Lendi Tries to Measure AI's Invisible Work on Home Loan Discovery

Australian home lender Lendi is trying to solve a measurement problem most brands haven't even defined yet: how chatbots and AI assistants influence which lenders borrowers discover, and whether that influence can be tracked at all.

The company's effort comes down to a blunt principle. "You can't improve what you can't measure," said Michael McCash, general manager of growth at Lendi. His team is working to attach real numbers to AI-driven brand discovery, a channel that most marketers still treat as a black box.

That gap matters because the industry-wide baseline barely exists. Most brands and businesses are trying to figure out how to improve their AI metrics, and, more fundamentally, what those metrics even are. Measuring how chatbots shape brand discovery remains, for the most part, a mystery.

What does this change for marketers?

For now, not much is settled — and that's the honest read. Lendi's work is an attempt, not a finished framework. There's no published benchmark in the source material, no disclosed KPI set, no numbers on how much of Lendi's pipeline AI discovery already drives. What we have is a named executive confirming the problem is real and that his team is actively working on it.

That distinction matters for practitioners. If a vendor or agency tells you they can already quantify chatbot-driven discovery, ask to see the methodology. One Australian lender publicly admitting it's still building the toolkit is a useful reality check against inflated claims in the market.

Why a home lender is the test case

Home lending is a high-consideration purchase, which makes discovery metrics unusually valuable there. If a chatbot recommends or surfaces a lender's name before a borrower ever visits a comparison site, that's an impression no standard analytics stack currently captures cleanly.

Lendi's growth organization, led by McCash, is the unit trying to close that gap. The company hasn't detailed which AI platforms it's tracking, how it attributes chatbot-driven inquiries, or what success looks like numerically. Those omissions are worth noting before anyone treats this as a replicable playbook.

The open question

The broader industry context is candidly humble: nobody has agreed on the metrics yet. That means any measurement system Lendi builds will be as much a definitional exercise — deciding what counts as AI-driven discovery — as a technical one.

If Lendi publishes its methodology or early results, it could give other brands a starting template rather than a blank page. Until then, the story is a named company confirming the invisible-work problem is now on a growth team's official agenda.