Blogs 2026

How a PA with no technical experience built a live refund platform with AI

Written by Makers | Sep 2, 2026, 4:25:09 PM

Ready to eliminate "low joy, high time" tasks in your team?

“I’m not technical in any shape or form," Samantha Chilvers told us at the start of her journey. Yet, just a few months into Makers’ AI Native programme, she had built and deployed a live web application supporting 230 Holiday Extras members. The dashboard has already cut her admin time by 75%, giving her around 19 hours back each month.

This is the story of how Samantha moved from manual email tennis to architecting her own digital solution - and how you can take control of your own "low joy, high time" tasks.

The non-technical myth

Before the programme, Sam had a clear picture of the kind of person who built software.

“Someone with a science degree or someone sitting writing lines of code. Definitely not me.”

As a PA supporting seven directors, Sam was used to problem-solving, anticipating what people needed and making work run more smoothly. But she did not see those skills as technical.

She says returning to formal learning was intimidating at the outset. Other learners had more experience with AI, and Sam was used to introducing technical questions with: “I’m really not technical, but can you help me?”


What changed was not her job. It was her sense of what she could do within it.

From ‘email tennis’ to one dashboard

Alongside her PA role, Sam runs the Holiday Extras Social Club, an employee benefit used by 230 members. It plays a key role in bringing Holiday Extras’ culture to life, helping colleagues make the most of experiences and events.

Refund requests originally arrived through emails and forms. Sam would need to understand each claim, find the right attachment, check the ticket cost and membership allowance, calculate the refund and go back to the member if information was missing.

At busy times, it meant hours of what Sam calls “email tennis” - on top of her day job.
Through the Makers’ AI Native programme, Sam started by mapping that process. She broke it into steps, identified what could be automated and worked out where human judgement still mattered.

She then built the Holiday Extras Social Club Refund Dashboard: a live web application that brings submissions, member data, business rules, evidence and human review into one workflow.

Members can now log in, see their remaining allowance, submit a claim and upload their evidence in one place. The system applies the relevant refund rules and moves claims into a structured queue. OCR functionality is also being developed to extract information from uploaded documents automatically.

The result is a process that once took around six or seven minutes per claim now taking approximately 90 seconds.

And Sam did not stop at making it functional. She designed the dashboard around Holiday Extras’ culture and made it simple enough that members would not need to work out how to use it.

“I didn’t want it to be a glorified Google Sheet… I wanted it to be inviting and fun.”

She designed the dashboard around Holiday Extras’ culture and made it simple enough that members would not need to work out how to use it.

Built to work. Built to be safe.

Building something quickly was only part of the challenge.

During development, Sam realised some of the tools being suggested to her would take the application outside Holiday Extras’ approved technology environment.

Instead of taking the fastest route to launch, she brought in the company’s Director of AI Engineering, worked with technical and leadership teams and moved the application into the approved ecosystem.

“I want it live, but… I’ve got to follow this process.”


Human review is built into the workflow too. Sam still checks claims before refunds are processed, particularly where automated document reading could introduce an error.

For Sam, this became part of learning how to build well: understanding the process, thinking through what could go wrong and putting the right guardrails in place.

Looking Forward

The biggest impact may go beyond the 19 hours Sam gets back every month.
The dashboard has changed the way she looks at work.

She is already exploring how AI could help the PA team deal with expenses and is building a shared toolkit to bring information currently scattered across Sheets, Drive and documents into one place. She also shares what she learns with her colleagues, helping others spot their own “low joy, high time” tasks.

Sam did not need to become a software engineer to build useful software. She already understood the problem, the people and the process.

Makers gave her the skills and confidence to turn that knowledge into something people now use.

“I’m building things that people use.”

Ready to eliminate "low joy, high time" tasks in your team?  

Discover how Makers’ AI Native programmes give your non-technical teams the tools, confidence, and guardrails to build real digital solutions.