The end of the year and the beginning of the new one is a time for reflection, summing up results, and making plans. Let's recall the predictions we made for 2024 and see how accurate they turned out to be.

Artificial Intelligence continues to hold the leading position. As we predicted, this is a long-term trend, and we will be observing its evolution in all spheres of our lives for many years to come. A clear example for several years now has been medicine: AI helps improve diagnostic mechanisms, optimize the process of creating new drugs, and generally contributes to improved quality of medical services while reducing costs. For example, one of our long-standing clients, a European medical software producer, has been using various tools for recognizing doctors' prescriptions for a long time, but the application of AI has taken them to a new level: if before, special techniques were needed to change image contrast, adjust lighting, and recognize text, now all that's needed is to correctly write a script. The rest will be done by the neural network itself: digitizing the prescription, understanding the meaning and deciphering the medical terminology, checking if the dosage is correctly calculated, and estimating the cost of the medicine. In another project of Noveo, a web portal for one of the most prestigious clinics in France, AI is used to optimize internal processes and reduce the time spent on writing, storing, and managing documents.

However, the possibilities of AI are also widely used in other areas: a wide variety of voice assistants and chatbots have definitely entered our lives; recommendation systems will soon know our preferences better than we do ourselves; automatic translators already allow us to understand the meaning of any text in nearly any language in the world, and many believe that the profession itself will eventually disappear. A similar fate is predicted for legal consultants, based on systems similar to our recent project: an application "fed" which a large volume of legal documents, based on which the neural network gives clear, evidence-based answers even to rather vague user questions.

The market for AI assistant applications is growing, capturing more and more areas of usage – and Noveo is not standing still! We have almost finished (the project is at the proof of concept stage) developing our own smart messenger: a web application based on Gemma LLM, which parses data from different sources (emails and messengers) and makes beautiful and easily readable summaries from them. Such an assistant allows you to easily navigate the avalanche of information that falls upon us every day, minimizes cases of forgotten or missed messages, and helps to allocate time more efficiently. We have already appreciated the capabilities of the messenger and are preparing to launch it on the market as our own product – after all, each of us deserves simple and convenient communications!

Continuing the thought about AI expansion: its current state of development allows it to be used in areas where human potential is limited – for example, in maritime transport. And this is not only about logistics and accounting optimization, which has become almost routine: we are currently working on a system that assesses the integrity and configuration of secured containers on a sea vessel and predicts potential problems with further transportation. These capabilities are available to us thanks to a combination of drone footage and the analytical skills of modern systems.

So, our bet on AI development paid off – but not only that: requests for data migration to the cloud also meet our expectations – we predicted their further growth and were not mistaken. Amazon Web Services still leads the list of cloud service providers, and being certified AWS partners, we see firsthand how the number of migration projects continues to increase.

This trend is becoming increasingly large-scale: today, not only data is migrating – entire spheres of activity are "moving." An obvious example is education: the process of its transition to remote mode began during the pandemic and continues to gain momentum today. Not only courses and materials are moving online – exam taking is now often provided only remotely too. And new forms give rise to new challenges: deepfake technology allows anyone to take an exam for someone else, even via video. Online platforms will have to compete with this dark side of AI (yes, it’s again about it): they must be prepared to recognize such situations and actively combat them.

However, not only education is moving to remote work: many tasks in various business areas can now be performed online, and the mass exodus of employees to home offices is one of the characteristics of our time. The remote work format has firmly taken root in our lives and processes, and many companies are not trying to fight the current, but are directing it: they make remote onboarding of new employees as simple and clear as possible using specially developed tools and services, measure the time spent on work tasks, and gather employees live only for joint events such as conferences and team building activities. Having caught this trend, one of the world's most famous hotel chains entrusted Noveo with the development of an application for remote booking of conference rooms in its hotels for such events.

Let's return to the topic of data migration to the cloud: this year, a new, opposite request has emerged – for data migration from the cloud back to hardware. Thus, our constant partner (an international IT company specializing in data storage solutions) entrusted us with a project on integration with devices for creators: they usually collect very large amounts of data, and sometimes it is more convenient if they are right at hand, without having to go to the cloud for them. And you don't need to worry that the devices won't cope with the scale of work or overheat – the customer is implementing an AI-based system for its devices (yes, we can't do without it talking about trends): if the device starts to overheat, make noise, or malfunction, data about this is automatically sent to support, and the user can remotely receive recommendations for correction or contacting a service center. The request is essentially  interesting, but whether it will become a new trend – time will tell.

And let's close the topic of AI (this time for good, we promise): if, placing your hopes on systems like Github Copilot or Codeium, you are waiting for the programmers to be displaced by all this magnificence they’ve created, we will disappoint you – programmers will be in demand for a long time to come. Yes, code generation tools are gaining momentum, but they are still too young for full-fledged assistance. From the experience of our developers, using assistants for routine tasks reduces programming time by a maximum of 15%: they solve problems by far not optimally, and often generate non-working code altogether. Solutions in the field of AI-driven Design and Development – Vercel, Builder.io, Locofy – are interesting, but there are still more limitations than real benefits. We will be watching their development in 2025!

If we talk not about global trends, but about specific technologies, then no miracles happened, and our predictions came true 100%:

• On the frontend in web development, React still prevails, while Vue failed to become a leader and is slowly losing ground. PHP (mostly Symfony, less Laravel) is most often used for the backend, followed by Node.js (mainly the Nest.js framework) and to a lesser extent Python.
• Testing continues to focus on automation – Python is often chosen for this purpose, but Java is not far behind.
• Native mobile development noticeably loses to cross-platform solutions on React Native and Flutter; native development can now be found mainly in giant companies in various sectors.
• An interesting discovery of the outgoing year for us was the MAUI framework based on .NET: we tested it by implementing for a large European network of HR agencies a number of applications for accounting and control of shift work in two business areas: HoReCa and cleaning. Each application is created separately for a specific group of users (managers, employees, consumers of services) and allows them to quickly and easily solve all tasks on their phone. For example, applications for managers typically include a calendar showing employees availability, an internal messenger, a task tracker with the ability to assign tasks to employees, etc. After working with this framework, we assess it as a very successful solution and expect new projects on it.

Summing up, we come to a rather pleasant conclusion: we turned out to be quite decent forecasters! Will our expectations for 2025 be as insightful? We'll find out in a year ;)