Let’s ideate over coffee.
Manual software configuration is still the norm — but not for long.
Most enterprise software ships as a minimal foundation. The real capability lives in the ecosystem of plugins, integrations, and settings layered on top. Getting from “purchased” to “productive” requires niche platform expertise that most teams simply don’t have.
Three to six weeks per platform. Multiply that across every SaaS tool a modern organization runs, and the “configuration tax” becomes staggering.
We buy software for the outcomes, yet setup stands aggressively between purchase and value: invisible, underestimated, and almost entirely manual.
That’s about to change, driven by the convergence of four maturing capabilities:
- Contextual understanding: AI extracting requirements from plain conversation — no platform jargon required.
- Direct execution: Agents calling APIs, browsing marketplaces, and executing steps directly, not just making suggestions.
- Multi-step planning: Breaking complex goals into ordered steps and self-correcting when errors pop up midway.
- Human oversight: Gating high-risk actions behind clear approvals so speed never compromises control.
Open standards like the Model Context Protocol (MCP) turn this from a neat demo into production-ready tech by giving agents and platforms a common language.
The Trust Model Is Key
Read-only actions (auditing, browsing) can run freely. However, configuration changes need reviewable, human-readable summaries before execution — and high-risk tweaks deserve dry runs. Every action must end up in a clean audit log so rollbacks feel like a single operation, not an archaeology project.
The Opportunity Ahead
A few forward-thinking vendors have started opening up AI-assisted setup, but it’s still the exception. Most platforms expose solid APIs for workflows, but rarely for admin tasks like installing extensions or applying templates.
Vendors that open this surface aren’t taking an architectural gamble; they are making a strategic product decision. Those who make autonomous setup the default will define the next decade of enterprise software.
The future isn’t a faster wizard or a smarter checklist. It’s software that onboards the way a great colleague would: ask a few sharp questions, propose a plan, get a nod, and execute.
No six-week delays. No specialized consultants on standby. Just software that configures itself.
The question isn’t if this happens — it’s who builds it first.
Disclaimer: Drafted alongside my AI writing partner to sharpen and articulate my thoughts.
