Almost every company that has been around longer than a few years has an app in its portfolio that people talk about in hushed tones. It supposedly works, but every change takes forever, costs a fortune and carries the risk that something else will break. For a CEO, COO or IT Manager this is not a technical problem — it's a business problem that blocks growth and ties up capital.
The pain every decision-maker knows
The symptoms are always similar. A new, simple feature you promised a customer is estimated at weeks of work. Only one person on the team understands how the app works — and they've just handed in their notice. Maintenance costs rise year on year, while the business value stands still. This is not an outage, it's a slow suffocation.
Why classic IT turns it into a dilemma worth hundreds of thousands
In the traditional model you have two costly paths. The first is continued patching — cheap on the surface, because each successive change is ever more expensive and ever riskier. The second is a rewrite from scratch — a months-long project by a team of expensive programmers, with a budget counted in hundreds of thousands and a real risk that the new version won't even reproduce all the functions of the old one. Both scenarios are painful, because they rest on the same assumption: every line of code requires the time of an expensive specialist.
The decision framework: four criteria
Before you decide, assess the app along four dimensions. It's a simple conversation you can have in a single board meeting.
- Business value — is this app still the heart of the process you earn money on? If so, it's worth investing. If it's a relic, consider winding it down.
- Technical debt — can the code be developed, or is every change a gamble? The deeper the debt, the stronger the case for a thorough rebuild.
- Risk — what happens if the app goes down for a week? The higher the operational risk, the less room for experiments without a contingency plan.
- Cost and time of change — how much does the simplest modification really cost today, and how long does it take? It's the best barometer of the system's health.
The third way: modernisation supported by AI agents
The dichotomy of "patch or rewrite" comes from a time when every change depended solely on the number of hours of expensive programmers. Today that assumption is out of date. At ESKOM.AI we approach modernisation differently: we combine experienced specialists with the support of dozens of specialised AI agents that take over the tedious, repetitive and time-consuming part of the work.
In practice this means that analysing the existing system, mapping the functions, generating new modules and a full set of tests — unit, integration, end-to-end, performance and security — come together in an automated software development process. What once required weeks of tedious work, we now deliver in days, and the total cost is often a fraction of what a classic project would cost.
The concrete: what a good decision looks like
Imagine a company with 80 employees and an order-handling app where changing the price list took three weeks. Instead of rewriting everything from scratch, modernisation covered only the most heavily loaded modules and added the missing integrations. The result: that same price-list change is now a matter of hours, and the budget closed within a predictable, far lower figure than a full rewrite. The key was the right decision at the start — based on the four criteria, not on emotion.
What next?
If you have an app that slows the company down instead of driving it forward, you don't have to guess. Start with a short audit that shows which of the four dimensions the problem lies in — and whether AI-supported modernisation is a faster and cheaper road for you than a classic project. Write to us and together we'll assess the state of your system and present a measurable plan.