Right now, AI is in its pure growth area.
Every CIO is deploying more budget to AI solutions in a drive for efficiency and to keep up with the competition. But even if IT budgets continue to grow at their current epic rate, the pace of applications we are buying just can’t last. And that includes AI Agents. Even though in many ways, we’re just getting started.
Just as CIOs have cut the number of traditional SaaS vendors over the past several years, the same will have to come to AI. We can’t all manage 200+ AI agents in production – and pay for all of them. It’s too many to track and manage.
Even tiny Team SaaStr is now spending a nominal $500K annually on our 11 core AI agents. When in essence a bootstrapped team with a handful of employees hits that spend level, you know the economics are ultimately going to shift dramatically across the enterprise.
Is that cheaper than the headcount we’ve replace with AI? Yes. Yes … But. Yes, but in the aggregate it’s still substantial. And subject to consolidation risk. Especially at the CFO/CIO level.
The AI startup ecosystem is heading for a reckoning. While CEOs are throwing budget at AI projects like never before, we’re likely to see a wave of consolidation that could happen faster and more dramatically than the SaaS shakeout of 2022.
When We All Have 100+ AI Agents Per Department, What Happens Next? Consolidation.
The 2022-2025 SaaS Consolidation Playbook
We’ve already seen this movie in traditional SaaS. Since 2021, companies have been aggressively cutting their application portfolios. The average enterprise reduced SaaS app count by 18% between 2022-2023, according to Zylo data, and isn’t adding any more net now:
The pattern is consistent: CFOs and operations teams gather annually with one mandate – consolidate. What started as “best-of-breed” strategies in 2019-2020 became “platform fatigue” by 2022. Integration costs, security reviews, and vendor management overhead killed the ROI on point solutions.
Salesforce’s $27.7B Slack acquisition and Microsoft’s push into every workflow aren’t coincidences. They’re responses to customer demand for fewer vendors, not more features.
The Math Problem No One Is Talking About — Yet
If your company is spending $500K to $1M annually on AI tools per department, you’re going to ultimately have a conversation about consolidation. The “Gongification” we saw in sales tech is coming to AI, but accelerated.
What happened with Gong? When budgets contracted in early 2022, Gong was often cut first. Customers loved Gong. But by then, some of that functionality was also in Salesloft, ZoomInfo, Outreach, etc. CROs didn’t want to give up Gong. But they often had to, since call recording was in other apps.
Fast forward to today? Gong stepped up and rapidly became a platform for revenue management. And became the one they kept. Not the one they cut. More on that here:
The core issue? Everyone thinks their AI tool will save labor costs, but you can’t all get credit for the same labor savings. When five different AI vendors are each asking for $60K-100K+ a year to automate parts of your sales process – SDR tools, BDR tools, conversation intelligence, pipeline tools – the math stops working.
Again, yes it’s cheap than a human. So that gets the budget nod — today. For now. But ultimately, the pressure will come as we spend millions and millions on different, paid AI agents from different vendors that seem to overlap.
The Rippling Model Emerging in AI Agents?
Smart companies are already moving toward the Rippling approach: “Give us five agents and an orchestration layer, and we’ll handle it all.” Why manage five point solutions when one vendor can deliver the complete workflow?
This trend is already visible in e-commerce, where multiple AI tools are rapidly consolidating into single platforms. Sales, support and marketing are becoming one in e-commerce, with one interface and one customer experience. The writing is on the wall for standalone AI startups that can’t keep pace with this integration demand.
Surface Area > ROI Efficiency
The lesson from robotics applies directly to AI: ROI efficiency isn’t enough. You can build an AI tool that delivers incredible ROI and automates a process with 50-75% labor reduction, but if you’re only touching two workstations in a 300-person operation, it doesn’t matter.
Successful AI companies attack the largest percentage of labor in their target workflows. It’s not just about high ROI – you need high ROI on a substantial quantum of headcount. CFOs and operations leaders can only manage so many implementations per year.
What This Means for B2B Startups
The window for standalone AI tools is closing faster than most founders realize. The companies that survive will either:
- Expand their surface area to cover larger portions of workflows
- Get acquired by platforms building comprehensive solutions
- Become the platform that orchestrates multiple AI capabilities
The billion-dollar single-person AI company remains a myth. Economics has a way of competing away excess profits, and when one company offers five integrated AI agents while you offer one, the competitive dynamics shift quickly.
The Takeaway
If you’re building in AI, ask yourself: Are you automating a workstation or transforming a workflow? The companies that answer “workflow” will be the ones writing acquisition checks in 18 months.
The consolidation is coming. The only question is whether you’ll be doing the consolidating or getting consolidated.
Until then, sell the heck out of your AI agent. Just also make sure you are one of the Top 5 your buyer puts in production. The rest are ultimately at risk.
