My co-founder Brian Schmidt is an operationally minded finance and legal pro. He's constantly looking for workflows to make him more productive and save time on manual tasks. We recently chatted about three practical workflows you can apply today that he's found useful using AI:
- Automating our support rotation
- Adding a bit of delight to his weekly finance presentation and
- Debugging SQL queries
In the conversation below, Brian walks us through each workflow using different tools. Take a look and apply these workflows to your own day:
Automating our support rotation (without writing code)
As a startup, one of the best things we can do to deepen customer loyalty is to be really responsive and personal with customer support. We believe that if a user takes the time to ask us something, they deserve a thoughtful and fast reply. To make that happen and to give everyone a turn managing support, we created a daily support rotation. Every weekday, one person takes the lead on replying to customers, and an engineer is responsible for any escalations. For a long time, Brian would go into our team calendar, figure out who was on, and then write a custom Slack message every morning. But even small daily tasks like that add up. So he built a workflow with Lindy, the AI agent builder, to automate the process.
Here’s how it works:
- Every weekday at 8 a.m., Lindy checks our team calendar.
- A custom AI prompt reads the support schedule and understands who’s primary and who’s secondary.
- Then it posts to Slack, tagging the right people, like:
“Good morning! Today’s support coverage is Stella (primary) and Travis (secondary). Tomorrow it’s Brian and Brad. Have a great day! ⭐️”
Bonus: The AI even includes custom Slack @mentions, using member IDs Brian added to the prompt. This might sound small, but it saves time every day, and anyone on our team can now adjust it without touching Zapier or code. It’s a perfect example of how AI can make tiny moments more efficient and more human.
Adding delight to data with weekly image generation
Every Thursday, Brian shares a presentation called Hoopy Numbers at our team meeting. It’s a look at the key stats we care about: signups, tasks captured, product engagement. But he always kicks it off with a playful cover image that sets the tone. Before AI, he’d use Unsplash and grab something seasonal. Now? He uses ChatGPT to generate original art.
Some recent faves include a hopeful robot painting a digital sunset, inspired by Gone With the Wind, a lonely whale floating past a giant “Hoopy Numbers” sign (this one took a few iterations… whales are tricky), and a Valentine’s robot in love, followed the next week by a hungover one.
The thread with ChatGPT keeps getting smarter. Because he’s used the same thread for months, it remembers our inside jokes and running themes. Sometimes all he has to say is, “Do a fun one this week” and it just gets it.
It might seem small, but these tiny moments of whimsy create team rituals. I genuinely look forward to seeing what image shows up in the deck each week. It’s a simple way to make data feel less dry and more human.
Writing (and debugging) SQL without second-guessing
Here’s where things get really practical. Brian leads ops at Hoop, so he spends a lot of time slicing and dicing product data. But he’s not a data engineer. And while he’s written his share of SQL, it’s not his favorite pastime. Now, he doesn’t have to be fluent because ChatGPT is his real-time pair programmer. He has an ongoing ChatGPT thread called “Hoopy SQL,” where he’s trained the model on our table structure. So instead of re-explaining the schema every time, he just says things like:
“Write a query to get me the number of email-based tasks for this user.”
ChatGPT writes the query. If it throws an error? He just pastes the error message back in:
“Error in line 25, column 35.”
And ChatGPT instantly debugs it and gives him a revised version. It’s like having a judgment-free tutor sitting next to you all day long. Even better: because the thread has memory, he can say things like “Use the same logic as that onboarding query from last week” and it understands the reference. What used to take an hour now takes 15 minutes... and NO context switching between a million docs or Stack Overflow threads!
“The thing about AI is that it’s not about being fancy, it’s about getting to utility.”
What this all adds up to
These workflows aren’t revolutionary. But that’s the point. They show how AI doesn’t have to be loud or flashy to be powerful. Sometimes it’s about:
- Saving five minutes a day, every day.
- Removing friction so more people on your team can contribute.
- Creating delightful little rituals that make your culture stronger.
- Unlocking a capability you didn’t have before (or dreaded doing).
As Brian says, “The thing about AI is that it’s not about being fancy, it’s about getting to utility.” Couldn’t agree more. Let’s use these tools to make work smoother, lighter, and more joyful. And if you’re curious, try adapting one of these for your own team. You might find your own version of a Hoopy Numbers robot along the way.