Knack:
Apprenticeships, matched
Designing a two-sided apprenticeship platform from a blank page to launch in three months. The product helps young people find a trade that fits them, and helps employers find candidates who know what they're signing up for.
The challenge
AI is closing off the entry-level office jobs that used to give young people their first step into work. Meanwhile, trades and care roles can't be filled. The channels employers use to recruit apprentices were built for adults with CVs, not for a 17-year-old who doesn't know what being a sparky actually involves.
Knack started as an experiment with almost no runway. I had one week to put a working prototype in front of a potential client, and three months to launch a new brand, a mobile web app for young people and a self-serve dashboard for employers. Because it was an experiment, every stage of the build had to prove the product was worth continuing.
Two things we had to prove
This was a two-sided bet, so we needed evidence from both sides before investing further. These weren't just UX questions. They decided whether Knack had a business.
Before we built the full product, we launched a waitlist with a short questionnaire. Our goal was 2,000 sign-ups before launch, and we reached 4,340, more than twice the target. That showed real demand from young people.
The questionnaire also meant we could preload their answers at launch, so they could confirm their details instead of starting from scratch.
I ran discovery conversations with 10 employers across mining, aviation, energy, defence, transport and logistics. Despite very different industries, the same problems came up again and again:
- 8 of 10 said candidates start without understanding what the role really involves, and early drop-off follows.
- 6 of 10 are overwhelmed by applications but struggle to find the right people.
- Most struggled to get women to apply at all, with some customers seeing 10% women applicants against targets of 30 to 50%.
Volume was rarely the problem. Fit was. That changed what we built: Knack became a quality-over-quantity matching product instead of another job board.
How I approached it
As the Lead Designer, I led Knack from discovery through to launch. I worked closely with engineering, product and our sales team, and I mapped both sides of the product in parallel.
What I delivered
Knack went from a blank page to a launched product in three months. That included a new brand, a landing page, a mobile web app and a self-serve employer dashboard. The experiment has already validated demand on both sides, and revenue is starting to flow.
Key learnings
Time the waitlist to the launch
The waitlist proved demand, but conversion into the app hasn't been good so far. We started the waitlist too soon and kept paid spend running after we hit 2,000, and from the results I suspect much of it has gone cold. Next time I'd stop paid spend once we hit the goal and start it again when the product is live.
Lock the brand before launch comms
Knack started under the WORK180 brand. The new name was agreed two weeks before launch and the brand was created one week before, which meant a new domain for emails. Those emails landed in people's spam and promotions folders. Settling the name and domain earlier would have avoided this.
Set up the AI workflow first
AI is great for speed and for getting prototypes out, but it does get things wrong. Fixing those mistakes cost us more time and credits than setting up our workflow properly from the start would have.
Scaling the experiment
Now that employers can self-serve, the sales team is ramping up outreach this month. The pipeline already includes several large national employers. The company goal is $1M in revenue from Knack by the end of the financial year.
On the product side, I'm focused on bringing more of the waitlist into the app, re-engaging candidates when new matches or employer interest come in, and adding self-service payments. We're tracking every change against our weekly metrics.