Artificial intelligence is taking another major step forward by moving beyond the familiar chatbot format to systems that carry out a range of tasks on our behalf.

This new wave, known as ‘agentic AI’, is designed to learn from feedback and operate with less human direction, doing far more for users than simply answering questions.

One of the most prominent examples is OpenClaw, an open‑source platform that allows users to build their own AI agents. These agents will take action for you: they might book your next holiday, do your shopping, or monitor live information overnight and report back to you the next morning.

This ability to automate real work has captured the attention of businesses – and investors.

From mid to late January 2026, the agentic evolution caused the market to “change how it looked at AI”, Milford Senior Analyst Andrew Curtayne said in episode 68 of Bridge talks Business.

“All of a sudden, people found they were able to use AI to automate work processes and really start saving time. This led to a massive repricing of the software sector,” Curtayne told Ryan Bridge.

Investors began questioning the need for expensive enterprise tools and the use of costly white‑collar workers for tasks that agentic AI can increasingly handle itself.

“So you saw various software companies down by up to 30%, 40%, or 50%. It also spread into non-software companies such as financial institutions – insurance brokers and legal companies started selling off.”

But Curtayne said the market reaction has not been uniform and does not mean every company stands to lose. Far from it.

Recalling discussions at technology summits he attended in the United States recently, Curtayne said Nvidia Chief Executive Jensen Huang was clear that the agentic evolution will significantly increase demand for AI – particularly as these systems operate continuously rather than requiring constant human direction.

“Token demand, which is the currency that large language models work on, is going to rocket exponentially. We’re already seeing this in the data – token demand has gone up about four times in the last eight weeks alone,” Curtayne said.

“Theoretically, this is good for the semiconductor or the chip companies. It should be good for the likes of Nvidia or TSMC.

“Software companies are going to have to quickly adapt to this new world. They’re going to have to deploy their own AI agents, which they’re all trying to do. They’re going to have to try and capture a share of this revenue from these AI agents and make sure that they can tie it into the products they offer, and can generate what’s called consumption revenue, which is revenue from people using AI.”

Despite how consequential the agentic AI evolution is, Curtayne said the ongoing war in the Middle East remains the biggest influence on Milford’s portfolio positioning. The geopolitical uncertainty and oil market shock mean “technology has taken a little bit of a backseat”.

Within the tech sector, Curtayne said Milford remains selective: positive on chipmakers benefiting from rising AI usage, more cautious on software stocks where earnings certainty has declined.

“The exciting thing is that there are opportunities out there,” Curtayne said. “We’re seeing a lot of companies selling off too much for perceived risk, so we’re finding some interesting investment opportunities.”

Catch the full conversation between Ryan Bridge and Andrew Curtayne in Bridge talks Business episode 68.