There’s quite a bit going on inside this thing. Research, analysis, your writing, the role, the company, and the steps between spotting a job and sending an application. We’ve been building those parts and getting them to work together. Deep industry discussions and work with our partners have helped sharpen where the effort goes. Plenty of moving pieces. One aim: make the whole thing useful to you.
Built behind the scenes
This is where some of that time went.
The gap between updates? Here’s one good example. We built an internal routing tool that lets us adjust which AI does which job, and what happens if it can’t deliver. You're never going to get these results directly out a consumer chatbot.
- 12+
- AI workflows
- 18+
- specialised models
- 5
- routing stages
- The right model for the job. More than a dozen AI workflows can be dynamically adjusted across a pool of 18+ specialised models. Research, analysis and writing ask different things of AI. We can tune each workflow to suit.
- New release? We put it to work. We experiment with every new release to find you the best results at the best value. The shiny new thing still has to earn its place.
- A five-stage safety net. A preferred model, then up to four fallbacks. If one has a bad day, the next can pick up the job. Built to keep things moving, rather than leave you high and dry at the first hiccup.
You get better-matched help, more resilience, and a platform that can improve as the models do. We get the complicated bit. Fair trade.
Quite a few moving pieces.
- We’ve been building a product that carries the context from research through to documents, interview prep and tracking.
- Each part has a job to do, and they need to work together.
- Industry research and discussions with our partners have helped refine the journeys and focus the features on real frustrations in applying for jobs.
- Plenty happening under the surface.
Yes, I left the house.
- Speaking engagements, mentor sessions and industry discussions have been part of the work too.
- Being recognised as an industry leader is lovely.
- Getting to talk through what we’re building with people who know the space? Very useful.
More going on behind the suggestion.
- More analysis layers bring your writing, the role and the company into the same picture.
- Better research-backed insight, more context for the suggestions, and interview prep that gets past the usual questions.
- More to work with when it counts.
Beyond the blank document.
- Research, resume and cover-letter work, interview prep, application tracking.
- The feature set has grown across the whole journey.
- Five clearer steps, with room to review and approve along the way.
- You still get the final say.
- That bit stays.
The wireframe got some clothes.
- A more considered interface, clearer next steps, and dialogs that give the task in front of you some room.
- The editor has had attention too, with a beta experience for reviewing suggestions.
- Your original words stay put until you approve a change.
- Fairly important for something called WroteItMyself.
The less glamorous bits count too.
- Account tools, specialist application workflows, and more checks behind the scenes have joined the build.
- Some of the newer tools are still being refined and tested internally.
- Plenty happening before we open the door wider.
The useful bits
So, what does it help with?
- Research
- Company context, backed by research
- Your voice
- Suggestions shaped by how you write
- Documents
- Resume and cover-letter review
- Interviews
- Preparation and questions to work through
- Tracking
- Applications, next steps and follow-up reminders
- Your say
- Review it. Approve it. Or put it your way.
Want a peek inside?
A few more refinements, a few more checks, then that insider peek for the email list. I’m very keen to get this into your hands. And perhaps write the next update before the weather changes again.
Join the email list for a special insider peek soon.
A little side quest
Some lessons travelled.
Our founder took lessons learned from comparing large datasets while building WroteItMyself, and put them to work on a rather personal one: 15,191 MRI images of his own head. The result is AdamBrain, an interactive viewer built from those scans.
Have a look inside AdamBrain