We’ve been keeping a list.
For six months, every conversation we’ve had with a founder or operator about software has landed in one of three buckets. Every one of them. Charter school directors, Shopify merchants, SaaS founders, regional services CEOs. Same three conversations. Once you can name them, you stop missing the referrals sitting in your inbox.
A bit of context first.
The software market split in February:
- Anthropic shipped Claude Cowork and roughly $285 billion in SaaS market cap evaporated in 48 hours.
- Atlassian reported its first enterprise seat decline ever and cut 10% of its workforce.
- Freshworks laid off 11% on May 5 and the CEO publicly blamed AI.
- Retool surveyed 817 builders and found 35% had already replaced at least one SaaS tool with a custom build.
- Lovable, meanwhile, hit 8 million users and 100,000 new projects per day, then had three security incidents in 13 months.
Every founder, operator, and CMO you talk to has felt this shift. Some are riding it. Some are scrambling to recover from it. Some are watching it pass them by.
Which means they’re in one of three conversations.
Conversation 1: Rescue
A founder built something themselves. Lovable, Bolt, Cursor, Replit, Claude Code. It works on a demo. They showed it to investors, to their team, to a customer. Now they’re trying to put it in production and the wheels are coming off.
The phrases sound like this. “It works on my laptop but breaks when more than five people use it.” “We pushed the AI to add one more feature and now nothing works.” “We have no idea what this code actually does.” “Our Stripe keys are in the front-end and we just had charges we can’t explain.”
The data is brutal. Veracode’s 2025 GenAI Code Security Report found 45% of AI-generated code samples contain OWASP Top 10 vulnerabilities. A Q1 2026 audit of 200+ vibe-coded apps found 91.5% had at least one AI-hallucination flaw, and more than 60% exposed API keys or database credentials in public repos. Aikido Security now attributes one in five enterprise security breaches directly to AI-generated code.
The tone is defensive, sometimes embarrassed, usually urgent. The founder validated the idea. Making it real is somebody else’s problem now.
A line that lands: “You proved the idea works. The hard part is the next 20%.”
Conversation 2: Replace
The operator is fed up with SaaS.
Their Mailchimp bill keeps going up. Their CRM doesn’t fit how they sell. They’ve got 14 tools that don’t talk to each other. Their team uses 10% of what they’re paying for. Their niche vendor got acquired and the product is going to hell.
Retool’s 2026 Build vs. Buy Shift report quoted Miles Konstantin at Harmonic: “Their support was so slow that it was faster for me to rebuild the product inside Retool than wait for support to get back to me.” Harmonic now runs 33 internal apps replacing third-party tools.
That story isn’t isolated. 78% of Retool’s surveyed builders expect to build more custom internal tools in 2026. 60% have built software outside IT oversight in the past year.
There’s a contrarian view worth knowing. Jason Lemkin at SaaStr argues the SaaSpocalypse is mostly the unwinding of 2021 over-hiring and over-valuation, not AI killing SaaS. Sarah Tavel pointed out that nobody is actually rebuilding Salesforce in Replit. They both have a point.
Nobody is rebuilding their ERP. But the niche tools, the workflow automations, the per-seat admin software? That wave is already moving.
The tone is calm and frustrated. The operator has done the math. They’ve decided to build. The only question is who builds it with them.
A line that lands: “You’re not buying software anymore. You’re buying ownership.”
Conversation 3: Catch-Up
A CEO knows their competitors automated something six months ago. They watched their margin compress. They got a deck from an AI consultant. They tried to hire an AI engineer and got crushed by the salary. They bought Agentforce or Copilot or Breeze and nobody on their team is using it.
The pressure is real. The 2026 SMB AI Outlook reports 57% of US small businesses are now investing in AI tech, up from 36% in 2023. McKinsey estimates agentic AI could unlock $2.9 trillion in productivity by 2030. Founders read these numbers and feel them on their P&L.
Agents are an integration problem dressed up as an AI problem. Gartner found in May 2026 that 80% of companies that reduced workforce via AI saw no correlation with higher ROI. Klarna publicly reversed parts of its AI-first customer service strategy after agents couldn’t handle complex cases. Every operator who said “we’ll just use AI” without fixing their data, their CRM, and their integration plumbing is now quietly rehiring.
The tone is anxious and vision-driven. They want to be ahead. They’ve spent money on the wrong things. They need someone who can connect the systems they already have so the AI they want can act on something real.
A line that lands: “The AI isn’t the hard part. The plumbing is.”
What to do when you hear one
Naming the conversation back to the person is more useful than pitching software.
For rescue: “You proved the idea. The hard part is the next 20%.”
For replace: “You’re not buying software anymore. You’re buying ownership.”
For catch-up: “The AI isn’t the hard part. The plumbing is.”
Do that and you’ve already done more useful work for them than 90% of the AI consultants they’re paying.
If they ask who builds the real thing, send them to us. FYC Labs has been doing custom software for 12 years. Most of our work right now comes through partners who learned to listen for these three.
We’ll take it from there.
by Brian Hammond (VP of Sales, FYC Labs)
More for FYC Labs