Let's fast forward to 2030 and imagine a typical Canadian shopping for a mortgage.

By then, their first question won't be fixed or variable—it’ll probably be, "Where can I get trustworthy mortgage advice the fastest, for free, on my phone?"

Eventually, millions of Canadians will end up confiding in AI mortgage chatbots to some degree, expecting speedy, accurate, pressure-free mortgage recommendations and rate quotes.

And these bots won't just be boring text; some will greet you with eerily human avatars, using technology like that from Synthesia.
​ ​ ​

Forward-thinking brokers and lenders see this wave coming, and many plan to add bots to their websites to service existing clients or attract mortgage leads looking for 24/7 advice.

Some early-movers have already gone live—True North Mortgage, for instance, fields a bot called "Morgan."
​ ​ ​

Truth be told, brokers have been dabbling in bots for years. We had one at my old shop, IntelliMortgage, but it was an old-school FAQ bot where you had to program in most of the knowledge, semantics and logic by hand. (If I could only have those 1,000+ hours back!)

Those relics are now being replaced by AI bots that gorge on vast data sets, tap into large language models, capture intent, and spit out instant answers.

The risks

Bots take enormous planning and refinement. Rushing out an off-the-shelf or weakly programmed AI bot is less “innovation” and more “liability grenade,” for at least four reasons:

1. The stakes are high: Many consumers who try a lender's or broker's bot and don't find it helpful may never use that bot—or mortgage provider—again.
2. Consumers who love their bot experience will tell their friends and family about it, creating word-of-mouth referrals.
3. Competition will be brutal, with developers racing to build better bots; those who succeed could dominate the DIY mortgage researcher space.
4. If the information provided by the bot doesn't comply with federal and provincial laws and regulations, and bot owners don't disclaim properly, the mortgage provider could have a compliance nightmare on their hands (don't rely on “the bot said it” holding up in court).
   ​ ​ ​
   The legal minefield is vast: truth in advertising, privacy, anti-discrimination, consumer-credit rules, APR disclosures, and a parade of provincial fine print.

Let's start with that last one, how provincial regulators look at robo advice.

Borrowers can set aside worries of rising rates this long weekend; Friday's data will leave nothing rising except maybe BBQ lids.

Indeed, Friday was a good day for low-rate lovers, courtesy of GDP temporarily falling off a cliff. And should next week’s jobs figures bomb as well, mortgage rates could tumble further.

The world heavyweight champ of inflation data, U.S. core PCE, nailed expectations dead-on. It also climbed to its highest in five months while U.S. Treasuries remain unchanged. That left many to wonder, why aren't bond traders more worried?

Plummeting exports and business investment did a number on Canada's economy, but that was predictable as November rain in Vancouver.

Today’s GDP release is proof that you have to dig below the surface to infer how mortgage rates might react. Hence why we spend extra time dissecting the numbers, instead of recycling headlines like a parrot that just discovered BNN.

The annualized headline number was hideous, no doubt. But under the hood, the domestic engine is still running. It's more jalopy than race car, but it hasn’t coughed up smoke just yet.

Thursday’s macro data looked like a Rorschach test for economists—everyone saw what they wanted and yields went nowhere. The real action comes Friday morning, with Canadian GDP and U.S. PCE inflation.

Wednesday featured no major macro drivers. That gave traders extra time to watch the White House try to muscle the Fed, like a backseat driver grabbing the wheel and insisting they know the shortcut.

U.S. bonds staged a rally Tuesday, and some of that momentum drifted north, dragging down our 5-year yield by 2 bps. (Bond prices and yields move inversely.)

For now, Canadas remain adrift at sea, awaiting trade or inflation clarity. Until our 5-year bond (now 2.96%) closes above 3.04% or below 2.90% (August's range), the daily 2-5 bps zig-zags we've been seeing are mostly noise.

Episode #2 of Stress Test This stars the man working to make Pine a household name in mortgages: CEO and Co-founder Justin Herlick.

As you're about to see, Pine does a lot of things right. Its method is part stopwatch, part spreadsheet: time and measure everything, take every wasted second out of the origination process, chop funding costs, and grow its net promoter score—which Herlick says maximizes word-of-mouth and lowers Pine's customer acquisition costs.

Pine is no tech slouch, but its people matter just as much as its automation. The company has low agent attrition because it does all the client chasing ("top of funnel work") for its reps, delivers ultra-low rates and fair compensation based on customer satisfaction.

We also dig into how Pine is leveraging AI, how it uses AI coding tools, how it scores agents in real-time, its philosophy on AI search, how it manages back-end funders, and where Herlick sees broker market share in 2035.

For mortgage pros allergic to marathon interviews, this one’s a quick shot of competitive insight—practical takeaways from an industry leader in our signature 9.9 minutes. Enjoy...