How we’re using AI + automation to scale

Hey, it’s Arik.

If you’ve been following along, you know I’m deep in the world of AI + automation, but not in the way most people are talking about it.

Over the past few months, I’ve been rethinking the way we run Spring. And it started with a simple question: “Where are we still doing things manually that don’t need to be?”

And that led me to N8N.

If you’re not familiar, N8N is an open-source automation tool that lets you connect different platforms, tools, and workflows. Think of it like Zapier, but with way more flexibility and control. And when you layer in AI agents and LLMs, it becomes something way more powerful than just a workflow builder. It becomes a creative assistant, an operations teammate, and a way to multiply output without multiplying your headcount.

Here’s what we’re building with it right now, what it’s replacing, and how you could start doing the same, especially if you’re running a brand and want to scale without all the extra overhead.


What We’re Automating Right Now (And Why)
We started by mapping out every single task done across the business. Department by department, role by role, and turned that into a full internal SOP.

Then we asked:
Which of these tasks can be automated?
Which of them should be?

The ones we prioritized were the high-volume, high-frequency tasks. The stuff that eats time and doesn’t need to.

Let me be clear, the purpose of this isn’t to replace people with AI, but about buying back our team’s time so they can focus on deeper work


Here’s a glimpse at what it looks like:
• Fully automated client onboarding:
When a new client signs, we use intake triggers to automatically create records, spin up folders, assign internal owners, and notify the right teams.

• Automated deliverable reviews + approvals:
We built routing logic that sends new creative work to the correct approvers based on project type, with SLAs and reminders baked in so it doesn’t fall through the cracks.

• Meeting follow-ups and tracking:
After client pulse calls, the system logs discussion points and even generates basic scoring based on sentiment and outcomes, so we’re not relying on memory or sticky notes.

• Finance trigger points:
When a key milestone is hit in a project, the system preps a draft entry for invoicing or reconciliation , with human review steps built in, so we keep control.


Why We’re Doing This:
The goal is to protect my team’s energy for the stuff that actually matters. And because they’re not wasting mental energy on low-leverage tasks, they’re able to think more clearly about systems, and we’re spotting weak points in our processes and replacing patch jobs with scalable solutions.

For example, if a project manager used to spend 30 minutes moving assets between tools, writing checklists, and notifying people, and now it takes 5, that’s 25 extra minutes they can spend catching bugs, spotting opportunities, or helping the team think ahead.

Every automation we build is scoped through that lens: Does this make someone’s job easier, faster, or more consistent?

We’re not automating everything. In fact, some things get worse when you force AI into the wrong spot. That’s why we build it case by case. We map every step, understand the logic, and keep people in the loop where nuance matters.

We also add safety rails to every automation, error handling, fallbacks, notifications, so nothing runs silently and breaks things in the background.

What This Means for Brand Owners

1. Start with one repeatable workflow
Pick a single task that happens every day or every week (ideally something with clear steps and a predictable trigger.)

Examples:
• Every time a customer fills out a form, an Airtable row gets created, a Slack alert goes out, and a customer support ticket is auto-assigned.
• Every time a new product is added to Shopify, it gets drafted as a feature in your email campaign board.
• Every time an ad concept is approved, an AI agent creates 3 script variations, drops them in your creative tracker, and tags the content team.

Use that to build your first N8N workflow. You can layer in logic, filters, and integrations as you go. The hardest part is just documenting what happens step-by-step. But once you’ve mapped it out, the build itself is mostly drag-and-drop.

2. Connect your systems
Connect systems that already hold your context.

For example:
• Connect Typeform → Slack → Notion to streamline hiring or intake forms.
• Connect your creative tracker → GPT-4 → Notion to generate first-draft captions or ad hooks for upcoming products.

It’s kind of like plumbing: once it’s connected, your operations start running more smoothly.

3. Add in AI as a support layer, not a full replacement
Let’s say you’re generating weekly UGC scripts for new product launches.

You can have a GPT agent generate three tone options from your product page, add them to a Notion database, and push them to Slack for review.

The human still provides final feedback and tweaks, but 80% of the heavy lifting is already done.

That’s the balance that works:
• Use AI to prep, draft, summarize, and route.
• Use humans to refine, approve, and direct.

4. Monitor, tweak, and expand
Your first few workflows won’t be perfect. That’s normal.

But once you’ve built one or two and they’re running smoothly, you’ll start spotting more opportunities.

See you next week,
Arik

Meme of the Week: