All Episodes
EP 78Special Episode•Oct 16, 2026

5 Recurring Trends on Brain MB: How ABW Brokers Are Using AI to Get Deals Unstuck

Dean Lawton and Jason Marshall

When the Data Speaks, Listen

51 out of 145. In just three months, more than a third of ABW brokers using Brain MB kept coming back to ask about one thing — and it wasn't rates. It was rental income.

That single data point tells you something important: files don't get stuck on the obvious stuff. They get stuck in the details — the policy nuances, the lender-specific limits, the fine print that separates a closed deal from a wasted week.

In this episode of The Mortgage Broker Podcast, Dean Lawton and Jason Marshall revisit Brain MB — ABW's proprietary AI-powered lender policy and rate intelligence tool — and dig into the 888 questions submitted by 145 ABW brokers between July 2nd and September 30th. What emerged were five clear, recurring themes that reveal exactly where brokers are getting stuck — and how the right tool can get them unstuck, faster.

What Is Brain MB?

Brain MB is ABW's AI-powered mortgage intelligence platform, built on a knowledge base of lender policies, rate data, and brokerage-specific resources. Think of it as the knowledgeable senior broker down the hall — available 24/7, never in a meeting, and never too busy to answer your question.

For brokers, the analogy is familiar: before tools like this existed, finding an answer meant hunting through broker kits, texting a BDM and waiting, or posting in a Facebook group and hoping someone with the right experience would respond. Brain MB compresses that entire process into a single, cited, real-time response.

"It's like three steps and a wait versus one question and one response." — Dean Lawton

The platform currently references 72 lenders with up-to-date policy data, a live rate explorer, and — for ABW brokers — a brokerage-specific layer that includes regional BDM contacts, submission requirements, and preferred lender information.

The 5 Most Common Trends on Brain MB

1. Rental Income — 51 Brokers

By far the most searched topic, rental income questions spanned three core areas:

  • How much rental income can be used (offsets vs. add-backs)?
  • What percentage does each lender apply?
  • How many rental properties will a lender finance before saying no?

A real example from the data: "Is there a maximum number of rental properties TD Bank will finance for one borrower?"

The response: Yes — TD limits borrowers to a maximum of five residential investor properties per application. But the value didn't stop there. The response also clarified that this limit includes properties financed by other institutions and properties owned free and clear — a detail that trips up even experienced brokers. It also noted that properties with more than four units fall under TD's Multi-Unit Residential program, redirecting the broker to the right channel entirely.

This is the power of context-rich responses: not just the answer, but the nuance behind it.

2. Non-Standard Properties — 45 Brokers

Mobile homes, manufactured homes, leasehold properties, micro condos, and acreage — these non-standard property types are where files quietly die. The primary question brokers were asking: who will even look at this?

Example: "Will MCAP finance a mobile home on lease land?"

The response was immediate and unambiguous: Double decline at MCAP. MCAP does not finance lease land under any circumstances, and does not lend on manufactured or mobile homes at all. But rather than leaving the broker stranded, Brain MB surfaced alternatives — Coastal Community Credit Union for BC properties, and Deja Nens for Ontario or Quebec situations — each with the specific terms attached.

"A clean no is worth a lot. It stops you from sending a file that was never going to get done." — Jason Marshall

The episode also highlighted an important behavior: when Brain MB detected a conflict in First National Excalibur's own records — citing minimum condo square footage as both 325 sq ft and 500 sq ft — it flagged the discrepancy and told the broker to confirm directly. That transparency, Dean emphasized, is exactly the behavior you want from an AI tool. It doesn't hide uncertainty; it surfaces it.

3. Rate Shopping — 37 Brokers

Rate sheets go stale. Portals lag. And when rates are moving fast, a week-old email from a lender can cost you a deal. This was the frustration behind 37 brokers' rate-related queries on Brain MB.

Brain MB's Rate Explorer allows brokers to filter by province, property type, term, insured/uninsured status, and more — and the rates are verified every morning, with a timestamp on every response. Where APIs are available, some lender rates update multiple times per day.

Example: "What's the lowest 5-year fixed insured rate right now in BC?"

Response (as of October 5th): 4.29% with Coastal Community Credit Union in BC, minimum Beacon score 600, with a note that several products were second-home specific — prompting the broker to confirm purchase type before quoting.

The takeaway isn't just the rate — it's the confidence to act. You know what you're working with, you know when it was verified, and you can make your call to the lender with a clear starting point.

An added layer for ABW brokers: active whisper promos are surfaced directly in both the chat and the Rate Explorer — critical when five overlapping promos with different start and end dates are in play simultaneously.

4. Self-Employed Income — 39 Brokers

Self-employed files are among the most nuanced in the broker's toolkit, and the questions in this category reflected that complexity: stated income programs, bank statement requirements, dividends, corporate income, and partial ownership structures.

One of the most instructive examples involved Scotia Bank. The broker asked: "Does Scotia Bank have a stated income program for self-employed borrowers and what are the requirements?"

The answer? Scotia Bank does not offer a traditional stated income program — but Brain MB didn't stop there. It identified Scotia's Business for Self Plus (BFS+) program, a confirmed income program using corporate net income after tax (NIT), and outlined the key parameters:

  • Incorporated companies only (sole proprietors are excluded)
  • Business must be 100% owned by the borrowers on the application
  • 40–60% of NIT less dividends paid is used as income allowance
  • Uninsured only; TDSR must be under 60%; Max LTV 80%

The question was based on a flawed premise — and Brain MB corrected it and found the solution anyway. That's the difference between a tool that answers questions and one that helps brokers do better work.

