The mortgage rulebook is being rewritten in parts. The running order.

Regulatory change usually arrives as a single document with a date, and firms treat it as a project. This one is arriving as a sequence, over about two years, and the effect is that each piece looks smaller than it is.

The change that matters most is the third one

Remortgage affordability and term reduction are useful. They remove friction from two specific transactions and every broker has already felt the difference.

The third change is structurally different. Allowing a firm to interact with a customer without automatically triggering the advice rules alters what a conversation is. For years the safe answer to "what would you do in my position?" was a careful non-answer, because the alternative was to have given regulated advice without the file to support it. That defensive crouch cost customers a great deal of useful help, and it cost brokers a great deal of business that went to a comparison site instead.

The opportunity is obvious. The risk is subtler, and it is this: the boundary is now something you cross rather than something you are permanently on one side of. A firm that could previously say "everything we do is advised" now has two modes, and has to know which one it was in, for every customer, on every date.

Three things this makes into engineering problems

1. Mode, recorded at the moment. Every interaction needs a durable answer to "was this advised". Written after the fact, from memory, by the person whose conduct is in question, it is worth nothing.
2. The transition point. Conversations move. The customer asks the question that turns information into a recommendation, and the system needs to notice, mark it, and change what is required of the adviser from that sentence onward.
3. The evidence pack, assembled either way. An advised case needs suitability. A non-advised interaction needs proof that it stayed non-advised — which is a harder record to produce, because it is evidence of an absence. Assembling the pack and the evidence work are the two halves of this, and both are easier when the record was made as it happened.

What is still coming

FS25/6 set out four priority themes and said policy development would commence by the end of 2026 and continue through 2027. CP26/18 consulted, from 9 June to 28 July 2026, on first-time buyers and underserved consumers, specifically borrowers with variable income, older borrowers, and those with past credit difficulties; the FCA’s review page listed no response to it when we checked on 24 September 2026.

Separately, lenders have had access to modification by consent on the loan-to-income flow limit since the PRA’s July 2025 statement — PRA-authorised firms by application, FCA-only lenders by individual guidance.

The common thread in all of it is assessment of people whose circumstances do not fit a standard template. That is, unavoidably, a data problem: variable income is a data problem, past credit difficulty is a data problem, and a lender's appetite for either is expressed as a rule that a broker has to hold in their head across twenty lenders. Which is where the hours go, measured rather than felt.

What this does not do

It does not make lending easier in the aggregate; it redistributes who can borrow. It does not change the Consumer Duty, which sits over all of it. And it does not tell a broker what their own firm's policy should be — the FCA has removed a constraint, not issued an instruction, and a firm may perfectly sensibly decide that everything it does remains advised because that is simpler to supervise.

What it does do is make the record of which a much more important artefact than it was in 2024. If you are choosing one thing to fix this quarter, fix that. Our general argument about why the trail is the product applies here with unusual directness, and the register is where we keep every rule we cite, with the date it was last checked.