As Software Gets Cheaper, Which Problems Become Worth Solving?

An old idea about efficiency

In The Coal Question, first published in 1865, William Stanley Jevons argued that more efficient use of coal could increase total consumption. Lower fuel requirements made industrial activity cheaper, encouraging greater use and opening applications that had previously been uneconomic.

That mechanism, associated with Jevons paradox, suggests a useful question for software. If the cost of solving a problem falls, could more problems become worth solving? This is a conditional analogy; the history of coal does not establish how software demand or profits will change.

Revisit the decision behind the rejected project

A project may have been rejected because its expected value could not justify development costs. If AI-assisted work reduces the effort required for that particular project, the balance may change. The size of any improvement needs to be assessed against the actual requirements.

Ask what stopped the project originally. Was it implementation cost, unreliable data, difficult integration or lack of user demand? Faster development helps only to the extent that it addresses the constraint.

Look for useful work that remained manual

The opportunity may be a narrow workflow serving a relatively small group. Consider these illustrative scenarios, rather than claims about specific Essential deliveries:

• An approval workflow that replaces repeated email chasing with visible responsibilities and status.
• Document processing that prepares extracted information for human review, with exceptions handled explicitly.
• A portal for a small customer group that reduces repeated requests for documents or progress updates.

Test the full cost of the improvement

Estimate what the workflow costs today: time spent, delays, rework and the consequences of errors. Compare that with the proposed improvement over a defined period. Record how many people will use it and what adoption would need to look like for the investment to make sense.

Include integration, data preparation, verification and training alongside implementation. Add recurring hosting, model usage where applicable, maintenance and support. A cheaper prototype does not automatically mean a cheaper maintained application. Compare an extension to existing software and an available SaaS product as well.

More applications bring more responsibilities

Every additional application can introduce dependencies, permissions, accounts and operating costs. Someone must keep it working as surrounding systems change. Generating more code can therefore increase the amount of software a business has to maintain.

Name the owner before launch. Confirm access to code, data and service accounts, define the support process, and make setup and deployment understandable to another engineer. Include a way to retire the application if its value disappears.

Start with the smallest valuable workflow

Choose one task and define a measurable outcome before building. For example, an approval tool might be evaluated against the current time from submission to decision, alongside the number of requests requiring manual follow-up.

Test the important permissions, integrations and failure cases, then observe whether people use the result. Expand when the evidence supports it. This keeps the investment tied to a business improvement and exposes assumptions before they become expensive commitments.

Agree a stopping condition as well. If users do not adopt the workflow, the integration proves too costly or the expected improvement fails to appear, the team should be able to pause or change direction without extending the commitment automatically.

Bring back a good idea with a new assessment

The practical lesson is to reconsider useful ideas when their economics change. Some will become viable; others will still have stronger alternatives.

Essential can help assess a workflow, compare the options and implement an agreed path. Bring the problem, the tools it touches and the reason it was previously set aside. Those are the starting points for deciding whether it now deserves investment.