Our Current AI Risk Matrix

Two blog posts in one week.

THIS one, written Friday, didn’t make it to your inbox (unless you're on the new mailing list).
Short read. Worthwhile message.

But it’s always the most recent post that shows up Sunday mornings - the OG email list.

And honestly, this collection of thoughts on a Saturday felt more important.
So I invested about 3hrs rewriting and reworking it - I hope you get as much out of reading it as I did writing it.

It’s about AI.
Everything is about AI.

And risk.
Also regulation, or lack thereof.
And it's about you, me, and our industry.

When it comes to AI, there are currently two types of mortgage professionals:

The dabblers…
Dabbling with a recipe, a travel itinerary, cleaning up an old photo.
Cool, we all start somewhere.

…and then we have the do-ers, the people already getting 2-3x more done every week.

Faster content creation.
Faster communication.
Faster organization.
Faster research.
Way faster idea generation.
Faster client preparation.
Faster.

More output.
More volume.
More leverage.
More.

Now before somebody jumps in with:
“Yeah but AI is supposed to save us time…”

Let me ask you something.

When we add lanes to highways, what happens?

Traffic gets worse.

When internet speeds increase, what happens?

We consume more content, at higher resolution.

When humans get tools that help them move faster, what happens?

We do more.
We build more.
We work more.

And this time is NOT different. The equation is always the same:

Faster + 24hrs = More
There's always the same 24hrs.
Therefore Faster always = More
Faster = More

Maybe that right there was the 'short version'?

Let's all be honest with ourselves here.

There is what we say we want…

…and then there is our actual track record.
There is what we actually do.

Nobody in this industry is suddenly down to a relaxed 20-hour work week because of AI.

Maybe a handful claim they are.

I doubt it.

More likely:
they’re producing more,
responding faster,
building more systems,
creating more content,
handling more conversations,
and increasing expectations for everyone around them.

And here’s the thing:

The gap between the dabblers and the do-ers is going to widen dramatically.

Much faster than the internet adoption curve from 1995-2005.

What took ten years back then…
might take one year now.

Maybe less.

Actually…
definitely less.

So yes:
lean in.
Move fast.
Adopt.

But also ask:
How?
Where?
When?
And with what level of risk?

Because right now there are very few guardrails beyond basic common sense.

And ya, common sense... we know, we know.!

And yet, most conversations, presentations, and AI tools inside the mortgage industry focus almost entirely on speed and efficiency.

Almost nobody is talking seriously enough about risk.

Hang in there, I’m not about to tell you to avoid AI.
This is not a post by a grumpy old man resisting change.

Far from it.

I am leaning in.
Further and faster with each passing week.

But in some cases, we may want to ask a few more questions before plugging a powerful new tool directly into our business, our workflow, or our client data.

Because AI does not just amplify productivity.

AI amplifies consequences.

And mortgage brokers happen to work in one of the most data-sensitive industries imaginable.

Tax returns.
Bank statements.
Employment letters.
SIN numbers.
Government ID.
Full financial histories.

Many independent brokers are effectively operating as:

sales professional
underwriter
marketer
compliance officer
document vault
legal department
cybersecurity department

…all from a laptop on the kitchen island.

And let’s be honest:

The average broker’s understanding of regulator policy, let alone PIPEDA obligations, is lacking at best.
Mine included.

To be fair, so is any sort of clear AI guidance from the (many) regulators themselves.

So here we are:
Struggling in an industry flooded with AI tools, being pitched nonstop through social media.

And remember:
there are VERY bad actors on social media.

Narco gangs in Mexico are literally pulling 100,000 likes on recruitment videos.
This is a real thing! What's that summer job you landed again kid?!?

So yes, that AI app you just downloaded might be perfectly clean…

…but the ad network inside it?
The “free plugin”?
The browser extension?
The random integration tool?

Maybe not so much.

That is often where the real malware lives.
The spyware.
The ransomware.
The breach point.

So the question is no longer:

‘Can AI make brokers more productive?’

Obviously it can.

The real question is:

‘Where should AI be trusted, and where should humans remain fully in control?’

That’s where a simple framework becomes useful.

Not all AI use cases carry the same level of risk.

Some mistakes are minor and easily corrected.

Others are catastrophic and permanent.

The two variables that matter most are:

Damage
&
Repair

How much damage occurs if the AI gets something wrong?

And how easy is it to repair afterward?

That creates four very different categories of AI use.

Low Damage + Easy to Repair?

This is where brokers should absolutely be experimenting aggressively.

• Marketing drafts

• Blog posts

• Email outlines

• Social posts

• Training simulations

• Brainstorming

• Content repurposing

• Image creation for posts

This is the modern productivity layer.

The brokers winning right now are not starting from scratch anymore.

They use AI to:

• Organize their day

• Prep for meetings

• Summarize information (from meetings, email lists, presentations, etc.)

• Repurpose content

• Generate ideas

• Accelerate execution

• Get unstuck quickly

AI gets them 80% of the way there.
Sometimes 99%.

Then human judgment takes over.

This is smart use.

Smartiness, not laziness.
Maybe I should trademark 'smartiness' but if my AI copy editor thinks it's good, it probably already has.

Be smart.
Be effective.
See the future coming.
Lean into the new uses as they arise.

Don’t become the person from the 1980s insisting we all need to memorize times tables because:
‘What are you going to do, walk around with a calculator taped to your hand all day?’

Well Mom…

As it turns out…

But as the damage potential rises, so must the level of caution.

High Damage + Hard to Repair is where things become dangerous fast.

• Uploading full client files into unsecured AI tools?

• Sharing borrower data without understanding where it is stored?

• Cross-border, cross-continent, cross-purpose data movement?

• Permanent storage of highly sensitive consumer documents… where exactly?

• Letting AI agents roam freely inside business systems... dicey at best!

And ya...

The original risk remains consistently high - those meddling kids!

Letting your kid install a free game on your work laptop... career ending in some cases.

And this is also where:

• PIPEDA problems emerge

• Regulator scrutiny increases

• Lender trust vanishes

• Lawsuits happen

• Future regulation gets written

Don’t become the test case.

Because the mortgage industry is probably one (more) major consumer data breach away from sweeping changes to document handling altogether.

And regulators always respond to the weakest operator in the chain.

Not the strongest.

So yes, lean in.

But lean in intelligently.

The future winners in brokering will not simply be the fastest AI adopters.

They will be the brokers who combine:

• Speed

• Judgment

• Security

• Compliance

• Restraint

AI is not replacing mortgage brokers.

But it is absolutely changing what separates a professional operation from an amateur one.

And over the next five months, not five years, that gap is going to become painfully obvious

DW

P.S. Join me for a 1-Day workshop in Calgary June 17 or Surrey June 18 for a deeper dive into this topic and so much more - Click HERE and type 'BTBB' in the header for complete details. Or click the city name - and commit!