Building Vietnam's AI-operated D2C HealthTech company

What I do

I run a D2C health and beauty company on AI, and I designed the system myself. I am the Founder, Chairman, CEO and COO of HOPE CORP (HOPE SERVICE CORPORATION), the company behind IKI Beauty & Wellness: inside-out health and beauty care for Vietnamese women and families. I also act as Chief AI Officer of my own company. AI here is not a side experiment next to the business. It sits inside daily work: customer care, content, reporting and risk control. I started as a software engineer and spent more than 12 years in Japanese companies, so process discipline and speaking in verified numbers come naturally to me.

Evidence from a real business

These are HOPE CORP's own numbers. They are reconciled, not rounded up for a deck.

• A team of 12 (9 full-time), working with a three-layer AI system connected to the CRM. More than 60% of the workload is automated.
• 1,294 paying customers from March to September 2026. 40.5% of new customers buy again, and repeat customers generate about 66% of revenue.
• Gross margin of roughly 60 to 65%. Net revenue for the first nine months of 2026 was about VND 3.57 billion (about US$137K), a run-rate of about US$170K a year. Advertising is only about 1% of revenue.
• A community of about 240,000 followers and about 9,700 Zalo members. IKI Academy has more than 400 lessons. The IKI Beauty & Wellness app is live on the App Store.
• Customer care: over the 30 days to 10 October 2026, median reply time was 1 minute and 80% of messages were answered within 15 minutes.
• AI quality is graded every Sunday by four automated test sets. Scores on 10 October 2026: 20/22, 16/16, 15/15 and 12/13.
• In October 2026, HOPE CORP became a member of Anthropic's Claude Startups program.

On control: AI drafts, and a person approves before anything is sent. The machine has no authority to move money. Automated advertising is off by default, with a spending cap of VND 1 million per day.

Frameworks I have built

AI reads wide, writes narrow

AI may read many data sources so it understands the whole picture, but its power to write, send or spend is narrow and always has a human check. Wide inputs make AI useful. Narrow outputs let me sleep.

A three-layer AI organization: executive, department, task

The executive layer sees the whole company, the department layer understands each function, and the task layer does small repeatable jobs. Separating the layers means each part can be measured, fixed or replaced without bringing down the rest.

The loop: DO, MEASURE, LEARN

Every job produces a real result, is measured against the plan, and the lesson is fed back into the system. Without the learning step, automation just repeats mistakes faster.

The six-piece model of an organization

Vision, strategy, capability, reward, resources, urgency. A missing piece produces its own symptom: confusion, false starts, anxiety, limited response, frustration, indifference. I use this to diagnose the team before I think about tools.

Publish only reconciled numbers, one set of numbers for every document

Every figure that leaves the company is checked against its source, and the same set is used for investors, partners and applications. A small true number beats a big one that needs explaining. Essay on this principle: Publish Only Reconciled Numbers.

Legal guardrails written in code

IKI is a dietary supplement. The rules on how we speak to customers are built into the content-checking system, so AI cannot cross them by accident, rather than relying on a reminder in a document.

Journey

• Kumamoto, Japan: studied Information Science (software engineering) at Sojo University. Scholarship every year, an outstanding graduation thesis, and the only Vietnamese student in a department of more than 150. Along the way I worked many jobs: night shifts at a convenience store, an auto parts factory, McDonald's, a restaurant, teaching Vietnamese, interpreting.
• More than 12 years in Japanese companies (FPT Software Japan, Scuti Japan): over 7 years as a Bridge System Engineer (BrSE), then Business Manager.
• Currently pursuing a Master's in Economic Law.
• 23 August 2023: founded HOPE CORP and built the IKI Beauty & Wellness brand.
• 2026: a team of 12 runs alongside a three-layer AI system; in October, joined Claude Startups; currently raising a Pre-seed round.

What I can speak about

• AI in D2C operations: from customer care to executive reporting.
• AI in health-related business under legal controls.
• Building a small team that uses AI as leverage.
• Raising capital with reconciled numbers.

A line I often say:

Everyone has their own time zone.

Work with me

I am raising a US$250K Pre-seed round for 20%. The planned use of funds is 40% to scale the brand, 30% to R&D for the beauty line and the flagship nutrition product, and 30% to channels and the repeat-customer program. I would like to talk with:

• Pre-seed investors interested in D2C and healthtech. Investor overview: /investor/.
• Channel and market partners who want to bring the brand to new customers.
• Manufacturing and R&D partners in beauty and nutrition.

Contact: contact@ikihealing.com. If you would like to start with one read, see Publish Only Reconciled Numbers.