For twenty years the CRM has sat at the centre of every agency’s tech stack. But when any agent can build software in an afternoon and AI can hold the conversation for you, what exactly is a CRM for? Avi Khan argues the category must reinvent itself – or disappear into the background.
There was a time when your CRM was your competitive advantage. If your agency had a serious database, automated workflows and a slick way to launch listings to portals, you were ahead of the office down the road still running on spreadsheets and shoeboxes of contact cards. The software itself was the moat.
It was expensive, it was complicated, and not everyone had it. That world is gone.
Today, an agent with no technical background can describe a piece of software in plain English and have a working version of it before their coffee goes cold.
AI has collapsed the cost of building technology to almost nothing. The tools that once required a vendor, a contract and a six month implementation can now be spun up by a motivated agent on a Sunday afternoon.
Adoption has followed: recent industry research found 82 per cent of agents are already using AI, and the number of brokerages with no plans to adopt it has fallen to under 2 per cent.
The technology gap between agencies is closing at extraordinary speed, and that should terrify every CRM company in the country. The moat has been drained.
Here’s the uncomfortable question for the category: if the software is no longer scarce, what is the CRM actually selling? For most platforms, the honest answer is two things. A place to store contacts, and a place to launch listings from. A data warehouse and a distribution pipe.
Neither is defensible anymore. Storage is a commodity. Portal feeds are a commodity. Email and SMS automation, the “drip campaign” that CRMs spent a decade selling as innovation, can now be replicated by a general purpose AI tool that also writes better copy.
When the core functions of a product can be rebuilt by its own customers, that product doesn’t have a moat. It has a subscription renewal date. The CRMs that survive the next five years won’t be the ones with the most features. Features are free now. They’ll be the ones that own something that can’t be copied.
Is propensity the new moat?
So what can’t be copied? The answer, I’d argue, is knowing who is going to sell before anyone else does. Every principal knows the maths of our industry: in any given year, only a small fraction of the homeowners in your database will actually transact.
The agent who wins is not the one who contacts the most people, it’s the one who contacts the right people at the right moment. That is a propensity problem, and it’s a problem that can only be solved with data at a scale no individual agent can assemble: transaction histories, tenure patterns, listing signals, behavioural data, life event indicators.
This is where the CRM has a legitimate future, not as a filing cabinet, but as a prediction engine. A platform that can tell an agent “these eleven households in your farm area have a high likelihood of listing in the next six months, and here’s why” is worth ten times one that can merely store those households alphabetically.
The value shifts from holding the data to interpreting it. From record keeping to foresight. The uncomfortable corollary: a CRM that holds your data but can’t score it isn’t an asset. It’s a liability with a login screen.
Which brings me to the most heretical question of all: does the CRM of the future need to be agent facing at all? We’ve spent two decades training agents to feed the machine, log the call, update the note, set the task, move the deal along the pipeline. Entire compliance regimes inside agencies exist purely to make agents do data entry. And agents, quite rationally, hate it.
The industry’s dirty secret is that most CRMs are graveyards of half entered data, because the interface asks salespeople to behave like database administrators. AI ends that bargain.
When every call can be transcribed, summarised and logged automatically; when the follow up drafts itself; when the system already knows what happened at the open home, the agent stops being the data entry layer.
The CRM stops being a place you go and becomes a thing that happens around you. And when an agent does need something, they won’t click through seven menus to build a filtered list. They’ll simply ask.
“Who should I call this morning?”
“Which of my past buyers are now likely sellers?”
“Draft the vendor report for Saturday’s auction.”
The conversation becomes the interface. The dashboard, that proud wall of widgets every provider demos, starts to look like a fax machine: a technology that exists because the better version hadn’t arrived yet.
The end of the age, or the start of a better one?
So is the age of the CRM coming to an end? The age of the CRM as we know it, the database with a dashboard, the listing launchpad, the system agents serve rather than the other way around – yes. That era is finishing, and quickly.
What replaces it is more interesting: an invisible, predictive layer that watches the market, scores the opportunity, does the administration, and speaks to the agent in plain language. Less software, more colleague.