The state of internal-tool design in 2026, five years after Retool broke through
Retool's Series C in September 2021 was the moment the category clicked. A $3.2B valuation for a company whose pitch was, essentially, *don't design your internal tools — just drag components onto a canvas that runs*. Five years later, that pitch is table stakes for one whole camp of internal-tool builders, and the industry has quietly split into three approaches that make wildly different bets about where the expensive work actually lives. If you're shipping admin dashboards, ops consoles, or ERP interfaces for a mid-market SaaS company in 2026, you're picking between: 1. **Build-first tools** — Retool, Superblocks, Toolpad, Appsmith. Skip the design phase; write config-and-glue-code inside the platform's canvas. 2. **Design-first with AI compression** — dxmax, Uizard, Galileo. Get the wireframe right in minutes, then hand off to whatever backend framework the team already uses. 3. **Framework escape hatches** — plain Next.js/React with a component library (shadcn/ui, MUI, Ant Design), often when the internal-tool volume outgrew a low-code builder. The interesting shift since 2021 isn't which camp is winning. It's that the design-first camp — which everyone assumed Retool had killed — is back. Not because Retool is worse than it was; because AI wireframing did to the design phase what Retool did to the build phase: compressed a multi-day cycle into minutes. ## The Retool bet: designers are the bottleneck, so skip them The original Retool argument was crisp. Internal too