@benbausili.com
Product Leader building things at InterWorks in Laravel. Data Strategist who wrote the updated Tableau white paper Designing Efficient Dashboards. Raising kids to see the world with love and complexity. Teacher at heart, theme park nerd. Remaining hopeful.
This has been my experience. The dopamine hits keep coming. It’s like classic video games that has you saying “one more turn” over and over.
benbausili.com/posts/ai-mak...
From my vantage point, right now I want all my developers to get management training. It feels like a lot of organizational problems.
The gap between "available" and "adopted" was bridged by the people in the room and not documentation.
Full writeup into how our internal workshop went: interworks.com/blog/2026/01...
A major takeaway was the need for support to be there to help people get started. Windows installation issues surfaced immediately. Without experienced users circulating the room, half the group would have abandoned before the first exercise.
We structured our workshop in three phases:
1. Markdown files (low stakes, talk to your data, generate a powerpoint, build a small website)
2. Data files (explore, join, create an excel, build a dashboard)
3. Independent problem-solving (bring your own challenge)
That distinction sounds abstract until you watch a non-technical employee build a working dashboard from their own company data. Not following a tutorial. Solving their actual problem.
One thing I've learned running Claude workshops at InterWorks: the lightbulb moment doesn't come from reading about it, it comes from seeing it work.
A chatbot generates text. An agent does work.
At Davos, Jensen Huang said Nvidia uses Claude Code internally.
The Verge is reporting that Microsoft developers have adopted it over Github Copilot.
Claude Code is having a moment and has moved from early adopter excitement to enterprise infrastructure.
New post on what actually works: benbausili.com/posts/entry-...
We need entry-level employees who act like founders, not gophers. People who ask "what should we build?" not "what do you want me to do?"
The companies hollowing out their junior pipelines today will spend years rebuilding institutional knowledge tomorrow.
Our last intern class blew us away—not despite AI, but because of it. Combined with mentorship and room to be ambitious, they tackled work that would have required mid-level engineers not long ago.
The old training ground was grunt work. The new training ground is decision-making under uncertainty.
Now I'm seeing the same logic applied to junior developers. Headlines about entry-level hiring down 25%. Companies announcing "no new engineers."
This gets it backwards.
Last September, I wrote about why we were betting on people while Salesforce celebrated cutting support staff.
Three months later, Salesforce admitted they'd been "too confident" in AI replacing humans. Service quality dropped. Savings evaporated.
I largely agree with these thoughts from Reid Hoffman (LinkedIn Founder). I think there’s a ton of value to unlock in this layer and the analysis potential is huge.
What organizational friction is slowing down your AI wins? How are you overcoming them?
The technical challenge? Gone. What's actually hard now is getting humans to agree on which template slides to use, what offerings to feature, how to structure client deliverables, what standards to adopt. The real work is understanding core problems and reaching consensus.
Second: I built a custom AI skill that generates company presentations using our branded template. Short conversation with Claude, complete .skill file, tailored to our standards. Done.
Time to deploy it so others could use it? Still waiting. I could push to Snowflake for easy hosting, but not everyone has access. I could use Railway, but then I need authentication, which means opening a ticket for an Azure app registration. The friction wasn't the AI. It was everything around it.
A couple back-and-forths later, I had something better than most enterprise dashboards I've seen. Maybe 20 minutes from frustration to working solution.
First: I got frustrated with Anthropic's built-in API reporting. Hard to see spend across multiple keys, especially with Claude Code spinning up new ones constantly. So I described the problem, added an admin key to my environment, and asked for a Streamlit dashboard.
AI isn't the bottleneck anymore. Your company is.
Two things happened in the last 24 hours that crystallized something I've been feeling for a while.
The tools won't do it for you. They amplify what you bring to them.
In a world of artificial agents, the most human thing you can do is choose how to use them well.
Full post: benbausili.com/posts/year-o...
At our company kickoff we saw a project manager automating status updates. A consultant using Claude Code to document Tableau dashboards. Several people standardizing processes from internal announcements to code styles. None of them waited for permission.
People want someone to hand them the answer. Meanwhile, the people who aren't waiting? They're discovering the future.
Over drinks, a friend recently told me: "A lot of people forget they have agency."
Last year was "the year of agents" in AI. But as I've helped skill up developers and roll out AI across our company, I've seen a lot of FOMO. And a lot of waiting.
AI has no face. It's not a person. It's not even a consistent thing. It's a mirror that becomes whatever you put in front of it.
The question isn't "What will AI become?"
The question is: "What will we make it?"
Full essay (with inspiration from Miyazaki): benbausili.com/posts/ai-has...
I finally got around to making ice tea with Good.Store’s Blue Fields Earl Grey and it might be my favorite thing ever. https://good.store/products/blue-fields-earl-grey-loose-leaf-tea?srsltid
Stop pitching the logic of AI. You're talking to the wrong part of the brain. My new blog dives into how leaders can use empathy to calm the emotional Elephant and guide their teams toward a future where AI augments, not replaces, our best people.
interworks.com/blog/2025/09...
This would be amazing
We're entering a new era where you can escape 'average' software.
AI is making it possible for anyone to build custom tools that actually solve their problems, without being a full-time developer.
interworks.com/blog/2025/08...