Our methodology

The Omni WAT Framework

The Omni WAT framework is how Omni.io scopes every AI automation engagement: we map your operations into Workflows (repeatable processes), Agents (judgment tasks), and Tools (one-off leverage) — producing an executable automation blueprint in one discovery session.

Map your operations
The three layers

Workflows. Agents. Tools.

Every automation problem in an SMB decomposes into these three layers. Get the decomposition right, and the build is fast, predictable, and legible.

Workflows

Repeatable processes that run the same way every time — built with n8n triggers, no judgment required, just consistent execution.

Sending order confirmation emails the moment a purchase lands
Updating your CRM automatically when a new lead is created
Syncing inventory between Shopify and a spreadsheet, or routing tickets to the right queue

Agents

Judgment tasks where the system must read context and decide — powered by LLMs (Claude, GPT-4o) with RAG on your product catalog, policies, and scripts.

A voice agent deciding whether a caller is a warm lead or a support issue
A chatbot deciding whether to answer, escalate to a human, or collect more info
A lead scoring agent deciding which contacts are worth calling

Tools

One-off leverage: single capabilities plugged in when needed — the integrations that give Workflows and Agents their reach.

Google Places API for location lookups
Shopify order status API for real-time order data
WhatsApp Business API for message delivery, Twilio for voice routing
How it's applied

One discovery session. One executable blueprint.

When Omni.io starts an engagement, we don't begin with a tech stack — we begin with your operations. We map every process in your business against the three WAT layers by asking three questions. The output is a scoped, executable AI automation blueprint, produced in a single discovery session.

1
What should always happen the same way?

Those become Workflows — deterministic n8n automations that fire on triggers and never miss a step.

2
Where does context change what we do?

Those become Agents — LLM-powered decision-makers grounded in your catalog, policies, and scripts.

3
What external systems does this touch?

Those become Tools — the APIs and integrations that connect the blueprint to the systems you already run on.

Why it matters

No black boxes

Most agencies build AI as a black box: you pay, something gets deployed, and you're left guessing what it actually does. The WAT framework gives you a clear mental model of what's being built and why. Every deliverable in your engagement is either a Workflow, an Agent, or a Tool — so you can see exactly where automation ends, where judgment begins, and which systems it all connects to. That legibility is why our clients can maintain, extend, and trust what we ship.

Worked example

Mapping a D2C Shopify brand

A Singapore D2C skincare brand on Shopify comes to us needing three things: customers kept up to date on their orders, incoming enquiries qualified before they hit the founder's inbox, and everything wired into their existing store. Here's how that maps.

W · Workflow
Order updates

"Every customer gets a confirmation and shipping updates" should always happen the same way — so it's a Workflow. An n8n automation fires on each new Shopify order: confirmation sent instantly, tracking updates pushed via WhatsApp at every fulfilment milestone. No judgment, no exceptions, no missed messages.

A · Agent
Lead qualification

"Is this enquiry a wholesale lead, a product question, or a complaint?" depends entirely on context — so it's an Agent. A chatbot on WhatsApp and Instagram DMs reads each message, answers product questions using RAG on the brand's catalog and policies, qualifies wholesale leads with follow-up questions, and escalates complaints to a human with full context attached.

T · Tool
Shopify integration

Both layers need reach into external systems — those are Tools. The Shopify order status API gives the Workflow live fulfilment data and lets the Agent answer "where's my order?" accurately. The WhatsApp Business API delivers every message. Each Tool is a single capability, plugged in once, used by everything above it.

Three requirements in, one legible blueprint out: one Workflow, one Agent, two Tools. That's the whole engagement, scoped in a single session — and the client can point at any piece and know what it does.

Map your operations with the WAT framework

One discovery session. A scoped, executable AI automation blueprint for your business — Workflows, Agents, and Tools, all in plain sight. Or write to us at joshua@omniio.ai.

Book your discovery session