System sustainability, availability and AI – where to focus development for long‑term value

The product and team have seen major developments in this time. We have consultants with far more experience and higher-quality solutions, enabling us to implement far more complex projects and deliver standard projects more quickly. When I arrived, we were just carrying out our first automation projects – today, these are our bread and butter. We work with customers that have high demands on application availability, work with a large volume of data and expect modern and robust solutions.

System sustainability is among our key topics. We want our system to support our customers’ business, both today and in ten or twenty years. It’s important for this solution to grow together with a company and be prepared for their evolution. And we’re able to cover all of this thanks to our configurability, modularity, transition to the cloud and more.

When I arrived, we were working on version 7. My task was to finish its development and deploy it to production for our customers. One of our main goals was to have our application in the cloud – in AWS – starting from this version. Our route was to use Kubernetes and break the system down into services.

Then we started modularising the application’s monolithic core and separating the front end, that is, the application’s display portion, from its business logic. Honestly, I expected it all to go more quickly. But I soon realised that with a system this large, you can’t just turn the rudder – it takes a while for the ship to truly change direction.

One key part of this was internal education – ensuring we all knew the new architecture’s impacts. Overall we focused on internal team communication. With sixty developers “on deck”, there’s no other way. When I took over the division, we were half that number.

We’re also adding language models – chatbots – for our customers; these can advise all day and in multiple languages, instead of our consultants. And we’re considering Copilot for configuring transaction processes. We optimise processes with mathematical solvers, which can effectively handle the most complex logistics and manufacturing tasks. They find the fastest or least labour-intensive solutions in the order of seconds, respecting all constraints.

From a developer perspective, we’ve been using AI and some Copilot alternatives for several years, for example for coding or making automated tests.

A customer with their information systems and applications in the cloud doesn’t have to take care of the operation and maintenance. That frees up the IT team for activities that directly support the customer’s business. We’re the ones managing their applications and integrating them into other systems. Because Aimtec uses the AWS platform, high availability and downtime-free operation – backed by a full 99.9% uptime guarantee – are ensured. Security is another major benefit. We have specific cases of on-prem customers where their server burned up or they suffered a “successful” cyberattack, and then returning the system to its original state was very complicated. Additionally, if a customer wants to spin up new applications or expand their system, this takes very little time in a cloud environment.

Yes, we run penetration tests, handled by a specialised, experienced external firm. Their people work as hackers and specialise in the identification of application vulnerabilities. Their output is a detailed list of weak points that could potentially be abused. But we don’t verify security using this method alone. We also regularly utilise vulnerability monitoring software, and we have an internal security expert who actively oversees our overall strategy.

We’ve set up automated monitoring to enable us to maintain the high cloud availability mentioned earlier. We monitor the application’s status in real time, and if a problem is detected, the system instantly alerts our 24/7 support, which starts working on it. This often lets us discover problems before the customer even registers them. We have predefined scenarios that can appear, and we respond as they arise. Automated tests help us as well. So we can put each new version into production in good conscience. To support our long-term platform sustainability, we’re transitioning to PostgreSQL – an open-source database with an active community. PostgreSQL is used by a number of large companies. And more importantly, it greatly simplifies the high-availability configuration we’re talking about.

We focus on two main directions for AI – its use for customers, and its use for the developers creating the product. In that first direction, we’re implementing machine vision and image recognition for quality control and automated stocktaking. During quality control, the system analyses photographs to auto-identify e.g. defects, incorrect shapes, colour deviations and missing parts right at a work centre. How automated stocktaking works is that cameras in the warehouse continuously monitor stock levels, automatically emit a refill request when they drop and can check inventory based on the pictures they take.

We’re also adding language models – chatbots – for our customers; these can advise all day and in multiple languages, instead of our consultants. And we’re considering Copilot for configuring transaction processes. We optimise processes with mathematical solvers, which can effectively handle the most complex logistics and manufacturing tasks. They find the fastest or least labour-intensive solutions in the order of seconds, respecting all constraints.

From a developer perspective, we’ve been using AI and some Copilot alternatives for several years, for example for coding or making automated tests.