AI is making software faster and easier to build but deciding what is worth building remains a business question. A recent Forbes article featuring Helmes explored how AI is changing the traditional build-versus-buy decision. For Helmes, the starting point is the business process: understanding the problem, the existing systems and data, and where AI can create real value, including through focused tools that work with the systems and processes companies already rely on.
For years, the enterprise software decision appeared to come down to two choices: buy an established application or commission a custom one. Generative AI is creating a third possibility – companies can build focused AI solutions around the systems and processes they already use.
That possibility runs through a recent Forbes article, “From Dashboards to Decisions, AI is Rewriting the SaaS Playbook”, in which contributor Ron Schmelzer shows how the software-as-a-service market is changing because of generative AI. The story features Helmes Lithuania CEO and Partner Šarūnas Putrius and Helmes Senior AI Consultant Bart Kappel. Their experience points to a distinction that can get lost amid promises of faster development: being able to build a tool is not the same as having a good reason to build it. The real value still comes from understanding the business process and making something that can be trusted and maintained.
A common case is a company whose existing applications perform nearly every required task, except for one troublesome handoff between systems. Staff members may patch the gap with an AI-generated script or spreadsheet. That fix might work until the source data changes, an unusual case appears, or the person who created the workaround moves on, and the institutional knowledge they had is lost. What looked like a shortcut becomes another system to support.
For Helmes, the work, therefore, starts before a developer types a prompt. Our teams examine where information originates, who needs access, how a proposed tool will connect to existing platforms, what could go wrong, and who will maintain it. AI can accelerate coding, testing, and documentation. It can’t establish business priorities or make a fragile design dependable.
That approach can make relatively small applications valuable. As Forbes reports, one Helmes-developed assistant searches more than 50,000 service manuals and answers questions using an organization’s technical documentation. Another helps prepare structured audit findings from documents, with people at the organization reviewing the results. Neither attempts to replace a company’s wider software environment. Each addresses a specific task within it.
But before choosing even a small tool, there is a decision to make. Some gaps can be closed by configuring existing software; others warrant a custom integration or AI capability. Some tasks may not justify automation at all. Helmes’ AI Value Scan service, for example, makes that distinction by examining business processes, people, data, and existing technology before ranking AI alternatives by expected impact, feasibility, effort, and strategic fit.
In one engagement with a global auditing company in 2025, Helmes considered 14 AI initiatives and selected three priorities, delivering a decision-ready roadmap in four working weeks. The first AI use case went live in under three months and was then scaled across the organisation. In 2026, with a mid-sized manufacturer, Helmes narrowed 14 possible use cases to four priorities, complemented with a 90-day action plan. The result in each case was not a longer list of things to develop, but a clearer set of choices the business could act on.
This is the opportunity behind the shift Forbes describes; companies don’t need to assume every software gap requires another large subscription or replacing trusted systems. They have more options to build selectively, as long as the result fits their processes and can withstand daily use.
AI has sped up and expanded what companies can build. The harder question is what they should.