The problem
Complex income cases are where mortgage brokers earn their fees, but they are also where the research burden is heaviest. A client who is self-employed with two years of accounts showing different profit levels, a contractor paid via a limited company, a director drawing a low salary with dividends, or someone with a mix of employed income and rental receipts represents a genuinely different challenge from a straightforward PAYE case. Every lender treats these income types differently, and those differences matter enormously to whether the client gets the mortgage they need.
The manual approach is time-consuming and unreliable. You can ring BDMs, check Trigold or Twenty7Tec, or wade through sourcing system criteria notes, but the information is often out of date, incomplete, or hard to compare across lenders at a glance. A lender that was the right fit six months ago may have tightened their criteria. A new lender may have entered the space with more generous terms for exactly the income type you are looking at. Keeping on top of all of this while managing a busy pipeline is genuinely difficult.
The consequences of getting it wrong are significant: a declined application damages the client relationship and wastes everyone's time. A client who ends up on a suboptimal deal because you did not know a better option existed reflects badly on the advice given. And the time spent manually checking criteria for every complex case is time not spent on new business or client service.
The system
Step 1: Define the client's income profile precisely (Claude)
Before researching anything, use Claude to help you define and categorise the client's income in precise terms that match how lenders think about it. This avoids the classic mistake of researching the wrong category.
Prompt example: "I have a mortgage client with the following income situation: they have been a limited company contractor for 3 years, paid via their own company. In the last tax year their company turned over £95,000. They drew a salary of £12,570 and dividends of £42,000. The previous year dividends were £38,000. They are applying for a £320,000 mortgage on a property worth £400,000. How would mainstream UK mortgage lenders typically categorise this income? What are the main different approaches lenders take to contractor income, and what documentation would each typically require?"
This gives you a clear taxonomy of lender approaches before you start the detailed research, so you know exactly what you are comparing.
Step 2: Research current lender criteria (Perplexity)
Use Perplexity to research the current criteria for specific lenders or lender types. Because Perplexity pulls from live sources, it is more likely to reflect recent changes than a static criteria database.
Prompt example: "Which UK mortgage lenders are currently most favourable for limited company contractors where income is assessed using gross contract rate rather than salary plus dividends? What are their typical approach, documentation requirements, and any known restrictions as of 2024?"
Follow-up prompt: "For clients with director income using salary plus dividends, which lenders use both years' accounts and average them, which use the lower year, and which use the most recent year only? Please list by approach with specific lender examples where possible."
Verify everything Perplexity returns against current sourcing systems or direct lender criteria before using it in any recommendation. Treat AI research as a starting point for verification, not a substitute for it.
Step 3: Build a comparison matrix (Claude)
Once you have researched several lenders, use Claude to organise the information into a comparison format that makes decision-making easier.
Prompt example: "I have researched the following lenders for a contractor client: [paste your research notes]. Please organise this into a comparison table showing: lender name, how they calculate income, maximum loan-to-income multiple for this income type, required documentation, any notable restrictions, and an overall suitability rating for this client's situation. Flag any gaps in the information where I need to verify directly."
This structured output makes it much easier to identify the best options and explain the comparison to the client.
Step 4: Draft a criteria summary for your file (Claude)
For compliance and file note purposes, use Claude to draft a structured criteria summary explaining the rationale for the lenders considered.
Prompt example: "Please draft a file note summarising the criteria research conducted for this case. Include: the client's income type and how it is categorised, the lenders researched, the key criteria differences identified, the lenders that are potentially suitable and why, and any lenders initially considered but ruled out. Use clear professional language suitable for a compliance file."
The results
Before: A broker handling three or four complex income cases per week spent an estimated three to four hours per week on criteria research, typically spread across multiple BDM calls, criteria database checks, and manual notes. Information was often inconsistent between sources and not easily comparable.
After: The same research is structured in under an hour per case, with a clearer comparison framework and a compliance-ready file note produced as part of the process. One broker reported that the structured AI-assisted research approach led them to identify a lender option for a contractor client they would previously have missed, resulting in a mortgage approval that would otherwise have been declined. The ongoing time saving is estimated at two to three hours per week across the caseload.