The agents are coming (whether we like it or not)

“Computer says no”.

These words will bring a pained smile to many Brits over the age of 30. The soundbite was coined in a mid ‘noughties’ comedy sketch from Matt Lucas and David Walliams. Walliams, the face of ‘customer service’ at an amorphous large organisation, sat expressionless, face to face with a customer played by Lucas, entering his questions into a computer. No matter, the question, the answer from Walliams always came back the same: “Computer says no”.

It hurt us to watch because computers back then were running simple yes/no algorithmic programmes for many day-to-day customer service tasks. Either you could have a loan or, as in Lucas’s case, you couldn’t. And because there were supposedly computers running software like this up and down the UK making decisions – where risk aversion was baked into the programme – the stigma built. We became pretty good at saying ‘no’ to things.

A quick glance at a few charts tells me the UK is not particularly productive. And a few other charts suggest that our public sector isn’t exactly hitting efficiency targets either. Only this week I’ve watched, serving as a Juror, as the UK Crown Court system struggles to recover from a very long backlog with technology that slows down, rather than speeds up, the court process.

I spend quite a bit of time in Brussels too, and I don’t think I’ll offend anyone there if I say that system efficiency isn’t something we see much of. Conceiving of, drafting and analysing legislation is still a very manual process.

Breakthrough

And so I was delighted to hear global tech leaders herald the arrival of AI ‘agents’, a new force in AI capability. Agentic AI are end-to-end process tools with autonomous, decision-making capability. My words, no one else’s*. AI Agents are similar, but with less autonomy across more specific tasks. Where Chat GPT can answer almost any question you ask, an agent will know when to ask the question, what to ask and how to use the answer to perform a pre-determined action. Agentic AI might decide not to ask the question at all and find a more efficient solution elsewhere.

Here's what Salesforce’s Mark Benioff had to say only last month: “Agentic AI is a new labor model, new productivity model, and a new economic model," he said. “I've never been more excited about software, more energized or motivated, inspired. And I think that we’re about to move into a completely new world.”

“I've never been more excited about software, more energized or motivated, inspired." Salesforce's Mark Benioff.

Unveiling his company’s future vision at CES only a few weeks ago, Jenson Huang, CEO at Nvidia, said 2025 will be the year of the Agent. “AI Agents are likely to be a multitrillion-dollar opportunity.”

This comes at the perfect time for the UK Government, suffering from a bleak picture on economic growth. Earlier this week Starmer unveiled the AI Opportunity Action Plan, with substantial investment into a structure that could put the country ahead of the race to benefit from this technological revolution.

Agents embedded into our workforces and processes, could take on mundane tasks that we are naturally reluctant to do. It can perform these tasks at a scale that no human can do and do so with greater accuracy and efficiency.

Agents in communications means we can listen to every issue in every market simultaneously and report on it with an automated daily email.

Agents in customer service means every customer being treated personally, with a personalised and reasoned interaction. No more 'yes' or 'no's.

Agents in the court system could mean auto transcription, automated jury communications, automated court booking systems. Less time wasted.

Agents in the NHS. Well is could be transformative. And note I talk of agents here. Not even Agentic AI.

The Computer will no longer say ‘no’. It’s going to say, ‘let me look into that for you, Mr Wilson’. In fact the risk of the ‘no’ comes from our workforce. The same employee that saw one GPT hallucination and now says no to any AI-led system change.

I’ll leave the parting shot here to ex British politician and newly appointed Oxford University Chancellor William Hague who has teamed up with Tony Blair to make a case for investment in AI.

“Any Government that clambers into the good ship AI Growth will find that it has to pay some expensive fares, throw some other items overboard and get used to working with some new crew – all difficult unless you remember the alternative is drowning.”

*This time, all words are my own except where indicated.