AI in the Driver’s Seat

Idan Bar, Head of Matrix’s AI Center of Excellence, sums up a year of rapid technological acceleration – in which artificial intelligence shifted from reaction to initiative and served as a life-saving tool in the operational arena and a disruptive force in the civilian arena – and outlines what lies ahead in 2026.

An interview on technology in constant motion.

The Breakthrough of 2025: Agentic AI – We Are Becoming the AI’s Copilot

I always like to look at startups as an indicator, and in 2025 many startups were founded that focus on agentic AI, with a sharp increase in investment in this field. For example, according to Morgan Lewis, in Q3 2025 alone, $17.4 billion was invested in applied AI, including agents – an increase of 47% compared to 2024. According to the same source, AI overall accounted for nearly 50% of all global startup funding in 2025, up from approximately 34% in 2024.

AI is changing, shifting from reactive to proactive. Large language models are now working far more independently. Whereas in the past we sat in the driver’s seat and used AI as an assistant, today the roles are reversing: AI is in the driver’s seat, and we are becoming its copilot. We’ve been promoted – or demoted, depending how you look at it – to a supervisory role.

 

A Positive Surprise of AI: Helping Doctors Save Lives in Combat. A Negative Surprise of AI: Fake News That Makes You Doubt Everything. It’s Dystopian.

The use of AI in wartime medicine was a pleasant surprise. The IDF’s Surgeon General, Brig. Gen. Dr. Zivan Aviad Beer, presented data showing that during the Iron Swords war, the survival rate of wounded soldiers doubled compared to the Second Lebanon War. This was due, among other things, to AI enabling better training, rapid problem-solving in the field, and more. AI also drove dramatic change in the treatment of combat trauma. For example, in CBT therapy, GenAI now makes it possible to easily create personalized scenarios, enabling true medical personalization. The military also deployed a chatbot that trains soldiers to identify signs of distress in their peers, raising awareness and vigilance of soldiers in cases of mental distress.

On the other hand, I was deeply disturbed by the scale and intensity of GenAI-generated fake news. It’s simply unbelievable. Much of what we see on Telegram today is AI-generated misinformation. It’s completely dystopian, and in my view, one of the most damaging developments we’ve experienced as a society.

 

Looking to 2026: Agentic AI Will Gain Momentum with a Compound Annual Growth Rate (CAGR) of Approximately 42.8%! And Ethical AI Is No Longer a Privilege.

In the coming year, we’ll see agentic AI continue to gain momentum, largely driven by the Model Context Protocol (MCP) and similar protocols. At the same time, there is a significant gap between organizations that are ready to adopt it and those still afraid to ‘release the autopilot’ into their environments.

While the startup ecosystem is flourishing, most organizations are still struggling to implement agentic AI in a practical way. Two major barriers are slowing adoption: suitable technological infrastructure and organizational readiness. Some organizations aspire to adopt advanced technologies, but in practice their data is scattered across employee laptops, GPU (Graphics Processing Unit) infrastructure is lacking, and systems aren’t built for autonomous workloads and processes.

On the other hand, there are organizations that are technologically ready but lack the necessary organizational maturity – processes, culture, risk management, and internal alignment – for change. True adoption requires intent, technical capability, and financial resources.

Given the current pace of technological growth, time is running out for organizations. For example, a report by Fortune Business Insights forecasts that the agentic AI market will grow from around $7.3 billion in 2025 to $88.3 billion by 2032, with a CAGR of 42.8%. The upshot of this is that organizations that continue to hesitate will lose the competitive advantage, harming efficiency, and may even becoming irrelevant. As a result, we are likely to see significant acceleration in the adoption of Agentic AI in 2026.

Another issue that can no longer be avoided is privacy and ethical AI. Regulations will become stricter in the coming year, and organizations will need to adapt to a new reality in which responsible AI is not just a ‘nice-to-have’, but a basic requirement for operating in a competitive market.

 

About My Role: AI Is Like a Puzzle in Constant Motion

I lead Matrix’s AI Center of Excellence and work across every aspect of AI project development. In a company like Matrix, this can mean huge projects as well as small projects – the range is wide. It requires me to constantly stay up to date, track trends, read key blogs, and be aware of all available solutions in the market.

In a role like this, it is imperative to be very up to date, because ultimately, we are supposed to provide solutions to problems, and in order to provide the optimal solution to each problem, you need to know all the products, all the platforms, and all the options that exist. Only then can you put together the exact solution to the problem, which will suit the needs and resources of the organization. The role also involves managing a group of people who implement the solution or architecture that I, as an architect, create, and of course participating in many pre-sales meetings.

Beyond technical leadership, I also provide guidance on AI across the company. The focus here is less on building the solutions themselves, and more on creating connectivity and synergy. I identify capabilities that already exist within Matrix, including those not yet brought to market, and connect the right people to leverage them. I serve as a senior point of reference for AI topics across the organization; they can always consult with me, and I help business units and subsidiaries advance independently in the AI domain.

My days are diverse and packed: I prepare podcasts to listen to on my commute, and dedicate significant time to internal and external meetings with clients, architecture responsibility, and helping with solution design. I also deliver professional lectures and talks both within and outside of the company.

What I love most about my work is the variety. AI is like a puzzle in constant motion – you’re always solving problems, and the problems change every hour. The composition of each solution – the domain, technology, customer, data – is never the same. And because the technology evolves in giant leaps, we are constantly discovering new pieces of the puzzle.