FundMore

The Canadian open-banking rulebook is in the Canada Gazette.

On Friday, Finance Canada pre-published the proposed Consumer-Driven Banking Regulations. 60-day comment period closes August 26. The cost-benefit math is now on paper: $457.7M in costs over 10 years against $13.2B in benefits. Roughly 29 to 1.

Three observations every Canadian lending executive should hold:

1. The in-scope data is broader than the headlines: deposit accounts, payment products, registered and non-registered investment accounts, and secured and unsecured lending accounts. Account profile, balance, transaction, and product data within each. The raw-data layer just became a regulated utility.

2. The derived-data carve-out is the most strategically important line in the document. Data that is significantly enhanced by a participating entity to increase its usefulness or commercial value is excluded from the sharing obligation. Differentiation moves from holding raw data to enhancing it.

3. The Anti-Fraud Regulations pre-published the same day require express consent for EFT capabilities. Read the two together: CDBA institutionalizes consent for data; Anti-Fraud institutionalizes consent for money movement. Consent moves from paper artefact to production signal.

The decision layer is where this pays back. Foundation models give every lender the same baseline ability to read inbound CDBA data. The specializing signal that turns generic reading into a defensible credit decision is your own policy and your own underwriters' corrections applied over thousands of files. That signal has to stay with the lender, not with the vendor.

FundMore deploys as agents over a lender's existing LOS. No rip-and-replace. FundMore never originates, funds, or brokers a loan, which means the institution's borrowers stay the institution's borrowers and the data the institution processes through FundMore is governed by the institution's policy.

The raw data is a utility. The decisioning is the differentiator.

What is your institution's plan for the 60 days the comment period leaves open?

https://hubs.li/Q04p5ncF0

Open banking just stopped being theoretical in Canada.

On June 9, FirstOntario Credit Union went live on open banking under the Consumer-Driven Banking Act. One of the first credit unions in Canada to activate consent-based financial data sharing in production on Day 1 of the framework.

Three things in that announcement that matter for every Canadian lending executive:

1. FirstOntario picked its open banking partners eighteen months before the rule required it. The lenders who treated 2024 and 2025 as a building window are live now; those still waiting for "final clarity" are watching from the sidelines.

2. FirstOntario did not rip-and-replace its core to get there. The capability runs on top of existing systems through FIS Everlink and Flinks Outbound, which already has 300+ fintechs connected. Extending the stack, not rebuilding it, is the model that scales.

3. The infrastructure clock is real. Bill C-15 received Royal Assent in March. Technical standards are coming in the next few months. OSFI Guideline E-21 comes into full effect on September 1. Every open banking partner becomes a third-party relationship to inventory under B-10. The institutions that were built early are also documented early.

For underwriting, the strategic question is what runs on top of the new inputs. Consent-based data feeds compress document collection; the decisioning layer is where the competitive advantage lives.

The model is not the moat. Foundation models give everyone the same reading and reasoning capability. What separates good underwriting from generic underwriting is whether the AI is trained on this lender's policy and corrected by this lender's underwriters. That specializing signal is the moat, and it sits with the lender, not the vendor.

FundMore deploys as an agent within a lender's existing LOS. No rip-and-replace. No 12-month IT project. Pure infrastructure that never originates, funds, or brokers a loan, which means it does not compete with its customers for borrowers.

FirstOntario showed what early looks like. The lenders building on top of open banking infrastructure now, with policy-trained AI applied to the cleaner inputs it creates, will define the underwriting standard for the decade.

https://hubs.li/Q04lD_mk0

The headless LOS has arrived, and lenders are finally in the driver's seat.

For two decades, lenders bought workflows they didn't write and bent their operations around them. That era is over. When Blend and Salesforce shipped MCP-powered platforms this spring, they weren't announcing the future. They were announcing they would no longer be defining it.

With MCP, lenders can now author their own workflows, change underwriting logic in days instead of quarters, and stop paying rent on a process that was never really theirs. The vendor still owns the substrate: compliance rails, audit trails, and data integrity. But the workflow belongs to the lender now.

The lenders who act on this will build real competitive advantage.

Stop renting your workflow. Start owning your operation.

https://hubs.li/Q04kQ53w0



---

The Headless LOS Has Arrived. The Lender Defines It Now. A month ago, Blend launched Autopilot MCP.

Most AI governance in lending is just photographing a moving target and calling it a policy.

Six weeks of committee work produces an approved tools list that is already outdated by the time it is ratified. FundMore CEO Chris Grimes calls this Control Debt: the compounding gap between how fast AI evolves and how fast your rules about it can be rewritten.

The solution is not less governance. It is smarter governance. Lock down the invariants like data handling, accountability, and decision verification, and let the implementation move. Govern the direction, not the inventory.

The lenders who freeze their stack to feel safe are the ones who will fall behind.
Read Chris's full article to learn how to build for adaptation instead:
https://hubs.li/Q04knVBX0

The average loan touches 20 systems before it funds, and none of them know about the file as a whole. That gap between good tools is where time, money, and borrowers get lost.

Our CEO, Chris Grimes, breaks down why the industry's "best of breed" approach created an architecture nobody owns and what it actually takes to fix it.

"You don't have a tool problem. You have an architecture problem."

👉 Read the full article: https://hubs.li/Q04hKxx20