Copy article

Mortgage AI in 2026: "Human Assisted" Future or Automated Redundancy?

ended 12. December 2025

Stuart Cheetham at MQube has called it early: AI will "define" the mortgage market in 2026. Buoyed by a 36.9% spike in the volume of lending, he insists that while technology's foothold will strictly increase, it will "enhance, not replace, human input", acting merely as a "human-assisted function."

It is comforting sentiment for the intermediary market, but does it hold water? In my 10 years of experience with corporate automation, my experience with "operational efficiency" and "resilience" always favoured by leadership over time, does not include maintaining the same headcount. The promise of "human-assisted" tech is often a Trojan horse for systems designed to eventually get past human judgement altogether. If an algorithm can process a borrower's risk profile in seconds, the commercial pressure to eliminate the "human assistant" is irresistible.

Perhaps more worrying is Cheetham's turn to "tokenisation" to make mortgage debt tradable. Mixing AI-driven origination with complex debt repackaging reads uncomfortably like a high tech version of pre-2008 financial engineering. If we are automating the lending decision and the trading of the debt, then who is actually watching the risk?

We want to hear from the frontline:

  • Do you think the "enhance not replace" narrative, or is this the transition phase between direct-to-consumer, broker-free lending?
  • Cheetham highlights "tradable mortgage debt"--are we innovating liquidity, or are we just using AI to obfuscate the risk in ways that the market doesn't fully understand yet?
  • When AI "defines" the market what happens to the complex cases? Can a human broker still argue a case or does the algorithm have the last word?
  • Is the sector actually ready for this or are there lenders who are attempting to wrap fancy AI over old-fashioned IT that cannot process a basic PDF document?
  • Will these efficiency gains lead to better rates for your clients, or simply healthier margins for the lenders?

8 responses from the Newspage community

Copy all

Star Quote
Copy

Rohit Parmar-Mistry
Founder at Pattrn Data
The 'human-assisted' narrative implies a permanent partnership, but in the boardrooms of big lenders, the reality is closer to 2001: A Space Odyssey. We aren't pilots; we’re just 'Dave', tolerated until the algorithm decides we are a cost inefficiency. In my 10 years of corporate automation, 'human in the loop' at huge firms isn't a long-term strategy; it is a data-harvesting phase to train the software before the redundancy notices go out.

Cheetham’s 'tokenisation' vision is even darker, effectively The Big Short 2.0. In 2008, the market imploded on CDOs that humans barely understood. In 2026, we are handing the economy to black-box algorithms that no one can audit.

If an AI originates the loan, assesses the risk, and trades the debt, who is left to act as the Mark Baum figure calling out the madness? When the system marks its own homework, we aren't innovating liquidity. We are simply automating the next financial crisis.
Copy

Editor
Darryl Dhoffer
Founder at The Mortgage Geezer
Its worth taking note for sure, that AI, could easily prevail and take a consumate lion share of direct mortgage business from the consumer. However the bulk of AI business may only come from customers who have the perfect credit profile, easy affordability, no major overhanging unsecured debts with no complex income, and easily identified own savings for deposits - anything slightly more complex, will only encourage consumers to speak with a human advisor, which may incorporate some "AI" assistance. We are yet to see the real threat AI dominance - is it coming or is AI in the mortgage world, never going to be ready for the consumer requirement.
Copy

Editor
Kate Underwood
Founder & Chief People Strategist at Kate Underwood HR and Training
Intel inside, idiot outside is exactly what you get when lenders forget there’s a real, worried human on the other side of that mortgage application. Yes, AI can enhance the process – crunch numbers, spot risk, kill the dull admin, but it cannot replace a broker who can hear the wobble in someone’s voice and challenge “computer says no”. We still need the human factor – people buy from people, we don’t like automated helpdesks, so why would we want an AI mortgage broker making the call on a 30-year debt?
From an HR lens, the real danger isn’t the tech, it’s leaders hiding behind it. The minute your staff feel they can only say “the system won’t allow it”, you’ve killed accountability, damaged trust, and set your culture up to fail when things go wrong.
Copy

Editor
Scott Gallacher
Director at Rowley Turton
Whilst it's true that technology can enhance work rather than simply replace workers, this is arguably only a short-term factor. Over time, any genuine productivity gain must translate into fewer people doing the same amount of work. And in the mortgage market—where overall consumer demand is relatively fixed—those efficiencies almost inevitably point towards job losses rather than redeployment.

Put bluntly, no leadership team is going to invest heavily in AI only to maintain the same headcount for the same volume of lending. The “human-assisted” narrative may be comforting, but history tells us it tends to describe the transition period, not the destination.
Copy

Editor
Mitali Deypurkaystha
AI Governance Strategist & CMO at AISensa
We’ve heard this promise before. UK banks said online banking would support branch staff, not replace them. Then the branches disappeared.

Now the mortgage industry is repeating the same script with AI. ‘Human-assisted’ sounds reassuring, until the algorithm decides it’s quicker, cheaper and ‘good enough’ to cut the human out. The real test comes when the broker and the machine disagree. History tells us who usually loses.

We all watched The Wolf of Wall Street. Humans packaged debt, sold it on, and told themselves it was safe. Until the system collapsed in 2008. What’s frightening is that we’re now using AI to automate that same behaviour at scale. Haven’t we learned that faster, smarter tools don’t fix bad ideas? Without human accountability, this isn’t innovation. It’s amnesia.
Copy

Kundan Bhaduri
Entrepreneur, Investor and Landlord at The Kushman Group
What truly terrifies me here is the tokenisation element, Cheetham's vision of mortgage debt repackaged into tradable securities managed by algorithms that operate in darkness. This is the 2008 playbook remixed with artificial intelligence, and it should alarm anyone who remembers what happened last time we let complexity obscure risk. When machines originate the mortgage, assess the risk profile, and facilitate its transformation into financial instruments that traders barely understand, who precisely is watching for the catastrophe?
Copy

Colette Mason
AI Ethics Consultant at Clever Clogs AI
Automating mortgage paperwork is sensible. There's a lot to be shunted from pillar to post, basic form validation and chasing missing facts, querying discrepencies. As an applicant it's a stressful and onerous task. As a broker grappling with this week in and out, it's boring and tedious work needing exceptional attention to detail. Streamlining makes a lot of sense.

However, AI has a woeful track record with unbiased, transparent information provision and simply cannot be relied on for making judgements. In my experience, the C-Suite are not power users of AI. They delegate research and operational delivery to colleagues which means they are not best placed to decide if an AI system is fit for purpose.

The cost savings blinkers soon take over, and the value humans bring to processes gets significantly undervalued. Given that mortgages are one of the most important decisions we make as humans, it is not in our interests to have firms playing fast and loose with AI decision making.
Copy

Patricia McGirr
Founder at Repossession Rescue Network
The mortgage market is moving with the quiet rhythm of the next Big Short. Automation is being sold as progress, but every gain in efficiency makes the human in the loop more disposable. The lesson last time was not just greed, it was blindness. Nobody watched the system that was making the calls, and by the time the damage surfaced it was families, not traders, who paid the price. Now we are mixing AI-led lending with tokenised, tradable mortgage debt and calling it innovation, without universal audit trails, mandated human override or regulators who can see inside the code. If AI is going to define the market, someone must define the guardrails. Progress that erases the human safety valve isn’t innovation. It’s abdication. And when the reckoning lands, it won’t speak in tech jargon, it will show up in arrears, repossessions, and households blindsided by decisions made in rooms they never saw.