Imagine a negotiator who works 24 hours a day, 7 days a week, 365 days a year. They always follow their training and can perform hundreds of task simultaneously. That’s what the agentic revolution is going to mean for businesses across the UK.

So what is the agentic revolution? A good place to start is to ask AI itself, so here is what it told me

The Agentic Revolution refers to a transformative period or concept in which artificial agents, such as advanced artificial intelligence (AI) systems, increasingly take on roles traditionally performed by humans, especially in decision-making, problem-solving, and creative tasks. The term “agentic” relates to the capacity of these AI systems to act autonomously, making decisions and executing tasks without constant human intervention.

In our language this is about autonomous agents who can book viewings, book valuations, call out new listings, chase feedback, give clients updates on progression, analyse the whole of the market to calculate the exact sweet spot for a price reduction and do it all through either text (e-mail / sms / whatsapp) or voice.

The technology we are piloting / building is exactly in this space, voice agents who have an objective, guard rails to stop them doing things we don’t want and then are set loose.

The death of bolt on Prop Tech

You might have tried AI chat bots and thought “yes, that’s interesting” but it’s not ready yet. This is because the current generation can tell me generic information about an area, like I was on ChatGPT. It’s brilliant but completely lacking context. It needs data, and by data, I mean your data.

Imagine that every home in one postcode (LS8) or one postcode sector (LS8 1) that was listed for sale was fed to the AI. So it knows how long every home was on the market, whether it reduced in price and following the price reduction it sold. It had every data point on every home from it’s council tax band, to it’s square footage. Now it analyses your stock which hasn’t sold, and calculates the optimal level of price reduction for each of your properties. It knows that facts tell and stories sell, so it creates specific case studies of properties that have reduced and the subsequent impact in terms of sale.

It analyses all of your price reductions and sees what supporting activities would be required to achieve a sale. It sees that when 50 phone calls are made, to support the e-mail campaign, we can generate an additional 3 viewings which gives an increased likelihood of sale. It also sees that a rightmove premium listing helps in terms of getting a seller to agree to the reduction.

Now it prepares a written report, complete with case studies and sends it to the customers, offering a call to discuss it.

It can even make the phone call to the customer, using a cloned voice to have that conversation with the seller. I hope this example demonstrates why the data, whether it’s external (the on market data from Twenty EA, the property data from Twenty EA) and your data is a powerful combination which means that everything must relate back to one central source of truth where everything happens. If you don’t have access and flexibility for this data, if it isn’t organised and structured correctly then you cannot leverage this technology.

We launched Greenhouse OS to make this possible for us and to bring others with us on the journey which is about to start accelerating fast.

Yesterday, I cloned two of my teams voices so that we can have our voice agent with a truly authentic Yorkshire accent. We are very much in the first few steps of building out these AI agents, but this technology is coming and it’s going to change the way service is delivered, leads are handled and marketing is done. Until very recently, I believed it was going to take 2-3 years for this to happen in a meaningful way which for most companies is going to be true. However this year I believe that the early adopters will have deployed AI agents and it’s going to start posing huge questions to them and to their competition.

No new technology or data is now required, it’s simply about the time it takes to make the prototypes, test them and deploy them.

A generational challenge to leaders

I came into estate agency in 1998 and the rise of rightmove and the fall of the newspapers is realistically the biggest change I can see that has occured. The amount of foot fall has lowered, gradually and much more time is now spent dealing with leads, but fundamentally the job hasn’t changed.

It’s not that I think the fundamentals of the job will change with the agentic revolution, it’s more that the choices that leaders will now have are going to be very different. The skills in an organisation will have to become different. We will have to be thinking much more about designing our process, testing and deploying AI agents and measuring their success. The beauty of this will be that once it has been done, the AI will keep following the process, unlike it’s human competition.

My plan is to approach this problem from two ends.

One end is MVP (Minimum Viable Product) deployments of the technology to learn how to do this in a controlled way with narrow scope.

The second is to imagine the most extreme deployment of this technology then gradually peel back the layers until I get to something which I can consider to be an appropriate target for my business.

You might think that the most extreme deployment of this technology will NEVER be for your business, that the personal touch is what makes you special and therefore AI can’t replace what you and your team do. You may be right, but there are others who will adopt the most extreme version, probably new incumbents into the market (see The Innovators dilemma as to why this is likely to be the case).

The most extreme version is basically everything done by AI with very few exceptions, probably the ‘in home’ elements. So that would leave viewings, valuations and key handover as the only elements done by an agent. Is there a scenario where it’s literally the valuation and everything else is done by the seller themselves?

If the AI described in my first section can do the job of VMU’s better than a human then why would this end up being an inferior service? What’s the price point that it could operate at?

The last generation of low cost agencies failed because their service was poor and their cost point compared to their revenue point didn’t add up. What is their cost point was a fraction of it’s previous level and their service was really, really good (but not personal).

Independent agencies have a huge advantage however. Brand and data. If we grab the nettle now and implement these technologies ourselves, combine it with our expertise, brand and data we can deploy these technologies which won’t just make us relevant now but for the next 20 years as well.

What is clear is that we all need a plan.