A new report says - probably not
Canada’s mortgage industry risk overestimating its progress on artificial intelligence, according to a new report from Section AI that finds almost all employees are still using AI at a very superficial level.
Surveying 5,000 knowledge workers at large firms in the US, UK and Canada, the study shows only 2.7% of staff qualify as “AI practitioners” who have embedded AI into their workflows and see meaningful productivity gains; just 0.08 per cent are “AI experts”.
In total, 97% of the workforce is using AI poorly or not at all. A quarter say they save no time with AI, 44% save less than four hours a week, and 40% say they would be quite happy never to use it again.
The core problem is not prompt-writing, but what the authors call a “use case desert”. Some 26% of respondents have no work‑related AI use case, 60% say their use cases are beginner‑level and, when 4,500 reported use cases were analysed, only 15% were judged likely to generate real ROI.
The most common “most valuable” use is using AI as a Google replacement (14.1%), followed by drafting and editing text. Task and process automation, where most mortgage value sits – data extraction from bank statements, pre‑populating fact‑finds, triaging complex files, checking documentation – accounts for just 1.6% of use.
Perhaps most worrying for Canadian mortgage firms, the report finds a sharp gap between leaders and the frontline. Eighty‑one percent (81%) of C‑suite respondents believe they have a clear, effective AI policy; only 28% of individual contributors agree.
While 80% of executives say tools are readily accessible, only 32% of non‑managers report clear access; just 27% have received company AI training and 7% are reimbursed for AI tools. In practice, the staff doing the repetitive origination, underwriting support and compliance work are the least supported to use AI – precisely where the technology could have the greatest impact on cycle times and cost‑to‑serve.
The report’s message for Canadian mortgage leaders is blunt. Adoption metrics – licences issued, log‑ins, training completion – are flattering but misleading.
To turn AI into a genuine advantage, firms will have to manage use case development as a core competency, focus investment on individual contributors in operations and underwriting, and shift training from “how to prompt” to “how to redesign workflows”.
Until then, AI will remain a clever assistant at the edge of the mortgage process, rather than the engine of real efficiency that many business plans now assume.
