How to Stop Losing Past Client Deals Before It’s Too Late
Andrew Penner breaks down why 82% of past clients use a different lender on their next deal and how his platform, myMilo.ai, uses behavioral intent signals to alert LOs 6-8 weeks before a past client pulls credit elsewhere. This episode is for LOs who know their database is their biggest asset but have no active system to monitor or re-engage it. The conversation also covers how to balance AI-driven automation with authentic human relationship-building.
Your past client database is leaking future deals right now — not because of bad service, but because you have no early-warning system to know when clients are back in market before they call Rocket.
Takeaways you can run this week
- Audit and organize your past client database this week — even a clean Google Sheet with names, contact info, and close dates beats flying blind. You cannot manage what you cannot measure, and no platform can help you without structured data.
- Set up credit monitoring on your closed loan portfolio if your company doesn't provide it. Andrew recommends Sightline (Sticum) as an individual LO option to receive alerts before a past client's credit is pulled by a competitor.
- Upload your database to myMilo.ai (mymilo.ai) to immediately start engaging ~70% of past clients monthly via white-labeled home value reports and equity tools — expect 1.5%–3.5% to surface as high-intent prospects each month.
- Use the 'Rule of 25' as a planning benchmark: per ICE research, 25% of homeowners plan to refi or tap equity in the next year, but lenders retain less than 25% of those. On a 1,000-loan database, that's roughly 250 opportunities — calculate your own number and treat it as a pipeline target.
- When Milo or any platform flags a past client as high-intent (e.g., saving homes on search sites, visiting mortgage comparison pages), call that client personally within 24 hours — do not hand that conversation to an AI autoresponder. Andrew's rule: 'Our past client database is sacred.'
- Start building structured data on your clients and your own communications now (recorded calls, logged emails, detailed CRM notes) so that future AI tools can be trained on your actual relationship style — LOs who do this early will outperform those who haven't when AI-assisted outreach matures in 1-2 years.
Useful? Get the full Vault free — plus Marketing Worth Stealing, weekly.
GET FREE ACCESSThe playbook
- Step 1: Organize your database — CRM, spreadsheet, or any structured format with complete contact and loan data.
- Step 2: Enable credit monitoring on that database (company system or a tool like Sightline/Sticum).
- Step 3: Load the database into myMilo.ai; the platform begins sending white-labeled home value reports and tracking on-platform behavior immediately.
- Step 4: Monitor the intent alerts Milo sends to your CRM — these flag clients showing high-intent signals (comparison shopping, home saves, equity calculator use) typically 6-8 weeks before a credit pull.
- Step 5: When an alert fires, make a personal phone call or send a personalized message — do not delegate first contact to automation.
- Step 6: Structure and log all client interactions (calls, emails, notes) in your CRM now to build training data for future AI personalization.
Worth quoting
“82% of customers use a different mortgage lender on their next transaction — there's clearly a systemic problem that wasn't being addressed.”
“Our past client database is sacred — we've been really careful to walk that line cautiously in Milo.”
Best for
LOs who have 3+ years of closed loans but no active system monitoring when those clients are back in market.