August 3, 2026
On February 17, 2026, Retool published its '2026 Build vs. Buy Report,' surveying 817 of its own builders and customers from late 2025. We're not going to pretend this is neutral market research -- Retool sells low-code tooling, and this report is effectively a sales document. But that's exactly why the numbers are worth reading closely. When a vendor whose business model depends on people building custom software finds that 35% of its own customers have already replaced at least one SaaS tool with something they built themselves, and 78% expect to build more in 2026, that's not marketing spin working in Retool's favor by accident. It's a signal about where an entire category of buyer behavior is heading, reported by the company closest to it.
The headline numbers get the attention, but the category-level replacement pressure is more useful for planning. Per Retool's report: workflow automation tools face the most replacement pressure at 35%, followed by internal admin tools at 33%, BI tools at 29%, CRMs and form builders at 25%, project management at 23%, and customer support at 21%. That ordering isn't random. It roughly tracks how bespoke each category's logic tends to be versus how commoditized the underlying tech is. Workflow automation and internal admin tools are frequently just database views, approval steps, and status fields wrapped around a company's specific process -- there's no deep moat in the SaaS version, which is exactly why our own Internal Tools & Admin customers tend to start there. CRM and customer support sit lower on the list, and we think that's honest: those categories carry more integration surface area and more institutional data gravity, which raises the cost of getting a rebuild wrong.
Retool's report includes specific voices worth sitting with. Borys Aptekar, GTM AI Product Manager at ClickUp, said the team realized they could build these tools themselves and save on multiple subscriptions -- a straightforward cost argument, the same one driving most of the SaaS replacement conversations we see. Miles Konstantin, Head of Automation at Harmonic, gave a different reason entirely: he rebuilt a product inside Retool because a vendor's support was so slow that rebuilding was faster than waiting. That's not a cost story, it's a control story -- teams tired of being stuck in someone else's support queue for their own operational needs.
Balance matters here, and Retool's report includes it. Pierre Yves Calloc'h of Pernod Ricard is quoted saying there's no way you can go live with a vibe-coded solution. We're not going to wave that away, because he's right about a real category of problem. Enterprise-scale, heavily regulated, or genuinely mission-critical systems -- think global supply chain orchestration, financial close processes with audit obligations, anything touching regulated data at scale -- deserve real caution, real review, and often real engineering. Our own Buy vs. Build vs. ViibeStack breakdown and our post on 5 signs you should buy the SaaS, not vibecode your own tool both make this same point: build-your-own isn't a religion, it's a tool with a scope. The data itself backs this up -- 35% replacement and 78% planning more is significant, but it's not 90%, and it's not universal. This is uneven adoption concentrated in specific categories, not a wholesale abandonment of SaaS. Calloc'h's skepticism and Aptekar's enthusiasm aren't actually contradicting each other -- they're describing different categories of problem.
Here's where we have a genuinely sharp opinion, and it's not just competitive noise. Retool is a low-code platform. Every one of those 817 builders still had to assemble their tool on a visual canvas -- drag components, wire up queries, configure logic blocks. That's a real skill floor. It's lower than writing code from scratch, but it's not zero. And even with that floor in place, Retool is reporting 35% replacement and 60% of builders shipping tools outside IT oversight entirely. If that's what a visual-canvas skill requirement produces, the natural question is what happens when you remove the canvas altogether and let someone describe what they need in plain English. That's the entire premise behind ViibeStack's AI App Builder -- and we'd expect the categories where Retool's builders already feel confident to be exactly the categories where a prompt-based approach removes the last real barrier, not a new one.
Given the category breakdown, our advice for a team with no engineers on staff is to follow the pressure data, not fight it. Start with workflow automation and internal admin tools -- the two highest-replacement categories in Retool's own numbers, and the two where the logic is usually just your company's process, not a technical moat. Our Workflow Automation and internal tools resources are built around exactly this starting point. BI and project management are reasonable second moves once the first tool is live and trusted -- see our guide on migrating spreadsheets to a ViibeStack app if that's your current BI setup. CRM is winnable, but treat it as a deliberate project rather than a weekend build -- we've written specifically about building your own CRM without engineers because the data gravity there is real. Customer support is the category we'd wait on longest for anything customer-facing at scale -- it's last on Retool's own list for a reason, and it's the one place where Calloc'h's caution applies most directly.
Don't treat this report as permission to rebuild everything. Treat it as evidence that the two highest-pressure categories -- workflow automation and internal admin -- are now safe enough, common enough, and proven enough that waiting has become the riskier choice, not the safer one. Pick one tool in one of those two categories this quarter, the one costing you the most in subscription fees or support friction, and replace it. Everything else on Retool's list can wait for the second or third round.
Sources