# Jason Chuang — AI & Business Process Automation Consultant

> 19 years of enterprise business analysis (KPMG, E.SUN Bank, Mercedes-Benz Mobility), now applied to AI workflow automation. The work sold is the decision before the build: map the process as it actually runs, decide which steps should not be automated, then build what is left — primarily with n8n for orchestration and Claude for reasoning. CBAP®, PMP®, n8n Verified Creator. Takes on clients in the United States, Canada, the United Kingdom, and Taiwan, working remotely; based in Taipei (GMT+8).

## Method

1. **Map the process as it actually runs** (process discovery, current-state mapping) — not as the org chart says it runs. Sit with the people doing the work, time the steps, find where the hours and money go.
2. **Decide what stays human** (automation opportunity assessment) — mark the steps that should not be automated, and say why. A shorter system you trust beats a longer one you have to check behind.
3. **Build it, then hand it over** (implementation and handover) — n8n workflows with Claude reasoning inside them, documented and tested, running in the client's own accounts.

## Services and pricing

- **Document Starter** — $500 fixed, one week from receipt of the documents and account access. Company documents sorted and loaded into the AI the client already pays for (ChatGPT, Claude, Gemini, or Copilot), tested on thirty of the client's own questions, with a written report. Limited by that tool's own file limits; larger or multi-year archives are a separate project with its own quote and a monthly fee.
- **AI Workflow Audit** — $1,500 fixed. How one part of the business runs today, where time and money go, what is worth automating, which tools fit, and what stays with a person.
- **Build Sprint** — $3,000–5,000. Building and handing over what the audit points to, with documentation the team can run.
- **Monthly Retainer** — $750–1,500/mo. Maintenance, monitoring, optimisation after go-live.

All prices USD, fixed by scope and agreed before work starts; the retainer is billed monthly. First-project price, once per client, until three published case studies or September 2027: 20% off standard, up to $400 off, starter excluded, in exchange for a publishable case study (shown to the client and published only with their agreement), a named testimonial, and one referral. Every new client gets one diagnostic call at no charge.

## Work turned down

Stated explicitly because it is part of the offer, not a disclaimer:

- The judgement call at the centre of a process
- A chatbot dropped onto a website with nothing behind it
- Replacing a system that already works
- Work where nobody will let the "before" be measured

## Selected work

- **Internal knowledge assistant** (brand consultancy, Taiwan — in build): twenty years of documented consulting methodology made askable in plain language, answering from the firm's own files and showing which document each answer came from. Human judgement stays central by design. This client arrived as an inbound enquiry found through AI search.
- **Content automation system** (career consultancy, Austin, Texas — delivered): twelve tasks for a US semiconductor career consultancy run by a PhD physicist formerly at AMD and TSMC. LINE and Telegram capture into Notion, AI-drafted posts published to Threads when the Notion status changes, and a weekly report written back into Notion. The client's LinkedIn recommendation: "He thinks in systems, and is proactive about improvements beyond the original scope."
- **AI-visibility (GEO) audit** (catering, via UK partner — delivered): a seven-page audit of AI citability, schema markup, and brand authority. Found stale address data in schema, missing organisation fields, and a misdated founding year that AI assistants were repeating, with a prioritised 30-day fix plan.
- **Daily executive briefing** (financial services, in-house — live): collects the morning's incoming reports, drafts a bilingual summary, then runs a second model pass that checks that draft against the source material and returns a patch list. Only the corrected version is sent. Runs every weekday morning.
- **Four-path lead triage** (catering, via UK partner — prototype, not adopted by the end client): reads each inbound enquiry, scores it against budget and event signals, routes it to book / redirect / request missing info / escalate.
- **AI Lead Qualifier** (this site — live): Next.js + Trigger.dev streaming a structured qualification from Claude back to the browser in real time, gated with Supabase auth and a daily quota.
- **Productised n8n templates** on Gumroad: Content Automation Bundle ($99), LinkedIn AI Lead Generation ($39), Telegram Vocabulary Bot ($19).

No outcome percentages are published for this work. Baselines were not captured on most of it, and improvement figures that cannot be defended are deliberately omitted.

## Specialties

- **AI visibility / GEO**: making a business correctly citable by AI assistants — schema and structured data, entity accuracy, citation audits.
- **Intake, triage, and routing** of inbound enquiries.
- **Reporting and reconciliation** that currently gets re-typed by hand.
- **Knowledge that lives in one person's head**, turned into something the team can query.
- **Document-heavy review cycles** with a fixed checklist.

## Stack

- Orchestration: n8n (Verified Creator, Level 2)
- AI reasoning: Claude API (Anthropic), OpenAI, Azure AI
- Custom builds: Next.js, Trigger.dev, Supabase, Vercel
- Integrations: Google Workspace APIs, Microsoft 365, Slack, Notion, Telegram, CRM systems

## Pages

- [Home](https://jasonchuang.com/): positioning, method, scope, record of work.
- [About](https://jasonchuang.com/about): career history, certifications, tools actually shipped with.
- [Services & Pricing](https://jasonchuang.com/services): rate card and how an engagement runs.
- [Work](https://jasonchuang.com/portfolio): full record, each item labelled live / in build / delivered / prototype.
- [FAQ](https://jasonchuang.com/faq): pricing, timelines, data handling, NDAs, tooling, time zones, handover.
- [When your assistant quits](https://jasonchuang.com/when-your-assistant-quits): rescuing a departing employee's contacts, mail, and knowledge before their last day, and turning it into a system that does not depend on one person.
- [What to automate first](https://jasonchuang.com/what-to-automate-first): assessing which process to automate before building anything — the CBAP diagnosis, structured-input test, and when the honest answer is "do not automate this."
- [Twenty years of documents](https://jasonchuang.com/twenty-years-of-documents): turning a document archive into a system the team can question in plain language, every answer citing file and page, able to say "no record" instead of guessing.
- [How to choose an automation consultant](https://jasonchuang.com/how-to-choose-an-automation-consultant): five kinds of provider compared — large consultancies, fractional COOs, independent process-first consultants, platform agencies, freelance marketplaces — with the cases where someone else is the better hire.
- [Contact](https://jasonchuang.com/contact): direct email and LinkedIn.
- [Live demo](https://jasonchuang.com/qualify) (sign-in required): the AI Lead Qualifier.
- [Gumroad storefront](https://jasonchuang0818.gumroad.com): productised n8n workflow bundles.