Clarity is a sales tool.

The objection always arrives in the same shape. Our buyers are sophisticated. If we simplify the language, we will look naive, and the head of model risk will assume we do not understand our own architecture.

It comes up in nearly every messaging session I run, usually from the most technical person in the room. It is also wrong in a way that has been measured.

The research says complexity costs you credibility

Daniel Oppenheimer ran five experiments on exactly this belief and published them in Applied Cognitive Psychology in 2006 under a deliberately absurd title, Consequences of erudite vernacular utilized irrespective of necessity. He took admissions essays and swapped short words for longer synonyms. He compared a plain translation of Descartes against an ornate one. He printed identical text in a hard-to-read font.

The direction never changed. As texts became harder to process, readers judged the authors less intelligent. The effect held regardless of the quality of the underlying writing, and it was mediated by processing fluency, which is to say the mere effort of reading was enough to lower the assessment. The paper won an Ig Nobel, which has caused people to treat it as a joke. The finding replicates the intuition of every good editor who ever lived.

So the sophisticated buyer does not reward density. Difficulty just makes them work harder for the same information and trust you slightly less at the end of it.

Finance already litigated this and lost

If plain language were a threat to technical seriousness, securities regulators would never have mandated it. The SEC adopted Rule 421(d) in January 1998 and published A Plain English Handbook that August, telling issuers to write the front of a registration statement in plain English and the rest in a clear, understandable manner. Warren Buffett wrote the preface. After four decades of reading filings, he wrote that he had often “had to conclude that nothing was being said”.

Europe went further and legislated a page limit. Under the PRIIPs Regulation, the Key Information Document for a packaged retail investment product may run to a maximum of three sides of A4 when printed, and must be written in language that is clear, succinct and comprehensible. A structured product with a payoff formula that takes a quant an afternoon to model gets three pages, by law.

Then hold your own materials against that standard. If a European regulator believes a retail investor can be equipped to decide on a capital-at-risk product in three sides of A4, your 46-slide architecture deck is a choice your team made, and the subject matter is taking the blame for it.

Your buyer is now legally accountable for understanding

This is the part that turns clarity from a style preference into a commercial requirement.

The FCA’s Consumer Duty came into force on 31 July 2023 for open products and a year later for closed books, and one of its four outcomes is consumer understanding: firms must communicate in a way that equips customers to make effective and properly informed decisions. The rules sit in PRIN 2A of the Handbook, with the final package set out in PS22/9.

Follow the chain. A bank buys your credit decisioning model. The bank now has to explain the outcome to a customer, evidence that the explanation was understood, and hold a named person accountable for the framework it sits inside. If your documentation cannot be translated into something a complaints handler can say out loud, then what you have actually delivered is a compliance problem with your logo on it.

And that gap is already visible in the data. The Bank of England and FCA’s 2024 survey of AI in UK financial services found that 46% of responding firms had only a partial understanding of the AI technologies they use, against 34% claiming complete understanding. The regulators attributed that largely to third-party models, where the buying firm has limited visibility. A third of all AI use cases in the survey were third-party implementations, up from 17% two years earlier. Eighty-four percent of firms reported having an accountable person for their AI framework.

Read those numbers together. Nearly half your market has only a partial grasp of what it bought from vendors like you, and almost all of it has appointed a named human who will be asked to defend the thing anyway. Banks own the accountability. Vendors own the explanation, and most are failing at it, which makes this the most winnable gap in the market right now. (A fourth edition of that survey is due, so a fresher figure is coming. I would expect the third-party share to be higher.)

Gartner puts a commercial figure on the same idea from the seller’s side: buyers with high decision confidence are twice as likely to report a high-quality deal as buyers with low confidence. Confident buyers are buyers who understood the material, which makes this an editing problem before it is a sales problem.

Precision and performance are different things

Simplifying does not mean stripping out terms of art. PD, LGD, false positive rate, drift, SR 11-7, threat-led penetration testing: these are load-bearing. A model risk lead reading “we monitor for drift” learns something specific. Replacing it with “we keep the model fresh” destroys information and signals that you do not know which word matters.

What comes out is the decoration. Adjective stacks that survive because nobody wanted to argue in the review cycle. Abstractions standing in for a number. Sentences built to sound institutional rather than to be understood, which is usually a sign the underlying claim is soft and the language is doing the covering.

Four edits that do most of the work

• Replace every abstract noun with the verb it came from. “Provides enhancement of operational efficiency” becomes “cuts review time”.
• Put the number in the sentence, not in the appendix. If there is no number, say so plainly and explain what evidence exists instead. Buyers forgive an honest gap and punish a vague one.
• Write one sentence your champion can repeat, from memory, in a committee where you are not present. Then test it by asking them to.
• Cut the first paragraph. It is almost always a warm-up written while you worked out what you meant.

The real test happens away from the demo. Ask someone three levels removed from the deal to restate what you do, and see whether they get it right.