Will your customers and AI tools understand your financial information?

Write Webmanager | July 2, 2026

Thanks to guest contributor Frances Gordon, Director of Narratology, PLAIN Ambassador, for writing this blog.

Clear financial information now has two audiences: people and the systems that help people

I have heard many financial firms say they use, or want to use, AI across their businesses — from customer service to marketing, from knowledge management to content creation. A quick search will give you an overwhelming amount of advice about setting up agents and agents-for-agents, MCPs, and so on.

But there’s something far more important you need to know: before these solutions can actually work well, you need to fix your content. This blog explains what financial firms need to do to make the most of AI applications for their business.

AI tools depend on the information they are given. Take the example of a customer-service chatbot. If your source content is unclear, long, inconsistent, or badly structured, AI will probably give answers that are vague or incomplete. It may also miss key details that customers need before they make a decision.

Of course, the same need for good information applies with or without AI. A customer may search your website, read a letter, open a policy document, use a chatbot, or call your support team. In each case, they need information that is easy to find, understand, and use. They need consistency and structure in the content for it to be useful. It’s a good analogy for what you need to do to make your content clear enough for AI.

AI tools need clear source content

An AI tool can make information easier to access, answer questions quickly, or summarise a long page. It can even point customers to the next step — but only if it has the right information to begin with.

Let’s look at an example. Imagine a customer asks,

Will I pay a fee if I repay my loan early?

Let’s look at a worst-case scenario where the information for the answer is found across several pages. One page may use the term ‘early repayment fee’. Another may use ‘prepayment charge’. A third may refer to ‘administration costs’. The product terms may include exceptions that are not explained in the help centre.

In this case, the chatbot would give a partial answer at best.

Before financial firms rely on AI tools, they need to review the content those tools use. The key questions are simple.

• Can the tool find the right, complete answer?
• Can the customer understand the answer?
• Can the customer see what to do next?

AI tools need structure — create simple, structured information on each page

Customers and AI tools both benefit from content that is structured well. At very least, think of each page as having a simple, consistent structure.

• A title that says what the content is about
• Headings and subheadings that describe the content beneath them and also reflect customers’ questions and tasks
• A short summary at the start
• One main idea in each section and minor ideas organised into paragraphs
• The most important information before the detail
• Consistent words for the same product, fee, action, or rule
• Clear next steps labelled as action points
• Links to related information only where they help — always give only relevant information

It’s not really about brevity. Long content can still be clear. Short content can still be confusing. The question is whether the information is organised around the customer’s task.

Inconsistent words create avoidable confusion — create a glossary

Financial services often use many words for the same thing.

A bank may use ‘blocked card’, ‘frozen card’, ‘restricted card’, and ‘temporarily stopped card’. An insurer may use ‘exclusion’, ‘condition’, ‘limitation’, and ‘not covered’ in related ways. A pension provider may use ‘charges’, ‘fees’, ‘costs’, and ‘deductions’ without explaining the difference.

Customers may not know whether these words mean the same thing. AI tools may also treat them as separate concepts, especially when the content is spread across different systems.

A simple terminology list can help. It should say:

• which term to use
• what the term means
• which terms to avoid
• when a legal or technical term must be kept
• how to explain that term in customer language — for example, any illustrative examples to use.

This is especially useful for firms that work across countries and languages. If the English source content is inconsistent, translation and localisation become harder. The same problem can then appear in every market.

High-impact information needs special care

Some financial information has a direct effect on what customers do next — for example, fees, charges, missed payments, or exclusions.

These topics need special care. They should not rely on dense paragraphs or vague headings.

For example, ‘Important information’ is rarely a useful heading. It tells the customer nothing. A heading such as ‘When you may pay an early repayment charge’ is more useful. It tells the customer what the section covers and when to read it.

Clear headings are also useful for AI tools. They help the system identify where an answer is likely to sit.

How I can help: preparing help content before chatbot use

As part of a team, I recently helped a financial firm to prepare its content, so that its chatbot would be more successful in answering customer questions about current accounts.

Our content review found specific ways we could improve consistency and structure. This involved restructuring content into snippets, so that each part of the document was self-contained and treated as structured data, and not free-flowing ‘copy’. We made changes such as adding headings and summaries, grouping content in a way that answered customer questions, and including clear actions for each question.

Quantitative improvements in accuracy followed immediately, as did an improvement in customer satisfaction metrics.

Clear content improves search, service, and customer confidence

Clear financial information works across channels.

It helps customers search your website. It helps customer service teams answer questions. It helps compliance and legal teams review documents more consistently. It helps product teams explain choices. It helps AI tools find and summarise information more safely.

For firms that operate internationally, this is even more important. A single product may need to work across markets, languages, channels, and regulatory expectations. Clear source content gives every team a stronger starting point.

The practical place to start is simple. Choose one important customer question. Find every page, document, and message that answers it. Check whether the same words, explanation, and next step appear across all of them.

That small review will often show where the larger problems sit.

About Frances Gordon

Frances Gordon is a content consultant and trainer, with an interest in using AI to make communications clear. She is the UK Ambassador for PLAIN, the Plain Language Association International, and an expert contributor to the British Standards Institution on the development of plain-language standards. She specialises in clear communication for complex and regulated environments, with a focus on accessibility and financial inclusion. Her work helps organisations make complex information clearer and easier to trust. She is Director of Narratology.

See Frances Gordon’s LinkedIn profile or visit the Narratology website