The story behind Claira™
Lynda Harris | February 18, 2026
Write Group Founder and CEO Lynda Harris talks to Wendy Hughes, Head of Marketing, about the creation of Claira™, Write’s new AI tool that reviews the quality of your writing.
Image by Write Group
When AI first arrived on the business landscape, what were your initial thoughts about it?
Straight away, I felt compelled to start investigating AI very thoroughly for two main reasons.
The first was obvious: AI-generated written content. As people whose jobs involve producing written content, or helping others to write better, we were thinking, ‘What does this mean for us and for our clients?’
The second reason was that I also felt very curious about what AI could do. We’re always encouraging people to follow good writing processes and to use appropriate tools to help. So the question was, ‘Should we recommend AI for writing or not?’
How have you seen attitudes to AI evolve since then?
Over a very quick timeline, we saw attitudes shift from being very varied and experimental — with some people saying, ‘I’m never going to use AI’ and others saying ‘Oh, it’s amazing, it writes all my emails for me’ — to much more widespread adoption.
As a company that had focused on writing quality for many years, we had to respond to what we were seeing. It was clear that AI was going to become an important part of business life, so we began by encouraging people to just jump in and have a go. We were doing the same.
But you also saw problems with AI-generated content?
Yes. Right from the start we saw questionable quality.
Sometimes AI-generated content looked amazing, and you’d think, ‘Wow, that’s perfect — exactly what I need.’ But all too often, if you dug deeper, you’d discover that the content really wasn’t quite fit for purpose.
Actually, that’s still the case now. It can be quite hard to articulate what isn’t working and that can be for multiple reasons. Sometimes AI misses out on important nuance and insight that only a human can bring. Or it might miss a reference to something in the past that would give important context.
We know that AI can emphasise pieces of content in unhelpful ways, draw poor conclusions, or bury key points beneath label headings. Sometimes it just ‘doesn’t sound right’.
The upshot of all that is the term ‘workslop’ — now used to mean AI-generated content that doesn’t work and often causes more work for the human who hoped to save time.
So you felt that the Plain Language Standard was becoming more important than ever?
Definitely. We began training our clients how to prompt effectively: to give AI the context it needed to produce effective, reader-focused content.
And we kept coming back to our Write Plain Language Standard. It’s what we’ve always used to guide people in how to write and how to check that writing is effective. With the rise of AI, we found that the Standard became more and more important.
Here’s an example. A client using AI to write a report was dissatisfied with the result. We asked questions from the Standard about purpose, audience, key messages, document structure, and headings (AI tends to do headings very poorly). Asking and answering those questions made a huge difference by uncovering key features of the writing that weren’t working well.
We were shifting more and more toward the Plain Language Standard being a vital writing and quality assurance tool for people using AI to produce their content.
Is that when the idea for Claira was born?
Exactly. I thought it would very helpful if we could apply the Write Plain Language Standard as an AI tool so that people could quickly check the quality of their own or AI-generated writing.
At first, we started experimenting with just uploading the Standard and telling an LLM to assess the quality of the writing against the Standard’s framing questions and its 10 elements.
We saw that the LLM could pick up many useful things . But we also noticed that it completely missed some things or made observations that were unhelpful. It became clear to me that, even when guided by our Standard, LLMs assessed quality randomly and needed much more explicit guidance than the broad statements in the Standard. In the absence of that explicit guidance, the LLM was drawing ideas and ‘rules’ from random sources that were not always clear or helpful.
That’s how Claira was born. We decided to build a tool that people could use to review their writing against internationally accepted plain language criteria — and generally trust the results. We began a 7-month process of refining the criteria on which Claira makes its assessment. It was a huge job, with many setbacks, but we’re proud of the result.
You didn’t do this alone though, did you? Tell me about AI Native.
Definitely not alone! We were lucky enough to know Tim Clark and Phil Vinall, who had a company that developed websites and complex back-end systems.
One day Tim asked me, ‘Are you guys using AI?’ We said, ‘Yes!’ because we’d already set up working groups and joined relevant organisations, and we had been experimenting with AI and Copilot.
Tim and Phil, fascinated by AI and the huge gains it offered, had been thinking of setting up a new company focusing exclusively on AI. We essentially became their testing ground for ideas.
They’ve been wonderful partners and we’ve gone on quite a journey together. They run regular AI training sessions for our team, and are our partner in developing various AI tools, including Claira.
AI Native is now a strong voice in the AI space, and we couldn’t recommend them more.
Write’s sub-brand is ‘AI+Human’, but Claira works without a human present. How does that work?
Claira was developed by expert humans. It’s a tool that has literally hundreds of hours of human expertise behind it — critical thinking and decades of plain language experience, along with lots of tweaking, testing, and refining. When you use Claira, you are using Write’s expertise.
So, it’s a tool. It does one job well — it reviews your work against best-practice criteria and provides observations and advice for you, the human, to act on. In this case AI+Human means Claira+you.
As a bonus, Claira acts as a writing tutor, directing your thinking to best-practice principles that will improve your skills as a writer or affirm the skills you already have.
So does Claira fix my writing?
No. And this is very deliberate.
One of the best ways to avoid the ‘workslop’ I mentioned earlier is to make sure you are thoughtfully engaged in your writing. Claira’s questions and advice are designed to quickly connect you with your writing, to have you think, process, and respond as you would if you were writing independently. Claira’s guidance speeds up the process and hones your thinking in the areas of your writing that need it most.
For example, Claira might say, ‘Have a look at your sentence length. You’ve got quite a few complex sentences in here that are obscuring your main ideas.’ It’s not rewriting the sentences for you. Claira is simply giving you advice about something that you, the human, then need to consider and, where appropriate, improve.
How does Claira differ from me just asking Claude or ChatGPT to check my writing?
If you just ask Claude or any other LLM to put something in plain language, or to assess whether it’s in plain language, you have no idea what criteria the AI is using. We know, of course, that LLMs are trained on billions of pieces of information, and we don’t know what they’ve chosen as their specific criteria to make that judgement.
As I mentioned earlier, even if you upload the Write Plain Language Standard yourself, every element in that Standard requires deep context that needs to be explicitly spelled out for an LLM. In our training workshops, we might spend 30 to 60 minutes on any one aspect of the Standard, really helping people understand it and giving them the knowledge and tools to apply it.
Unless your LLM chooses all that deep knowledge and context reliably and consistently, and you know it’s going to use only that information, you can’t guarantee the result.
You might still get some useful advice, but you don’t know what its judgements are based on.
Claira is born of expertise you can trust.
Does Claira ever get it wrong?
No AI is completely foolproof, unfortunately! As with all AI, Claira can potentially make random mistakes. But much of our development time has been spent reducing the likelihood of that happening.
So who should use Claira?
Anyone who writes any document — on their own or with the help of AI — can use Claira to check its quality against the criteria that are in the Write Plain Language Standard. These criteria also reflect the principles of the ISO International Plain Language Standard.
Claira is for everyone. We even use it ourselves in-house to check our own work!
Claira™ is available now
The tool represents 35 years of experience in clear writing combined with cutting-edge AI technology. Claira offers practical, reliable guidance for anyone who wants to communicate more clearly.