For years my routine after a face-to-face meeting was the same. Sit there taking handwritten notes, get back to my desk, and realise half the gold was already gone. The bits I hadn’t managed to write down while I was actually listening.
Then I started using a small pocket recorder. Suddenly I could capture an in-person meeting in full, the same way my online calls were already captured. That alone felt like a step change. But it was the next step that was the real value: the insight that I could now feed that recording to AI and have it pull out the gold for me. Not just a transcript, but now the ideas and angles I often miss when I am deep in a conversation. The insights. The action list! It could even now prepare for me (within in my email system) a fully drafted follow-up email and put in my drafts folder ready for me to review and send as soon as I was back to my computer.
That’s the whole point of this article. Not because AI is magic, but because of where the time and the value were actually hiding, and how little of it anyone has been shown.
The state of play
Here’s the reality of AI use today in our industry, and I’ve spent a while getting a read on it. A few AI sessions I’ve run, a lot of conversations along the way, and a survey across the market. Most brokers using AI are still working one prompt at a time. Some have a feel for memory, or have set up a project so the tool holds a bit of context. But they’re still in the middle of every step: the one who opens the window, kicks the job off, checks what comes back, then feeds in the next instruction. Nothing moves unless they move it. What almost nobody has built is a workflow, where the AI carries a job from start to finish and only comes back when it hits something that needs a real decision from you.
It works. A listing description here, a quick email draft there. But it’s a sliver of what’s sitting there, and almost everyone assumes the rest of the market is further ahead than they are. They’re not. In our most recent survey, around 70% of brokers were on ChatGPT (often now moving to Claude – but still in the chat function) and the majority hadn’t tried building a single workflow.
So if you’ve felt behind on this, you can let that go. The brokers getting real value aren’t using a secret tool. They’ve taken one step the rest haven’t.
Where the time is actually hiding for brokers
When brokers tell us where they want AI to help, the same areas come up. These are the ones worth your attention, and what to actually do with each:
- Seller and buyer meetings. Record the meeting and let AI do the heavy lifting afterwards. Capture why they want to sell or buy, the specifics that come up along the way, and save it all to memory so nothing is lost and your AI can work out the best approach to take with them next. From one recording it builds your post-meeting action list, drafts the follow-up emails, and flags any objections that surfaced, with strategies for handling them and a suggestion for what to send next.
- Buyer qualification. Stop spending hours on buyers who were never going to proceed. Have AI qualify them before you invest any time: do research on the buyer, their background and likely capacity, and how well they fit the seller you’re working with, so you walk into the conversation already knowing a real buyer.
- Research, the seller, the industry, the buyers. Build a repeatable research process and point AI at it. Seller and industry background, research on the buyers who have already approached you, and a scan of other potential strategic buyers to spot opportunities you’d otherwise miss. Turn it into a process you run on every deal at the press of a button rather than a from-scratch job each time.
- Client and buyer communication. Set up an agent that runs each morning before you start. It connects and summarises yesterday’s activity across each deal’s email threads, brings anything urgent to the top, orders the rest by priority, drafts replies ready for you to review and send, and strips out the noise and clutter. A 24/7 personal assistant that cuts your time in the inbox right down.
- Information memorandums. Not every broker drafts their own IMs, but if you do, this is where AI earns its keep. It drafts the narrative, brings sector and industry insight, and adds elements from its own deep research that the seller may not have thought to give you. It can even find angles for the IM by benchmarking against its own research and the tools it’s connected to. (Industry insight is a bigger topic on its own, and one we’ll come back to in a future piece.)
And this is just the surface of each one. In a future article we’ll show how these turn into deeper workflows, where multiple agents work together to get the job done: agents that know how to do the research that matters, draft the emails and documents you need in your voice and in the format you want, and learn each time you correct them, so you end up with an almost-perfect output the first time.
The setup that makes it work
Here’s the actual difference between brokers getting value and brokers spinning their wheels. It isn’t the tool. It’s whether you’re using AI as a chat box or as a system.
A chat box forgets you every time. A system remembers. It knows your business, your deals, and how you work, because you set it up once with that context and gave it a memory that carries across sessions. Connecting it to your inbox and files takes that further again, though not every broker can do that, and we’ll come to why in a moment. Even without it, the same question that gives a generic answer in a cold chat window gives a sharp, specific one when the AI already knows your world.
That’s the whole game. Not a better product. A proper setup.
The step that makes the biggest difference
If I had to name the single step that lifts AI from useful to genuinely powerful, it’s connecting it to your real files and inbox, so it works from inside your business rather than starting cold every time. Most brokers never do it. It takes an afternoon and it changes everything from that point on.
One honest caveat: some brokerages don’t allow you to connect work email to outside tools, and if that’s you, this step may be off the table. Don’t let that put you off the rest. The project setup and the memory layer still deliver most of the value on their own, and there’s plenty you can do with documents and information you bring in manually. The connection is the accelerator, not the whole engine.
The question worth asking yourself
So the question isn’t “which AI tool should I use.” Almost everyone asking that is still standing at the chat box. The better question is: have I actually set this up, or am I starting from scratch every single time?
The brokers ahead of you right now didn’t find a better product. They spent an afternoon building a proper setup and moved past the chat window. That’s the gap. It’s a setup gap, not a tool gap, and it’s a far easier one to close than it looks.
Where to go from here
If you want to go past the article and get this built properly, that’s exactly what our hands-on AI workshops are for. You leave with the setup actually installed and running, not a set of notes to action later. The sessions are for our current referring partners, brokers who have referred work to us in the last 6 months, and getting on the list is more straightforward than most brokers expect. If that’s of interest, reply or get in touch at [email protected] and we’ll point you in the right direction.
