1/ Integrated features into one of our core platforms (they’ve evolved fast), 2/ Replaced with open-source deployed on AWS (BI is a great example), 3/ Built internal workflows on N8N, powered by AI-generated services. AI is at the heart of this: Core SaaS tools are innovating at breathtaking speed. Deploying open-source is now painless — the knowledge barrier has collapsed with AI support for code and config. Workflow automation with N8N + AI lets us build fast and cheap. Result: 1/ 40% savings on non-core SaaS spend 2/ 30% reduction in total tools 3/ Direct impact to the bottom line AI isn’t just changing what SaaS can do — it’s changing why you even need SaaS in the first place.

Healthcare SaaS has been selling tools that promised to help—but real transformation will come when software stops assisting and starts doing the job itself. For years, SaaS companies have been selling tools that promise to ease the burden for healthcare teams. But in reality, care teams spend 15+ months in implementation hell, adding more headaches than relief. Most healthcare SaaS products have been a net negative. But now, things are changing With AI, we’re shifting away from Software-as-a-Service and toward Service-as-Software, where the software doesn’t just help—it does the entire job. Here’s what that looks like in healthcare: Before: Your patient engagement tool bombarded patients with generic messages to schedule annual wellness visits. After: AI handles personalized outreach (email, text, mailer) and the company only gets paid when the appointment is booked. Before: Scheduling systems worked 60% of the time, with staff manually handling the rest After: The AI system manages scheduling 100%. If something goes wrong, the software company fixes it or provides backup staff. Before: Denials management software flagged issues and sent them back to your team to sort out. After: AI receives the denial, understands what went wrong, fixes it, and ensures the claim is processed—end-to-end. We’re no longer selling tools that just make the team’s job a little easier. With AI, we’re talking about software that owns the entire process—and if it fails, the software company is on the hook. The Service-as-Software revolution is coming, and it’s going to bend the cost curve in ways traditional SaaS never could.

A lot of GTM automation is going to hurt business outcomes before people realize what happened. They will then roll it back. A great example of this I read about came from Elon Musk. They automated processes at Tesla that should not have existed in the first place. His algorithm has automation last after questioning every requirement, deleting process steps, and refining. They ended up automating too early and had to pull back. The same mistake is happening right now in SaaS GTM. Everyone is bragging about automation. Very few understand the implications of what they automate. Automating CRM data entry may produce cleaner CRM data. It also produces AEs who are less familiar with their prospects, their customers, and their deals. Updating CRM is not just admin work. It's a forcing function for seller thinking. Writing things down forces a rep to better understand. Automating deal coaching may create more feedback. But more feedback is not the same as better coaching. AI may flag talk ratios, competitor mentions, missing next steps, and stage progression. But coaching requires judgment. It requires understanding confidence, customer politics, urgency, motivation, deal quality, and whether the seller actually understands what is happening. I'm a proponent of using tools to aide productivity but too much is too much. If you don’t understand the implications of what you’re about to automate, Walter White said it best: Tread lightly.

Something important is happening in customer support. But because it sounds technical, most people just scroll past. So let's break it down. The support team at Craft Docs cut their ticket handling time by 10x. Not by hiring more people. Not by replying faster. Not by "working harder". But by building workflows that start with one scary word: 𝗔𝗣𝗜𝘀. This is usually the moment non-technical support leaders think "APIs? Not for me" 😬 But APIs are simply what makes automation possible across your stack. "What plan is this customer on?" → API asks the billing system. "Is this bug real?" → API checks the codebase. "Engineering needs to know about this." → API creates the ticket. Craft Docs had been using Zendesk for 5 years. Zendesk wanted to charge them $20,000/year for an AI integration that didn't feel powerful at all. So they built their own workflows. Now Zendesk is just a backend. All the actual work runs through their own tools. And they're probably leaving Zendesk entirely... because what's the point in staying with a vendor that doesn't let you build? That’s why I’m going to say the quiet part out loud: AI in support won’t compound if your tools don’t connect. As Balint Orosz, Craft's founder, puts it: "A lot of UI-first SaaS will be disrupted by API-first players." This is exactly why I’m excited about what we’re building at Plain. A future where support teams can design workflows across tools, not get trapped inside one. If APIs still feel intimidating, start with this question: When you choose a tool, does it help you move data and actions across your stack, or does it keep everything locked inside? If this resonated and you’re not sure where to start, DM me. I see what customers are building with Plain’s API every day. Full story by Gergely Orosz: https://lnkd.in/e_Pct8yW

This week I automated the process of identifying which clients are wrapping up a training and do not have anything else scheduled with us afterward. This week I built a small Power Automate flow that solves a problem we kept bumping into, but never took the time to automate. We store all of our client trainings in a single SharePoint list. Past, present, and future sessions all live together. The data was there, but the insight was not. The question we wanted to answer was simple: → Which clients are finishing a training this month and do not have anything else scheduled with us afterward? Manually, that meant filtering dates, scanning company names, cross checking future sessions, and then writing a follow up email. It worked, but it never happened as consistently as it should. So I automated it. Here is what the flow does: 1. First, it runs automatically on the first of every month. 2. It pulls all trainings that occur during the current month from SharePoint. 3. From there, it evaluates each company on that list and checks whether they have any trainings scheduled after the current month. If they do, the flow ignores them. 4. If they do not, the automation captures the company name and the name of their most recent training session and formats the results into a clean bulleted list. 5. Finally, it sends an email to our Director of Client Services with that list included in the body. Each bullet shows the company name and their latest training so follow up conversations are grounded in context. The email also includes a link to our full training library so she can easily dig deeper if needed. The outcome is simple but powerful. ★ Leadership gets a proactive view of clients who may need follow up. ★ Client services can prioritize outreach without pulling reports. No one has to remember to run a manual check every month. This is a good example of how automation does not need to be flashy to be valuable. Sometimes the best flows just make sure the right information reaches the right person at the right time, every time. If you are sitting on good data but still relying on reminders and manual checks, this is usually a sign there is an automation opportunity waiting. Let’s start building!