Technology Trends for 2025: AI, Governance, The Future of Development, and More
“Any sufficiently advanced technology is indistinguishable from magic.” – Arthur C. Clarke
At Mendix, we’ve spent enough time in the software development space to know innovation and advancement result from years of hard work and iteration. Still, we’ll forgive the idea that new and advanced technologies can seem magical in 2025.
After all, even a few years ago, who could have seen how quickly AI and GenAI would advance in all aspects or how automation would continue to grow, evolve, and change?
Software development and technology continue to evolve and move faster and faster. There’s no such thing as terminal velocity in these spaces. In times like these, it’s essential to keep your feet planted in the now while still looking toward the near and the far future.
To usher in 2025, experts across Mendix gave their opinions on the biggest topics. Here’s what they had to say on AI and GenAI, automation, governance, and the changing face of software development. Let’s take a look.
Artificial intelligence: The known unknown
AI has dominated the enterprise software development space, at least since ChatGPT launched to the public in November 2022. It’s not hard to find opinions on utilizing AI and GenAI in every aspect of an enterprise. However, most enterprises remain at inflection points around use cases, data, and more, while critical questions remain about how AI will affect development.
“AI right now feels in some ways like the new Internet of Things,” says Raymond Kok, CEO. “Everybody’s jumping on it, experimenting, and figuring out the ins and outs of the technology.”
Adds Gordon van Huizen, SVP, Strategy: “Increasingly, developers can focus more directly on the opportunities and problems in the business that they can address. In an enterprise context, product owners can work more closely with people in business teams – who care more about the problem than the technology. Multidisciplinary fusion teams – and “T” shaped people who can collaborate with the business, designers, and technology specialists – will deliver incredible value.”
“2025 might be the year we see enterprises push AI out into their external audiences,” says Sheryl Koenigsberg, VP of Global Product Marketing. “But with that, it’s critical for enterprises to measure and validate the ROI of what they’re doing.”
Raffaello Lepratti, Global Vice President of Industrial Manufacturing, continues, “There is an opportunity to connect GenAI services to data from different systems that typically are operating siloed. So, manufacturers can create a more holistic intelligence, that trained over time, helps increase efficiency and reduce costs. For example, a manufacturer has all information about parts, parameters, and tools used to manufacture a specific product. When this product needs to be inspected GenAI services can create a tailored inspection plan which reduces operation time and avoids potential reworking. There are a lot of costs to save with that approach, but GenAI must evolve and be tested in that context.”
“I’m looking forward to seeing how manufacturers data, that is typically stranded in silo’s influences GenAI use cases,” says Subba Rao, Director, Manufacturing Industries Cloud. “Within production, we’re seeing GenAI use cases move beyond chatbots and into more natural language conversational experiences that extend the value across and beyond factory processes. It’s an exciting space to see continually improve and innovate.
The redevelopment of enterprise software development
AI and GenAI are creating huge questions about the future of software development, but they’re not the only questions out there. What do the next few years of automation look like? Is it time for enterprises to look at development as a whole? After all, according to a recent Mendix survey, 75% of respondents agree that their org’s C-suite believes low-code is “the only option for coding” moving ahead.
Hans de Visser, Chief Product Officer, is excited for the future: “In 2025, we will continue to see the profound impact of AI on the software development life cycle. GenAI will further boost the productivity of traditional developers using coding frameworks. We’re seeing a strength and a jump in the level of productivity. How do we systematically leverage AI services that simplify developers’ lives? The goal should be to remove the burden, the repetitive things, and the stuff you don’t get excited about. We want to integrate AI and GenAI across the entire SDLC, but only where it’s meaningful for the developer. The gimmicks will fade away.”
“Now that GenAI is here, model-based software development paradigms such as Mendix will be the way to engineer software,” relates Kok. “Traditional programming languages are at the wrong level of fidelity for generative AI. Low code is an ideal fit because there’s no single path from a set of user stories to a functioning full-stack application. Software development becomes a human orchestrated process with genAI in place to move as efficiently as possible. To achieve this, enterprises need strong, strong technology partners around them to ensure they can deliver next generation development experiences to match this new way of software and application engineering.”
