Who they are

They run training courses, bootcamps and workshops across Lancashire, for businesses and for the people who work in them. The more courses they run, the more trainees they take on, and every trainee is a set of records that has to live in their learner-management platform before a course can properly start. Growth, for them, means more learners to set up and keep up to date. That admin scales in lockstep with the teaching.

The work, as it actually happened

At the start of every new course, one person sat down and typed every trainee's details into the learner-management platform by hand. Not a quick job: two to three days of one person's time, per course. And the courses kept coming, generally two a month.

There was a second version of the same problem. On the longer-running courses, every learner has reviews with their trainer, and the outcome of each review is written down in a spreadsheet. Every one of those then had to be copied back into the platform by hand — only a few minutes each, but multiplied across every learner, on every long-running course, after every review. It never stopped, and it all landed on the same desk.

Put together, that came to around eight days every month, the new-course setups and the review updates combined. The better part of two working weeks, every month, spent keying data into one platform before any of it taught a single learner. And it grew with every course they sold.

Why it hadn't been fixed already

The platform they use has no API and no integrations. There is no supported way to get data into it other than a person sitting at the screen and typing, which is why nothing off the shelf touched this and why it stayed a manual job for as long as they'd been running courses.

And there was a simpler reason underneath that: they didn't know it could be automated at all. As far as they were concerned, a system with no integrations was a system you keyed into by hand. The idea that the typing itself could be handed over wasn't on the table, because they didn't know it was possible.

What we did

We watched the real work. They recorded the process exactly as they did it, both the new-course setup and the post-review updates, so we were automating the real job rather than a tidy description of it.

We built it around browser automation, because there was no other way in. With no API to connect to, the automation does what the person did: it signs into the platform and enters the data directly. For the new-course job, it reads each trainee's details straight from their spreadsheet and creates the learner records in the platform, one after another. For the reviews, it takes the outcomes recorded in the spreadsheet and writes them back into each learner's record. The two-to-three-day setup job and the steady trickle of review updates, both done without a person keying anything in.

We reached a system other tools can't. Most automation tools need an API or a ready-made integration to work with. When a system offers neither, they can't help, which is why this job had stayed manual for years. Browser automation works differently: it drives the screen the way a person does, so it doesn't need the platform to co-operate. Very few tools can reach a closed system like this at all, and that is why a job the client had written off as impossible to automate turned out not to be.