The episode also touched on the evolving bank statement landscape — with most alternative lenders now requiring 12 months minimum, and some moving beyond that — and flagged a TD dividend gross-up program that Brain MB couldn't initially locate in its records, which was subsequently surfaced through broker feedback.

5. Beyond a Standard First Mortgage — 39 Brokers

Bridge financing, construction draws, second mortgages behind another lender's first — these are the deal structures that catch brokers off guard, especially when a client is on the phone in a stressful situation and needs answers now.

Example: "Does Scotia Bank offer bridge financing and what is the minimum term?"

Brain MB returned a comprehensive breakdown:

  • Yes — at prime plus 5% (variable based on amount and length)
  • Additional fee of $750
  • Maximum term: 90 days
  • Always registered on title
  • Scotia does not use 100% of net equity — they divide net equity by 1.1
  • Property must be MLS listed; private sales considered on exception for strong files

That last detail — the 1.1 divisor on net equity — is exactly the kind of fine print that blows up a bridge deal at closing. Having it surfaced proactively, before the file goes in, is the difference between a smooth transaction and a panicked phone call.

"If you're on a call with a client and they need bridge financing, you can write on that call, go to Brain MB, ask the question — and at least get some immediate feedback that could soften the client situation and save the day." — Dean Lawton

How to Ask Brain MB Better Questions

One of the most practical segments of the episode was Dean and Jason's rapid-fire breakdown of how to get the best results from Brain MB — advice that applies to any AI tool.

  • One lender, one issue. Don't stack three questions into one message. Single-focus questions returned accurate lender policy in 9 out of 11 cases in testing.
  • Use the lender's full name. "First National Financial" and "First National Excalibur" are different products. Be specific.
  • Always name the province. Without a province, the tool defaults to Ontario. Lenders have different rates and policies in different markets.
  • For rates, name the market — not just the lender. Include term (fixed or variable), insured or not, and purchase vs. refinance.
  • Give it deal context. Treat it like a conversation with a knowledgeable BDM. The more detail you provide, the better the response.
  • Follow up, don't cram. Ask one question, get the answer, then go deeper. Three questions at once risks missing the most important one.
  • Always check the citation. Every answer cites its source. If something surprises you, verify it — but use the citation to know exactly where to look.

The ABW Layer: What's Exclusive for ABW Brokers

For brokers within the ABW network, Brain MB carries an additional layer of brokerage-specific intelligence:

  • Regional BDM contacts — direct phone numbers and underwriting contacts for every lender, organized by BC, Alberta, and Ontario
  • Submission requirements — who the submission agent is versus the agent on file, by lender and province
  • Preferred lender tags — ABW's preferred partners are clearly identified, with active promos surfaced automatically
  • Unlimited chat access — ABW brokers (on Mortgage Architects or DLC) get full, free, unlimited access to the chat feature

For non-ABW brokers, Brain MB remains free to use with access to lender policy search, the Rate Explorer, and a limited number of chat queries.

On AI Limitations — And Why That's Okay

Dean and Jason were candid about where Brain MB gets it wrong — and why that honesty is actually a feature, not a bug.

The tool is designed to cite its sources, flag conflicts in policy data, and surface broker feedback when it contradicts the written policy. Experienced ABW brokers who knew from practice that certain deals could be pushed through on exception provided feedback that went directly into the knowledge base — and the tool adjusted.

"You can't 100% trust the AI. This is a funnel to get you to the answer quicker — not a replacement for due diligence and BDM conversations." — Jason Marshall

The feedback loop is built directly into the platform. After every query, brokers can return to the conversation and report how the deal turned out. That outcome data feeds back into the knowledge base, continuously improving the tool's accuracy — particularly around exceptions that never appear in any written broker kit.

Key Takeaways

  • Rental income, non-standard properties, rate shopping, self-employed income, and non-standard mortgage types are the five areas where Canadian brokers are most frequently getting stuck.
  • Brain MB provides cited, real-time answers to lender policy questions — reducing a multi-step research process to a single query.
  • Rates are verified daily, with timestamps on every response and API connections to some lenders updating multiple times per day.
  • The tool is designed to surface conflicts, correct flawed question premises, and flag uncertainty — not hide it.
  • The best results come from specific, context-rich questions: one lender, one issue, the province, the deal type.
  • Brain MB is free to use for all Canadian brokers, with an enhanced layer for ABW network members.
  • AI is a powerful efficiency tool, but broker due diligence and BDM relationships remain essential.

Why You Should Listen to This Episode

Whether you're a seasoned broker looking to streamline your workflow or a newer broker trying to compete with colleagues who have years of operational memory, this episode delivers something concrete: a data-backed look at where Canadian mortgage files actually get stuck, and a practical framework for using AI tools to get them moving again.

The five themes covered — rental income, non-standard properties, rate shopping, self-employed income, and complex mortgage types — aren't edge cases. They're the everyday challenges that separate brokers who close efficiently from those who lose time, and sometimes deals, in the research process.

Listen to the full episode to hear the real query examples, verbatim AI responses, and Dean and Jason's candid breakdown of where the tool works brilliantly — and where it still needs your professional judgment to get it right.

Stay Connected

Subscribe to The ABW Tuesday Mortgage Memo for weekly insights delivered straight to your inbox.

Interested in joining a network that helps you grow? Learn more about A Better Way Mortgage Group and discover how we can help you build a better brokerage.

Enjoyed This Episode?

Subscribe to get notified when new episodes drop.

Get notified when new episodes drop. No spam, ever.