If Data is the New Oil, then Connect is the new Gold

I’ve been quiet recently. I haven’t written an article since 1st May and I’ve posted on LinkedIn much less.

I’ve been working on expanding my knowledge about how business owners can use technology and AI to radically change their business, and I’ve been working on a case study with Preston Baker.

I have put in a massive number of hours to get to this point, and spent thousands of pounds on AI tools to help me prototype, fail and build. That has always led me to the position — until this morning — of prioritising that work over writing about it.

Results after 2 months

In the past 30 days we have booked 165 valuation appointments from customers who have interacted with this product and then booked a valuation. That’s 56% of all of our valuations for the past 30 days. (These figures move week to week as we scale the product; these are correct as at the time of writing.)

In terms of listings, 40% of all of our listings have come from this product. The gap between the two numbers tells you the appointments are of a slightly lower motivation than our average.

June was our busiest month of the year in terms of valuations, as we started to scale up the product.

It is hard to imagine that it can continue to be as effective as this, but if the net position works out at a 10-20% growth in the business with no additional hires, then it will be the most effective strategy that we’ve ever deployed.

EAI — Effective Artificial Intelligence

AI up to this point can be described as very cool but not, yet, very effective in terms of real, tangible business results. We can add up the number of man hours saved, but where is the quantifiable business benefit?

By “not very effective” I mean brilliant for synthesising information — note taking, content creation, AI summaries — but where is the ROI? Could I run a business with fewer people, or the same number of people producing much more business?

Since 2023, when Alex and I first stood up and told everyone we had created an autonomous agent that could prospect for valuations, we have been trying different use cases. This one is 10x anything we’ve tried before.

Now, the major caveat I need to offer is that I am two months into testing with customers, so I can’t say that its performance will be sustained. But at every test we pass through, it continues to produce really, really good results.

The second caveat is that I am not going to share publicly how it works.

I don’t believe it’s replicable by the major competitor CRMs, because of our unique data structure. There is also the fact that Greenhouse OS has an unbelievably powerful and flexible Salesforce API that allows me to create new fields in minutes, build custom automations, and build AI agents that can talk directly to the CRM.

However, competitors have a funny habit of shouting ‘we can do that too’ once they see that you are ahead, and I want to delay that for as long as possible.

What I do want to share are some of the principles that led me to build the product.

Principle 1: You don’t need to acquire new customers

What percentage of your database do you use every month?

I think if you are above 10% you are an outlier. In our database we know over 70,000 current homeowners, and every month we add around 600 to the total.

The question is whether or not you remove the owners who have moved and no longer own the home. We remove them by tracking completions at the Land Registry, automatically.

Because you are securing new owners every day through your normal trading activities, the most relevant question is not whether you need to spend money to acquire new seller leads. It’s what you are doing to identify the movers already in your own historic database.

The cost of acquiring a new seller from marketing — £5? £10? £20? £50? The cost from your own database is the cost of communicating with them.

Principle 2: Estate and Letting Agents don’t like prospecting

I have spent years persuading agents of the value of prospecting. I was recently with Tom Panos, who extolled the virtues of agents prospecting and shared the habits of the top 1% of estate agents.

He made the argument that a self employed agent has to prospect for 3 years in order to build an ‘attraction business’, where the business just comes to you. He presents it as a hard, challenging task that a robust mindset is required for. That tells you everything you need to know. To walk into being ignored and rejected is not something that most people want to do.

In order to do it, they require encouragement, training, incentives and cajoling.

From a business owner’s point of view it makes commercial sense, because great people are really good — much better than AI — at converting opportunities into business.

However, if AI can identify the people who are in market today, because they tell us they are, then the human and AI partnership is transformed.

Our High Skill / High Tech model is about putting the right task in the right place.

The team at Preston Baker who are responsible for prospecting have told me twice, unsolicited, that this is the best thing we have ever done.

Principle 3: You must have control of your own technology

What percentage of your service is currently delivered digitally? And is that percentage going to increase in the next few years?

I don’t know what it is for the average agent, but I suspect around 30% is probably digital at the moment. My opinion is that our customers want it to be higher, and as the next generation of tech is deployed it is going to become higher and higher.

So what if we all use the same app? What if our digital experience is identical to the next agent’s? How do we differentiate?

Connect exists because a business owner looked at a problem and experimented with digital systems — and we could do that because we have complete control over our tech. AI has levelled the playing field in a way that most people haven’t understood yet. It has only happened in the past few months, but building digital systems is now quick, flexible and cheap.

My vision for the future is about empowering businesses to create their own customer facing digital experiences. Think of a viewing feedback web app, or a client portal. They will be able to tell the AI how they want it to look, and then plug it in to Greenhouse OS themselves, or we can help them do it.

A single powerful, safe place to keep your data (Salesforce, the world’s biggest CRM). Greenhouse OS to build the estate agency and lettings tools to run your business. And the final 10% of magic created by the business itself, to give it a competitive edge.

The future of Connect

My goal is to make Connect a product for Greenhouse OS customers from the end of this year. I will prioritise existing customers first and then gradually roll it out to new customers into 2027.

Next week’s article will focus on the final three principles I considered as part of building Connect: why brand matters more in an AI world, why disruption is now the normal state of our industry, and why AI doesn’t have to mean agentic AI.