Insight 1Out now
Trust Issues
How private lenders are using AI, and where they still want a person in charge.
Read Trust IssuesState of Private Lending | 2027 Outlook
The State of Private Lending
What 200 senior private lenders in the US and Canada told us about speed, risk, AI and where the market is heading in 2027. The full report arrives in November. The first Insight, Trust Issues, is out now.
Presented by Mortgage Automator. Survey by Lux Insights.
The full report · Coming November
Faster, busier, still stitched together
The flagship report: where the market is now, where lenders say it is heading, and how you compare. Ten chapters, one chart per page.
Insights
Short reports, one question each
Alongside the full report, Insights go deeper on a single topic from the survey. The first one is out now.
Insight 2Coming soon
More Insights coming soon
New topics from the survey, released through the winter. Unlock Trust Issues and we'll let you know.
Insight 1
Trust Issues
How private lenders are using AI, and where they still want a person in charge.
Based on a survey of 200 senior private lenders in the US and Canada, fielded August to September 2026.
Page 2
AI is already part of the job
Most lenders are not deciding whether to use AI. They already do, somewhere in the file. The open question is how far it should go, and adoption is level across the US and Canada.
84%
use AI in at least part of their lending
37%
use it widely or extensively
89%
expect to use more in 2027 (38% significantly more)
Page 3
Trust drops as the stakes rise
The same lenders who trust AI to read documents are more careful about letting it inform credit decisions. Among the 41 lenders who use AI for both, 80% trust the document review output and 56% trust the underwriting output.
What we see in Lendr's AI Underwriter
People rarely change what the AI reads from documents: between under 1% and 12% of AI-filled values were later edited, depending on the field. Underwriting flags are weighed by a person, with loan-to-ARV (69%), down payment (68%) and credit score (62%) reviewed most.
Lendr platform data. AI Underwriter on 924 loans at 123 lenders since April 7, 2026. Demo and trial accounts excluded.
Page 4
Who is leaning in, and what they are building on
Heavy AI users, the 37% who have adopted it widely, look like ambitious, growing lenders. They are bigger, more growth-minded and happier with their process. AI is part of how they run, alongside a bigger appetite for growth and risk.
Their technology is split. Only 17% of all lenders run a largely integrated platform, and only 37% of heavy AI users do. AI is often added on top of separate systems.
48%
of heavy users are enterprise lenders (vs 15% of light users)
77%
are more willing to take on risk than a year ago (vs 33%)
49%
expect a significant increase in AI use (vs 12%)
Page 5
AI as a first line of defense
More than half of lenders saw an attempted fraud in the past year. Among them, AI tools and people reviewing files caught it most often, and most expect the threat to grow.
"Fraud detection is getting harder as fake documents can look convincing even when basic checks appear clean during normal underwriting reviews."Canada, mid-size lender
55%
saw an attempted fraud in the past year
65%
of those caught it with AI tools (63% with people reviewing files)
56%
are very or extremely concerned about AI-enabled fraud in 2027
Page 6
Where AI should lead, and where people should
A simple way to sort each AI use case: by how much is at stake if the AI gets it wrong, and by how much lenders already trust it.
| Level | What it means | Use cases (trust among users) | The person's job |
|---|---|---|---|
| Delegate | AI does the task; people spot-check | Document review and data extraction (75%), portfolio reporting (76%), routine borrower updates (80%) | Check a sample; fix and feed back errors |
| Review | AI drafts or flags; a person approves | Compliance monitoring (81%), servicing and collections (73%), borrower screening (66%) | Review every flag before acting |
| Decide | AI adds signals; people own the decision | Underwriting support (64%), fraud detection (65%) | Make the call; AI never approves or declines on its own |
Methodology
Who we asked, and how
Based on a survey of 200 senior private lenders in the US and Canada, fielded August to September 2026. The survey was conducted by Lux Insights, and respondents came from the wider market of private lenders, not from Mortgage Automator or Lendr customers. This is 2026 data with a 2027 outlook.
200
senior private lenders surveyed
120 / 80
US and Canadian respondents
±5.8%
margin of error on the full sample
78%
manage $50M or more