The way U.S. companies buy software development services is changing.
AI is changing expectations about cost and speed, while businesses have more options for building, buying and modernizing software. The challenge is deciding which investments will create real business value.
My advice is to slow down just enough to make sure the organization is solving the right problem.
Before approving a major software development or modernization initiative, I would ask six questions.
1. What business outcome are we actually trying to improve?
Do not begin with “We need AI.” Begin with the business problem.
Are you trying to reduce operating costs? Improve customer service? Increase revenue? Eliminate manual processes? Improve forecasting? Reduce errors? Replace an unsupported system?
AI may be part of the answer, but it should not be the starting point simply because it is available.
2. Can we improve the existing system instead of replacing it?
Before approving a major replacement project, understand what business logic already exists in the current environment.
Sometimes the best solution is a complete replacement. Sometimes it is modernization. Sometimes it is an API layer, new user interface, data modernization, workflow automation or AI capability built around the existing platform.
The answer should come from the business and technical analysis, not from a predetermined technology preference.
3. Where can AI provide measurable value?
Identify specific workflows where AI can improve productivity, decision-making or customer experience.
- Document processing
- Automated reporting
- Knowledge management
- Customer support
- Estimation
- Forecasting
- Predictive analytics
- Data extraction
- Workflow automation
- Software testing
- Internal search
- AI agents connected to business systems
Then determine what data, integrations, governance and human oversight are required to make that use case reliable.
4. What should we build internally, buy, or outsource?
There is no universal answer.
The right decision depends on the strategic importance of the capability, existing internal expertise, time-to-market requirements, cost, scalability and long-term ownership.
External engineering can be particularly valuable when a company needs specialized expertise or additional capacity without committing to building an entire internal organization.
5. How will we measure whether the project succeeded?
Define measurable outcomes before development begins.
- Reduction in manual processing time
- Faster customer response
- Lower operating cost
- Improved forecast accuracy
- Increased revenue
- Reduced system downtime
- Faster release cycles
- Improved employee productivity
- Higher customer satisfaction
If the business cannot explain what success looks like, the project probably needs more discovery before development begins.
6. Is the technology foundation ready for what we want to do next?
This is particularly important with AI.
AI capabilities are only as useful as the data, architecture, security, integrations and business processes surrounding them.
A company does not necessarily need to rebuild everything before adopting AI. But it should understand whether its current environment can support the use cases it wants to pursue.
Discuss your next software investment
Talk with Shinetech about your business priorities, existing systems and the engineering work needed to move forward.
Sources and Further Reading
Sources reviewed September 23, 2026.
Further reading selected by the Shinetech editorial team to accompany John’s six questions. These resources were not cited in his original text.
- GOV.UK Service Manual, How the discovery phase works. A public-service guide to understanding needs and constraints before deciding whether to proceed. Included here as a reference for discovery methods.
- NIST, AI Risk Management Framework. A voluntary framework for considering trustworthiness and risk when designing, using and evaluating AI systems.
- Shinetech Software, Who Takes Responsibility for Your Software After Launch? Questions about support, continuity and ongoing ownership.
Illustration created with AI for this article. It depicts a conceptual scenario, not a real client project.