It always seems sudden
The call comes on a Tuesday. The buyer's solicitor says their client is withdrawing. When you look back through the file, the signs were there: the buyer took a fortnight to instruct a solicitor, the survey was booked late, the buyer stopped answering the progressor's texts, and a sale further down the chain had been quiet for a month.
Nobody put those facts together because they were spread across notes, emails and different people's heads. Each one on its own looked like normal slowness.
Why the warning signs get missed
Agreed sales are usually managed as a list in date order, and each file gets attention when someone rings about it. There is no view that says this sale looks different from the ones that usually complete. Your team has instincts about which buyers are wobbling, but instincts do not get shared at the Monday meeting unless someone raises them.
The information that matters is partly structured, like milestone dates, and partly buried in text, like a note saying buyer seemed unsure after the survey.
| Signal | Where it hides |
|---|---|
| Slow solicitor instruction | Milestone dates, if recorded |
| Survey or valuation not booked | Progression notes |
| Buyer going quiet | Unanswered messages, call logs |
| Down valuation or survey issues | Emails and phone notes |
| Chain link stalled | Another agent's updates, often verbal |
What fall-throughs cost the branch
Your fee arrives on completion, so a sale that collapses late means months of work unpaid and a property back on the market with less momentum. The vendor's confidence in you drops, and they may take the relaunch elsewhere. Forecasts built on the pipeline become unreliable, which makes planning staff and marketing harder.
Some fall-throughs cannot be prevented. The ones that hurt most are those where an earlier conversation, a second viewing lined up with a backup buyer, or a push on a slow solicitor might have changed the outcome.
How we build the risk view
- Pipeline data: we read your agreed sales from the CRM, with milestone dates, days since the last contact from each party and chain information.
- Notes reading: progression notes and emails linked to each sale are read by a language model for signs your team would recognise, such as a survey concern or a buyer mentioning doubts. Each signal is quoted, so you can check it.
- Simple rules first: most flags come from clear rules your team agrees, such as no survey booked within a set period after the memo. Where you have enough past sales, we can test whether a scoring model adds anything, and we tell you honestly if it does not.
- Ranked list: each morning, the progressor and branch manager see the sales most in need of attention, with the reasons listed next to each one.
- Actions: from each flagged sale, the team can log a call, schedule a vendor conversation or mark backup buyers to keep warm.
The view advises where to look. It never contacts buyers or vendors on its own, and it never changes a sale status.
A different Monday meeting
Instead of reading the pipeline top to bottom, the meeting starts with the handful of sales that look unusual, and everyone can see why. Instincts get written down and shared. When a sale does fall through, you are more likely to have seen it coming and to have kept interested buyers warm.
Over time the branch also learns which stages cause most of its fall-throughs, which is useful when deciding where to put progression effort.
Is this your pipeline?
- Fall-throughs regularly come as a surprise.
- Your pipeline review goes through every sale in date order.
- Warning signs sit in notes that nobody reviews together.
- Backup buyers are not kept warm on shaky sales.
- Your exchange forecast is often wrong by the end of the month.