Everyone at Axiologik, employee or associate, must have an Axiologik CV. All consultancies use CVs as part of bids, tenders and proposals for clients and projects, and so this is a shop window into the skillsets we offer. They get reviewed and updated every quarter by the Practices. It's important that we set a high bar and have a consistent brand standard, and CVs all follow the same structure that highlights key skills and experience.
For a few years, our Talent Team would sit with the original document and recreate it by hand; as you can guess, this is a time-consuming task, demanding a lot of attention to detail. Copy the summary across. Read the whole thing again to pull out the technical skills. Reformat the dates. Check nothing had been dropped. It took about 30 minutes per CV on a good day, but when someone sent a scanned copy, or a CV built as an image, there was no text to copy at all, and it took a lot longer still. Multiply that across a growing business and pipeline and it stops being an annoyance and it becomes an operational constraint.
// 01Starting with the real problem, not the technology
We didn't set out to build an AI product. We set out to give our Talent Team their 30 minutes back. Cat and Jack described exactly where the time went, and it was the same manual steps every time: extract, restructure, verify, and repeat.
That is exactly the kind of work that suits intelligent automation, because the judgement sits in a few places and the rest is repetition.
The first version we built was an internal app built on Google's Gemini. It read the source CV, pulled the content into our standard template, and handled the image-only CVs by reading the pages directly rather than depending on copy and paste. That alone took the worst cases off the team's desk.
// 02A second pair of eyes, built in
The part we cared about most was accuracy. A CV is an individual's professional record, and it goes in front of our clients, so an AI that invents a qualification or a date is worse than useless. So, we built the quality check into the process itself.
The first pass produces the formatted CV. A separate, independent review step then reads that output back against the original and looks for anything the source doesn't support: an added skill, a changed date, a role that was never there. If it finds a problem, it sends the work back to be redone. We repeat this process up to three times and keep the best result. The final QC Check Report is returned, and we ensure that nothing goes out until it has been checked against the document it came from.
We later rebuilt the whole thing as a Claude skill. At Axiologik we are on the Claude Partner Network so using Claude internally fits how the team works and it made the review step easier to run and maintain. The principle held across both versions: format first, then verify, and never trust a single pass.
// 03The result
A CV that took 30 minutes now takes under a minute. The scanned ones that used to take longest are no longer a special case. The team spends the time it gets back on the work that needs a person and has supported our recent rapid growth: recruiting, talking to candidates, and getting more of them in front of clients. As the business continues to grow, that headroom has mattered more than the saving on any single CV.
So much time and energy has been saved by utilising what we have at hand, rather than manually doing it all individually. It massively helps keep everything consistent, the format as well as the language, and allows me to crack on with the stuff that I actually enjoy and WANT give energy to, rather than the mundane tasks.
CAT HARDY · AXIOLOGIK TALENT TEAM
// 04The wider point
This is a small internal tool, but it makes a point we put to clients often. The value in AI right now is rarely a grand reinvention. It is in finding the repetitive, well-understood work that eats into people's days, automating it properly, and building in the checks that let you trust what comes out. We do this for our own operations, and we help other organisations do the same through AxioIntelligence and our AI services.
If you have a process that looks like our thirty-minute CV, we can help you find it, cost it, and build something dependable around it. Start with the work that already annoys everyone. That is usually where the return on investment is.
