Context
Aircon King was doing roughly ₱30M (≈$535K USD) a month in revenue with almost no digital infrastructure behind it.
- Email was personal Gmail, not Google Workspace.
- Finance ran on manual ledgers and spreadsheets.
- There was no shared company environment — process knowledge lived in individual people's heads, not in systems anyone could hand over.
- The team's AI exposure was near zero.
They weren't broken — they had clear demand and a proven service. But they wanted to lock in ₱100M (≈$1.8M USD) per quarter as a sustainable floor, not a spike. That's a system problem, not a marketing problem. The demand was already there; the infrastructure underneath it wasn't.
The Challenge
₱30M/month on Gmail and spreadsheets is a person heroically holding everything together — not a system. Three structural ceilings sat between them and their goal:
- Finance invisibility. At ₱30M/month scale, manual accounting is a liability. You can't see margin by service line in time to act on it, you can't close a month without heroics, and you can't make a confident decision about where to invest in growth because the data isn't legible.
- No shared operating environment. Personal Gmail means no company identity, no role-based access, no shared drive, no way to onboard someone into a process that only exists verbally. Every workflow depended on the specific person who happened to run it.
- A capability gap. Any automation dropped into that environment would rot the moment the builder left, because nobody on the team could operate it, extend it, or build the next one.
We agreed on the diagnosis: to hit ₱100M/quarter reliably, the operation needed legible finance, a company environment the whole team actually works inside, and a team upskilled enough to keep building after the engagement ended.
Approach
The engagement ran April–May 2026 as one build with four connected pieces — build, automate, document, train — so that what shipped would survive the handoff.
- 1. Google Workspace from scratch. Provisioned the environment and migrated the team off personal Gmail: accounts, company profiles, groups, role-based access, shared drive architecture. This is the foundation piece — finance automation needed structured email and ownership, and every documented process needed a shared place to live.
- 2. Finance-process automation. Rebuilt the recurring financial work on n8n + Claude Code. Mechanical steps run on event triggers; judgment-heavy steps (categorization, anomaly detection, flagging) run AI-assisted with a human in the loop. Manual finance ops came off the plate — the hours-per-week curve bent down while data legibility climbed.
- 3. Company-wide SOPs. Authored the operating documentation for the new environment: how each team runs its workflow inside Workspace, how to navigate the shared structure, and how to keep processes consistent as people change. This is what makes a migration stick instead of quietly reverting to old habits.
- 4. AI training — beginners to builders. Took the core team from introductions to ChatGPT, Gemini, and Claude through to hands-on builds of real systems, websites, and dashboards using Manus. The goal wasn't AI literacy — it was a team that ships its own tools.
The unlock is the pairing: every system shipped with the SOP that documents it and the training that lets the team run it. That's what makes a handoff real instead of a dependency in disguise.
What Shipped
- Google Workspace live. Domain, accounts, company profiles, groups, shared drive — stood up from scratch and stable.
- Finance automations running. n8n + Claude Code handle recurring work that used to eat hours weekly, with human-in-the-loop for anything outside known patterns.
- SOPs adopted company-wide. Each team operates inside Workspace against written process, not memory.
- A team that builds. Core staff went from AI beginners to producing their own systems and dashboards with Manus.
Timeline
- April – May 2026. Core build: Workspace provisioning and migration, finance-process automation, SOP authoring, AI training program.
- End of May 2026. Systems handed off running.
- Ongoing. Weekly advisory — reviewing what the team builds, tuning the automation layer, and advising against the ₱100M/quarter target.
Reflection
When a business is already at ₱30M/month with manual ops, the growth bottleneck isn't demand — it's the system's ability to run without any single person heroically holding it together. Digital infrastructure doesn't create sales; it stops a winning business from leaking time, data, and decisions.
A system isn't done when it works. It's done when it works without you. The finance automation matters, but the SOPs and the AI training are what made it safe to hand over — the client kept me on weekly advisory because they wanted a thinking partner, not because anything breaks when I'm not there.
What I'd repeat
Pairing every system with its documentation and its training. Most automation engagements fail at handoff: the vendor builds something clever, leaves, and six months later the business has reverted because nobody inside could operate it. The sequence we ran on Aircon King — stand up the environment, automate the process, document how it's operated, train the team until they can build their own — is the order I'd propose again for any operation above ₱10M/month still running on manual-era ops.