How Keene Systems Uses AI to Build Business Software

Keene Systems uses AI throughout a project: to read and explain old systems, to draft specifications, and to assist our developers as they design, write and test code. It makes experienced people faster and more thorough. It does not replace them. Every requirement is confirmed with you, and every line of code is reviewed by a developer who is responsible for it.

Analyzing old systems

Most modernization projects begin with a system that nobody fully understands any more. The person who wrote it has gone, there is no documentation, and the business rules live in tens of thousands of lines of VB6, Access, FoxPro or older .NET code.

Reading that code used to be the slowest part of the job. A developer would work through it screen by screen for weeks before anyone could say with confidence what the system did. AI changes that step more than any other:

• It reads the whole system. Every form, every query, every report and every stored procedure, not a sample.
• It produces a map. What screens exist, which tables they touch, what each report shows and where its numbers come from.
• It surfaces the business rules. The discount that applies only to certain customers, the status that blocks an order, the calculation written down nowhere else.
• It flags the oddities. Code that is never called, two routines that do nearly the same thing, a rule that contradicts another.

Our developers then check that map against the running system and against the people who use it. The result is the understanding a project needs, reached sooner and with fewer rules missed. See our legacy application modernization services and taking over undocumented code.

Where AI helps most

• A system nobody fully understands any more.
• Code with no documentation and no original developer to ask.
• Hundreds of screens, reports or stored procedures to inventory.
• Requirements scattered across emails, spreadsheets and people’s heads.
• Conversion work, such as moving an application from one technology to another.
• A project where an early, accurate picture of the scope matters.

Writing specifications

A specification says, in plain English, what the software must do. It is the document you approve before anything is built, so its quality decides whether the finished system fits.

Our business analysts still do the part that matters most: sitting with your people and learning how the work is really done. AI helps with what follows.

• Drafting. Interview notes, existing screens, sample reports and the analysis of the old system are turned into a structured first draft quickly.
• Checking for gaps. What happens when an order is cancelled halfway? Who may change a price? AI is good at asking the awkward questions a tired reader skips.
• Checking for contradictions. Two requirements that cannot both be true are found on paper, not in testing.
• Keeping it current. When a decision changes, every part of the document it touches is updated together.

The analyst edits the draft, and you review it. You see what will be built sooner, and in more detail. See business analysis.

Assisting our developers

Our developers use AI the way a good tradesman uses power tools. The skill and the judgment are theirs. The tool removes the slow, repetitive part of the work.

• Writing code. Routine screens, data access and conversions from one technology to another are drafted by AI and finished by a developer.
• Writing tests. More of the system is covered by automated tests, because writing them is no longer a chore that gets cut when time is short.
• Reviewing code. A second pair of eyes on every change, looking for defects, security weaknesses and cases that were not handled.
• Explaining unfamiliar code. A developer joining a project, or returning to it after a year, is productive in hours.
• Documentation. Written as the system is built, so it exists when someone needs it.

The developer decides what goes into your system and answers for it. Code review, testing and your own acceptance testing all still happen.

What AI does not do

It is worth being plain about the limits.

• It does not decide what your business needs. That comes from conversations between your people and ours.
• It does not replace experience. Knowing which of three workable designs will still be sound in ten years is a human judgment. We have been making it since 1987.
• It is not trusted blindly. AI can be confidently wrong. Nothing it produces reaches your system without being read, tested and approved by a developer.
• It does not shorten testing. Faster building leaves more time for testing, not less.

Used this way, AI raises the quality of the work as well as its speed.

A recent example

A company had run an inventory and purchasing application on desktop PCs for many years. It worked, but it could not run in a browser, its reports failed regularly, and nobody was left who knew the code.

We used AI to read the old application and its report files and to catalogue what each part did. Sixteen report files turned out to be nine reports that people still ran. That analysis became the specification, which the client’s staff reviewed and corrected.

The system was then rebuilt as a web application on current .NET, with our developers using AI assistance throughout and checking every screen and report against the old one. The reading and cataloguing that normally opens a project like this took a fraction of the usual time, and the client was using the new system alongside the old one while it was still being finished.

The reporting side of that project is described on our Crystal Reports replacement page.

What it means for you

You understand your own system sooner. The first deliverable of a modernization project is a clear account of what you have today.

Projects move faster. Less time is spent on reading, retyping and routine code, so more goes into the parts that are specific to your business.

The work is more thorough. More of the old system is examined, more cases are specified and more of the new system is tested.

The people are the same. You deal with our analysts and developers, not with a tool. The team that builds your system is the team that comes back when you want the next feature.

The code is yours. Our contract states that you own what we produce for you.

How AI fits into a Keene Systems project

1. We talk with you first. What the business needs is settled by people, in conversation.
2. We analyze what exists. AI reads the old system; our developers verify what it found against the real thing.
3. We write the specification. Drafted with AI, edited by an analyst, reviewed and approved by you.
4. We build with AI assistance. Developers direct the work, review every change and test it.
5. You test and accept. Weekly demonstrations along the way, and your own acceptance testing before go-live.

The next step is a short, no-pressure conversation. Schedule a project exploration call or use the form on this page.