Lever

Inspiration

Inspiration for Lever came from a few sources:

1. A friend of mine who runs a consulting and staffing agency, sharing with me how he is a main AI evangelist at his company and he's done seven trainings on AI. Yet it's still not adopted outside of his team. Even though he's made skills that automatically take an applicant's information and put it into their resume templates, the entire HR team is doing all of that manually and by hand. People don't remember it exists and they don't even know what they can ask of it.
2. Another friend shared how she has to categorize HSA and FSA expenses. Even though she thinks her company might have access to Co-Pilot and ChatGPT, she can't keep track of all the windows. She just looks up if an expense is eligible on her phone on ChatGPT and then categorizes it on her computer because she can't find her way to another one of these windows and then back. That's the main blocker.
3. Clippy being a go-to tool that helped the workforce get up to speed with the newer technologies they were meant to be adopting.
4. Tools like Claude and OpenAI are releasing screenshot-to-skill automated pipelines but that's still not solving the in-the-moment challenges workers face when they don't even realize what's exactly possible with AI.

This showed a massive gap in how even though many companies and people are adopting and wanting to adopt AI, that last mile stretch of AI delivery is missing. That's what Lever solves.

What it does

It's a symbol that's always on screen to help people be able to identify how AI could speed up tedious workflows that they're already facing and use it in the moment without them needing to be an AI expert to leverage AI. People log in with their enterprise Codex account to be able to actually use Codex in their work when otherwise they weren't.

How we built it

It was 100% built within one Codex ChatGPT Solve 5.6 terminal session. I did a brainstorming doc ahead of time and that was an initial.md file that was uploaded into the folder and then everything else was completely generated by Codex. From research to design to all of the technical components. It did leverage a handful of different skills that I've used when building other projects previously to make this efficient.

Challenges we ran into

Sometimes it got a bit too much in the weeds and the UX of this is really important in my opinion for adoption so I got a bit too fixated on that.

Accomplishments that we're proud of

At one point I needed an example of using Excel and I was having trouble myself to figure out how to capture that nicely to display to others. I then actually just used Lever to help me work through that for this submission so I was able to use the tool I built to solve a problem I was facing already.

What we learned

It was really easy to get all of this built in one session so I didn't have to use multiple different terminal windows, which was nice.

What's next for Lever

Ideally a partnership with OpenAI to be able to help push this out to your current enterprise customers so they get the most of their subscriptions and then also expansion for personal use as well.

Built With

• codex
• gpt5.6