121 Units. Half the Team. One AI System

Abdel Khawatmi breaks down how he embedded AI into every layer of his mortgage business — from income calculations and LOEs to client communication and event ideation — achieving 210% net revenue growth YOY while cutting his virtual assistant team from seven to three. He also covers his hyper-local in-person event strategy and how he delegates without losing the client experience. Best for LOs who want a concrete, operations-level look at AI adoption, not theory.

Takeaways you can run this week

• Write down every task you do in a day, categorize each one as Sales, Service, or Support, then identify the first three that are repetitive and rules-based — those are your first AI automation targets.
• Train ChatGPT (or your AI tool of choice) with a detailed SOP for a recurring task (e.g., calculating Schedule E rental income or drafting a letter of explanation) and test it until it outperforms human consistency — then assign that prompt/SOP to a VA or processor.
• Use AI to generate unconventional event ideas for referral partners: Abdel prompted ChatGPT and got Botox events, mani-pedi days, and a daddy-daughter salon rental. Run one of these instead of a standard happy hour this quarter.
• Set a hard calendar block at 7–8 p.m. daily labeled 'Missed Calls and Emails.' Have your admin text you a list of missed contacts; call every one of them that night — even if you already connected via email.
• Have your team or admin make the initial outbound call to a new client or partner, then transfer to you live and stay on the line to take notes — or let the call record and auto-transcribe into your CRM so you never manually update a contact record again.
• Before meeting any new referral partner or prospect, have your loan officers research their social media and tag you in — then show up knowing one specific thing about them. Abdel's team found a financial advisor lead this way through a simple social post asking for recommendations.

The playbook

1. Step 1 — Map your business: Write down your daily calendar tasks and sort them into Sales, Service, Support.
2. Step 2 — Identify AI targets: Find the tasks in Service and Support that are rule-based and repetitive (income calcs, LOEs, CRM updates, call transcription).
3. Step 3 — Build the SOP: Record your screen with Scribe or a similar tool doing the task exactly how you want it done. Feed that SOP to your AI tool.
4. Step 4 — Test for consistency: Give AI the same task a week later. If it outperforms human error rate, it graduates from your desk.
5. Step 5 — Assign and layer: Hand the trained prompt/SOP to a VA or processor. Use the freed time to make personal outbound calls to clients and partners.
6. Step 6 — Be the guinea pig first: As the leader, demo every new tool in your own production before rolling it to your team. Share your results (revenue lift, time saved) to earn buy-in.
7. Step 7 — Add the human touchpoints back: Use automation to surface the right moments, then pick up the phone — show clients a Mortgage Coach total cost analysis, MBS Highway rate report, or housing market data in real time on that call.

Worth quoting

“I have never more been on the phone over the last year, since we really implemented artificial intelligence, than ever before.”

“The ROI — it's not return on investment. It's: are you being relevant, are you present, and can they relate to you?”

Best for

LOs who know AI exists but are still dabbling — and branch managers who want a repeatable system for rolling AI adoption across their entire team.

Links from the show notes

• https://www.linkedin.com/in/abdel-khawatmi-79a511144/
• https://www.instagram.com/got_mortgages/
• https://www.linkedin.com/in/katieshive/