Some professional-services work is already being handled by AI. If your next investment is in planning, review or judgment, start with what you need that capability to accomplish.
It might make an existing service better: more accurate plans, fewer omissions, less rework or lower delivery costs. Or it might become a service clients pay for separately, with the planning or review itself as the outcome they buy.
The same capability can serve both purposes. Better delivery may help a separate service take shape. But improving an existing project doesn’t, by itself, show that clients want to buy the planning involved as a distinct offering. A new paid service still needs evidence that its delivery is valuable and that customers want to purchase it separately.
That distinction matters when deciding where the next investment belongs.
What the tool already covers
Before investing in planning as work that sits outside AI tools, look closely at which parts those tools already cover—and which still require human effort.
FastRuby.io’s own free and paid offerings provide a concrete example.
According to its README, the free Claude Code Rails Upgrade Assistant skill can analyze an actual codebase and generate upgrade reports, provide before-and-after code examples, preview configuration changes and plan a migration step by step across versions. “Free” here refers to the published skill, not a claim that running it is free.
The README also describes limits. Deployment strategies are outside its scope. Investigating complex runtime patches and resolving dependency conflicts require manual work. Time estimates are rough guidelines. The skill supplies guidance and examples; the user implements the upgrade.
The same team’s paid The Roadmap to Upgrade Rails is described on its official page as an application-specific audit. Two developers each estimate the best- and worst-case time and cost for every version upgrade. They combine scripts with manual checks of gems, deprecation warnings, technical debt and upgrade obstacles to produce a 15-to-50-page, step-by-step plan.
Clients can use that plan to carry out the upgrade themselves or ask FastRuby.io to perform it.
Based on these official descriptions, the free skill and the paid Roadmap partly overlap in generating upgrade reports and plans from a codebase. That overlap does not establish equal quality or capability. It doesn’t show that the skill has replaced the Roadmap, or that clients will necessarily continue buying the paid service.
These are the supplier’s descriptions of its offerings—not independently verified performance, substitution, customer demand or business results. No tool comparison tests or customer purchase data were collected for this discussion.
Still, the same provider has put part of its upgrade-planning work into its own free tool. “AI handles execution; planning and judgment stay outside” is no longer enough to identify which specific capability deserves further investment.
And work outside a tool’s scope is not automatically work clients will buy.
Match the evidence to the investment
If the investment is intended to improve existing upgrade delivery, start with the projects you already do.
Are upgrade plans becoming more accurate? Are there fewer omissions? Is manual checking or rework falling? How much additional effort is still needed for work the tool doesn’t cover?
Those changes can support continued investment in an existing service. They help explain whether the capability is making delivery better, while leaving room to consider the effort and cost required to achieve those gains.
If you also want to sell the plan separately, there’s another question: why would clients choose that purchase rather than use available tools, assign an internal team or seek help elsewhere?
The answer needs to be tested against clients’ actual choices and their responses when offered something to buy. Uncovered technical work may help explain the value of an offering, but its existence alone doesn’t establish demand.
That doesn’t mean a small experiment must wait for completed sales. It means customer response is evidence in its own right. More accurate plans and less rework cannot stand in for it, and delivery improvements don’t justify further investment at any cost.
For an upgrade-planning service, both lines of inquiry can develop together. Improving plans within existing projects may reveal a useful separate offering. Testing that offering may clarify what clients value about the work. Neither possibility removes the need to show that the service delivers something useful and that clients want to buy it in that form.
Making delivery better can support investment in the service you already provide. It does not, by itself, establish that a new paid service is worth building.
Top comments (0)