“AI will not replace software or low code development anytime soon, but it may be possible to accomplish more with the same developer headcount,” states van Huizen. “Working with a next-generation copilot is like working with a relatively unskilled junior developer to take some of the work on board: they might be developing test cases or developing a small section of code defined by the developer.”
AI will play a big part in how evolution evolves, as will platform providers. “It will no longer be enough for solution and platform providers to simply help create better software,” Arjo van Oosten, Senior Vice President, Digital Transformation, states. “The recipe for success and closing transformation gaps will continue to move earlier. Vendors that help mobilize organizations and can bring other pieces (process, planning, and more) rather than just the platform are key. Organizations that can effectively raise their execution readiness ahead of or in conjunction with development will thrive.”
Further, organizations must continue to look at their stacks and where to innovate—and automate. From Chief Growth Officer Nick Ford: “I’m interested in how automation changes over the next year. The rise of BOAT points to a shift in how enterprises need to look at automation. Gone are the days of one-off, siloed solutions.”
Agentic AI: Not a secret
Part of the way AI-assisted development may manifest is a shift toward agentic AI. If you’re unfamiliar with agentic AI, van Huizen can help: “AI-based agents are an exciting development. That means software that can reason – that can plan and execute steps to get something done on behalf of the organization. Rather than being infused into an application, the AI is responsible for working out how to address a task or solve a problem in a more autonomous and proactive way.
It may be too early to throw a bot at everything, but as the technology develops over the next few years, it’s essential to consider what could become possible.”
de Visser relates, “I think 2025 may see a breakthrough of software agents. Those agents will be powered by a combination of platforms that support the building and editing of agents with large language models. That will allow for almost autonomous orchestration that people supervise. Developers become composers. Users become supervisors. People’s roles in software development are more important than ever.”
“Agentic AI will continue to grow and evolve,” agrees Kok. “That leads back to composability. With AI handling much of the heavy lifting, most app developers will be composers.”
Rao adds, “As agentic AI and development keep evolving rapidly, it creates difficulties with accurate costs and benefits forecasting. The organizations that can figure that out will be at an advantage.”
However, as with all things, there is some potential risk. “I’m interested to see the accessibility impact that moving to an agent model has,” states Koenigsberg. “Will AI agents use the right colors? Is there sound or movement in their output? Accessibility is a huge topic and something any platform vendor is concerned with. But it’s a space where the line between AI agent and human needs to be well-defined.”
Governance: Redefine ROI, create value, minimize risk
AI and GenAI certainly play a part in the evolution of governance, but with continued legislation and disruption, governance needs to be top of mind for nearly any enterprise.
“The definition of ROI will have to change,” continues Kok. “Traditionally, enterprises have focused on governance regarding enterprise application building, but AI adds additional layers. Many governance questions and issues are also very much applicable when assessing different AI models. Do they add any additional risk to your application builds? What about data? Is governance where it needs to be to avoid introducing hallucination and side effects? What if you’re looking to build out AI agents or use AI in a customer-facing way?”
Koenigsberg says, “At some points, AI can feel like a runaway train. You plug into some services, and suddenly, they’re being used a lot. If there’s not a lot of governance around them, you don’t know if it’s actually saving you anything.”
“Governance will continue to be a key consideration for CIOs along with data strategy, engineering, and security, especially manufacturing data and processes across IT and OT,” says Rao. “As manufacturers embrace AI and low-code technologies within their processes, there is an urgent need to start formulating data strategy for their organizations. Scaling AI and GenAI across an enterprise’s process will bring much-needed alignments and focus on data strategy, engineering, and security.”
2025: Paths to potential reward—and potential risks
There is a pull to get AI and GenAI into applications and as part of the development process as quickly as possible, but that comes with uncertainty. The status quo may seem like the safer path, but that’s also fraught with possible hazards. After all, what happens when competitors get ahead while you stand pat?
The correct option will vary depending on the enterprise but likely involves placing the right bets on use cases that match business and customer needs while exploring how AI can help enhance the process. It’s not magic, it’s just the next necessary step.
“AI will trigger a total rethink of how the software development life cycle is driven,” says Kok. “It’s about putting the right cogs into the right machine, adding the right incremental value.”