Designing automation for Zapier

Design the automation before
you build the Zap.

Zapier is excellent at connecting tools. It is not a tool for deciding what to connect, why, or when a human should be in the loop. Henry designs the operational logic. You build Zaps that implement a system you've actually designed.

The automation design gap

Why do Zaps get rebuilt
three times?

You can wire your first useful Zap in an afternoon. That's what makes Zapier magic — and what makes it dangerous. Six months in, you have forty Zaps, half of them forgotten, most of them silently broken, and when one of them fires at the wrong moment it takes out a customer experience nobody knew depended on it.

The real issue is not node count or plan pricing. It's that each Zap was built as a tactical response, not as part of a designed operation. Nobody asked: which processes should be automated? In what order? With what human checkpoints? Which ones should have stayed manual? Those are design questions, and Zapier doesn't ask them.

Henry is the design layer that asks them. He maps your operation using the AAAERRR methodology from Deliberate Work, specifies each step across three planes, and runs a Fit Assessment that sorts every step into automate / augment / leave-manual. The Zaps you build after that are the ones that should exist — and the ones that shouldn't, don't.

If you're a solo founder or small operator, this is how you stop paying the friction tax of Zaps that run the wrong thing perfectly. If you run a team's operations, it's how you keep your automation layer in sync with how the work should actually flow.

Process selection up front

Henry's Fit Assessment tells you which steps deserve a Zap before you build one. Not every process should be automated.

Human-in-the-loop design

Steps flagged for human judgment become Paths + delay + approval patterns in Zapier, placed intentionally instead of retrofitted after a bad outcome.

Trigger logic from step intent

Work Plane intent and Execution Plane triggers translate into clean Zap trigger + filter logic. Fewer edge-case misfires, fewer guard filters bolted on after the fact.

How it maps

How do Henry steps
become Zapier Zaps?

Henry Specification Zapier Element
Work Plane — IntentZap name and purpose; action step configuration
Work Plane — InputsTrigger app event and incoming data fields
Work Plane — OutputsAction app outcome and downstream data
Execution Plane — TriggerTrigger event, schedule, or webhook
Execution Plane — Mode (AI)Zapier AI action steps and AI-by-Zapier nodes
Execution Plane — Mode (human)Paths, delay, and approval steps before commit
Experience Plane — Quality barFilters, formatter validations, error-handling branches
Fit AssessmentAutomate / augment / leave-manual — which Zaps to build at all

These are design-to-execution patterns, not a product integration. Henry produces specifications you translate into Zaps.

The specification depth

What makes a step
Zap-ready?

A step specified across all three planes already contains the trigger logic, the data contract, the quality gates, and the human-or-machine assignment. The Zap you build from it works on the first connect — and keeps working as the operation evolves.

Work Plane

Intent, inputs, outputs. What this step accomplishes, what it consumes, and what it produces. The strategic "what" and "why" of every atomic unit of work.

Execution Plane

Who performs this step, how it gets done, when it fires, and in what mode — human, AI, or hybrid. The specification that makes delegation possible.

Experience Plane

What the stakeholder should feel. The emotional and perceptual design of each interaction — the layer most operations never specify but always need.

Design the operation.
Then build Zaps that earn their keep.

Start with a diagnosis. End with an automation stack where every Zap has a reason to exist.

Talk to Henry →

Not sure if Henry is right for you? Book 15 minutes with Joe →