Welcome. This list is my attempt to make sense of the common ways I turn to AI for help. It’s up to you to discover more patterns. Let me know what’s missing.
By the end of the post, you should have new ways to think about building with AI, Agents, and Zapier that will generate countless ideas.
1. Data Collector
Collect things
The idea is to have an agent collect things that can be used later by another agent. Don’t copy and paste. Instead, quickly spin up an agent. I use the Zapier Agents Chrome extension and I save things to a Zapier Table.
An example: I use an agent to collect metrics from an interactive demo app called Arcade and I save them in a table. I can use this data as knowledge for other agents.
2. Auto Reporter
Collect things > Agent tells you about the things > You learn about your collection of things
However you want to collect things, just collect things and put them in one place. Then, build an AI agent that uses your collection as a data source. Trigger on a schedule, tell it what insights you want to pull for what time frame, and the agent will send you its findings.
I use this pattern for monthly Youtube video metric reporting. The metrics are pulled into one place. I have it send me the report in Slack with directional guidance on content that is performing well.
3. On-demand Intern
Send a message > Agent does a thing > You get things done
Decide how you’ll wake your agent up: either by clicking a button, sending an email, reacting to a message in Slack or by clapping loudly (if you’re feeling rude). Your Agent can then take an action in whatever app or look at a connected data source for answers.
I use this pattern in email to block off time for something like prepping for a workshop.
Another example is for updating a tacking table for AI hackathons our team runs. When there is a conversation in Slack that’s worth tracking as an update for a particular hackathon, we reactji with the agent emoji and it will update the tracking table for us.
4. Data Enricher
Collect a thing > Agent researches things > You learn about each thing
You could collect leads, companies, websites, events etc. Then have an agent research what you collect. It can use one of your own data sources or the internet to research.
An example: I collect meetings, how about you? So, I use a call prep agent to prepare for any meetings my calendar has collected for the following day. It looks at external participants and tells me about who they are.
P.S. This agent has a template you can steal if you’re curious.
Wrapping it up
You don’t need to be fancy to get the most out of AI. In fact, subtlety is way more impressive. Make AI do things on the edges so you can stay at the center.
And the center’s worth building for.
Happy building,
Bryce
Let's Connect
Learn to build with Zapier
Subscribe to the Build Anything Newsletter.
.svg){kind=small}