How to Tokenmax at Work Effectively as a Product Manager

Show, don't talk.

For everything you do at work, you can use AI to do it better or faster.

The trick is not to get overwhelmed by demos, learning videos from team leads, or internal AI newsletters.

Pick your tool, and the only tool you should pick is VS Code or its equivalent in your org.

Every task can be its own folder or workspace. Spend a couple of hours learning how VS Code works and the concepts that make you an expert at it: workspaces, skills, and so on.

VS Code is the only AI superpower you need.

And when you have a task to do, use VS Code + AI to do it.

You want to write a talk track for a deck you're working on? Start a new chat, share the PDF version of the PowerPoint, and ask Copilot to generate a markdown talk track.

You want to create a new deck for an upcoming summit and you're stuck? Share some existing resources like past decks and ask it to come up with a slide flow to get started. It'll get you to 80% very quickly, and then you can add your own creative sauce.

• Never use AI-generated slides at work. Everyone is using decks that look the same and feel AI-generated, people will know, and even if they're good, it undermines the work.
• Always put your own spin on it. Bring your taste. Curate the presentation, add your own color scheme or your company's brand theme.

You want to build a new dashboard? Give the chat some starter queries and build them out one at a time, adding each back to your dashboard framework to verify accuracy. Repeat the process and you can go from zero to a fully functioning, accurate dashboard quickly. Don't try to one-shot an entire dashboard, that causes more waste in the long run. You need to bring determinism to things like dashboards so you can actually trust the data you're working with.

You wish there was an internal tool for sharing all the ways AI can be used at work? Build a full-stack app, deploy it on your internal stack, and share it.

Spin up an OpenClaw agents in your approved internal setup, use AI to figure out the deployment plan in your infrastructure, and use it to grow your products, generate growth ideas, and execute on them, with you in the loop via Slack or Teams.

• Each OpenClaw agents could be for a different work stream you are assigned to at work.
• One could be for general GTM work of your product, one for improving your features (tracks metrics, finds improvement ideas, and iterates over your feature spec). You get the idea.