The Six Pieces of an Organization: Miss One, Get That Symptom

In July I noticed something uncomfortable: the automated loops I had built were running once a week. I wrote back plainly that the team was not being proactive inside the loop I had set up, and that the intensity was far too low. They needed to run one to three times a day.

Nothing was broken. There was no bug to fix. But the feeling was a familiar one: an organization that moves without arriving. That feeling looks identical whatever the cause. People can be slow because they do not know where they are going, because they started in the wrong direction, because they are not yet good enough, because they see no reward, because they lack people or tools, or simply because nobody feels the need to hurry. Guess the cause wrong and you spend effort for nothing: you send a team to training when it only lacked motivation, or raise rewards for a team that still does not know the goal.

So I do not ask "who is performing badly." I ask: which piece is missing?

Six pieces, six symptoms

I use a six-piece diagnostic. It is a fairly common change-management framework. I did not author it and I do not claim to have invented it; my part is only the interpretation for a small company that runs largely on AI. Read it as a diagnostic chart: miss one piece, and that symptom appears.

• Missing vision: confusion.
• Missing strategy: false starts.
• Missing capability: anxiety.
• Missing reward: limited response.
• Missing resources: frustration.
• Missing urgency: apathy.

Here is how I read each piece, with one real example from HOPE.

Look far, then choose few

Vision, for me, means always looking three to five years ahead to see where the company is going. It is not a slogan on a wall. It is the answer to "what will we be in three years." HOPE's fits in one sentence: a beauty-from-within and wellness ecosystem for Vietnamese women. I once wrote: "I want us to build a system that leaves a legacy, not just look at today's small revenue." With that sentence in place, every small decision has a ruler to measure against: does this product, this channel, this feature take us there?

Strategy is the other half, and here I am deliberately strict: in any one year, only one to three core bets, the strongest ones. Short goals add up to the long vision; there is no shortcut. If I must say where HOPE is concentrating, I name only what can be measured. Customers come mainly from a community we own: paid advertising in 2026 is about 1% of revenue, roughly US$1 per paying customer. Operations run on AI, so cost grows with software, not headcount. And a clean brand that does not overstate what its products do. Choosing few has a price, and I do not hide it: about 66% of revenue currently comes from one plant protein line. I regard that as a concentration risk to watch, not an achievement.

A school and a game

Capability means turning the organization into a school, where the CEO must study relentlessly, but study exactly what solves the numbers problem, not whatever is interesting. For people, that is a daily habit of learning from those who have done it. For HOPE's AI agents it takes a more concrete form: each important agent has an automated test set built from real cases, run every Sunday. There are four today: operations advisor, 22 cases; lead routing, 16; Zalo customer replies, 15; scoring of sales-conversation quality, 13. The latest scores, on 10 October 2026, were 20/22, 16/16, 15/15 and 12/13. The rule that goes with them: an agent is given more work, and we switch to a newer AI model, only when its score matches or beats the previous one. That is a school in the real sense: exams, grades, and conditions for moving up. Missing capability shows up as anxiety, and clear scores are one way to make a team less anxious.

Reward means building a game so the team has the drive to fight. I will not describe reward levels or mechanics here; that is internal. What I can share is a design principle: the rules must be measured by real work. At HOPE, a customer belongs to whoever looks after them, with no grabbing, and whether care happened is measured by real messages, not by a "called" button anyone can press. In the same spirit, the system measures real response time: over the 30 days to 10 October 2026, the median was 1 minute and 80% of messages were answered within 15 minutes. A good game rewards only what the scoreboard cannot fake. Departments, internal rules, culture and benefits are all part of that game, and partners and customers need a game of their own too.

Leverage and the wall

Resources come down to leverage. The able are those who know how to use it, and the most successful companies all do; none of them does 100% of the work itself. For HOPE, AI is the largest lever today: a three-layer AI system handles orders, customer care, content, reporting and payroll, automating more than 60% of the work. But a lever needs a fulcrum. My principle is that AI reads broadly and writes narrowly: posts and customer replies are drafted by the machine and sent only after a person approves; the machine has no authority to move money; ad autopilot is off by default and capped at VND 1 million per day. A lever without a fulcrum lifts nothing; it only makes mistakes faster. Missing resources lead to frustration; with leverage, a small team can still move something heavy.

Urgency means always putting yourself with your back against the wall, so you fight with everything you have. Nobody can do this for you, and you should not wait for circumstances to force it; you set it yourself. I do it in a very concrete way: I state our weak points openly. Reconciled figures begin only in 2026, so I do not publish 2023 to 2025. Long-term customer retention is not yet measurable. Revenue is concentrated in one product line. Once said aloud, there is nowhere to hide. Alongside that, I re-evaluate the company weekly, monthly, quarterly and yearly, stay firm on the goal and keep walking that road. Apathy is the symptom of this missing piece, and it is exactly what I saw in the loop that ran once a week.

Merit, and the loop of do, measure, learn

Six pieces are not enough. I add a condition the textbook framework does not carry: the "merit" of a person in business, what we call phuc duc in Vietnamese. For HOPE it means doing business decently, keeping the product clean, and never overstating what it does. Our products are dietary supplements, and refusing to promise what we cannot prove is part of the product, not paperwork (see Legal guardrails are part of the product). It sounds moral, but to me it is very practical: an organization with all six pieces that walks the wrong road simply arrives at the wrong place faster.

All six pieces run inside one loop: do, producing real results with sources, checked against HOPE's actual situation; measure against the plan; learn, extracting the lesson and loading it into the system's memory; then repeat. This is also how I hold the supervisory role: the agents survey, propose, I approve, then they execute, report and improve, and I do not have to touch everything. The essay Only publish numbers that have been reconciled tells of one time the "measure" step rescued the "do" step; Returning customers are the real yardstick is about what to measure.

What investors and partners should take from this

Three things. First, when HOPE slows down, we diagnose before we act: each symptom points to one piece, and we know which piece to ask about. Second, the hardest pieces to talk about, capability and resources, are turned into things we can measure: weekly evaluation scores, response times, spending caps.

Third, I do not pretend all six pieces are complete. The thin part is resources beyond software. A company that depends too much on one founder is a company still short of leverage, so we are looking for connections in three directions: capital, channels and markets, and capability partners in manufacturing and research. The six-piece frame does not promise that no piece will ever be missing. It only ensures that when one is, we know what it is, and the next turn of the loop starts better than the last.