Dissatisfied

Podcast 2 — In conversation

Speaker A: I had an interesting moment in a client meeting last week. The client kept saying they were 'a bit unhappy' with the delivery. I translated that in the report as dissatisfied, and they actually preferred that word.

Speaker B: That is a good instinct. Unhappy is vague and personal. Dissatisfied is professional and specific — it implies the problem is in the delivery, not the relationship.

Speaker A: Exactly. You can say 'the client is dissatisfied with the current release schedule' and it sounds like a solvable problem. 'The client is unhappy' sounds like a feeling you need to manage.

Speaker B: Right. And dissatisfied avoids the melodrama. It is serious without being emotional. What about the difference between dissatisfied and discontented?

Speaker A: Good one. Discontented is a bit stronger and often implies a longer-term state of unrest — like a discontented workforce. Dissatisfied tends to be more specific and tied to a particular thing or outcome.

Speaker B: So you could say: 'She was dissatisfied with the presentation, but overall she was not discontented with the project.' Both words, different scale.

Speaker A: That is a great distinction. And here is a common mistake: people often soften it unnecessarily. Rather than saying 'I am somewhat dissatisfied', just say 'I am dissatisfied'. The word already has a measured tone built in.

Speaker B: True. In an annual review, writing 'the employee is dissatisfied with current growth opportunities' is direct, professional, and actionable. It does not need hedging.

Speaker A: And in user research, you might say: 'Forty-two percent of users reported feeling dissatisfied with the onboarding experience.' That is a clear, usable data point.

Speaker B: Dissatisfied is one of those words that makes you sound precise and thoughtful. It is the right choice whenever you want to describe a gap between expectation and reality in professional or formal writing.

Podcast 3 — Prompt engineering

INST: Today we are exploring how to use the adjective dissatisfied when writing prompts for software development. The word is powerful in prompts because it directly names a user state that developers need to model.

STU: I see — it is not just describing feelings, it is telling the AI what to build for?

INST: Exactly right. Let us start with a UI design prompt. Imagine you are designing an employee engagement platform:

Prompt: Design a React component called EmployeeSatisfactionMeter. It should display a 5-point scale from Very Dissatisfied to Very Satisfied. Highlight the selected option with a colour gradient from red for dissatisfied states to green for satisfied. Add an animated transition when the value changes.

INST: Using dissatisfied in the scale labels tells the AI exactly what the negative end of the spectrum looks like. The output will include the right colours, labels, and emotional tone without further instruction.

STU: That is efficient. What about a database scenario?

INST: Here is a schema prompt that anchors around the dissatisfied state:

Prompt: Create a PostgreSQL schema for a support ticket system. Include a ticket_feedback table with columns: ticket_id, user_id, satisfaction_score integer, is_dissatisfied boolean generated as satisfaction_score less than 3, resolution_time interval, and agent_id. Add an index on is_dissatisfied for fast filtering of unresolved dissatisfied tickets.

INST: The generated column is_dissatisfied is derived from the score. By naming it explicitly, the AI generates a schema that is self-explanatory. Any query using WHERE is_dissatisfied = true immediately communicates intent.

STU: That makes SQL readable too. What about an API endpoint?

INST: Here is a good one for a Node.js backend:

Prompt: Build a Node.js Express route GET /api/users/dissatisfied that returns a paginated list of users whose latest satisfaction survey score is below 3. Include filters for department, date range, and score threshold. Return JSON with total_count, dissatisfied_users array, and average_score.

INST: The endpoint name and response fields all carry the word dissatisfied. This keeps the codebase semantically consistent and makes the intent unmistakable to any developer who reads it later.

STU: What about a full application scenario?

INST: Here is a comprehensive application prompt:

Prompt: Build a full-stack web application called DissatisfiedUserTracker using React frontend and Node.js backend. The app should allow admins to view a live feed of dissatisfied users based on NPS scores below 6, drill down into individual feedback, assign follow-up tasks to support agents, and export weekly dissatisfied user reports as CSV.

INST: Every module the AI builds — the feed, the drill-down, the task assignment, the export — is all framed around the concept of the dissatisfied user. The naming becomes consistent throughout the whole application.

STU: And for a Python tool?

INST: Try this one for a classification pipeline:

Prompt: Write a Python script that reads a CSV of customer reviews, uses a pre-trained HuggingFace sentiment model to classify each review, and flags reviews as dissatisfied when sentiment score is below 0.35. Output a filtered CSV of dissatisfied reviews with columns: review_id, text, score, and dissatisfied_flag.

INST: The dissatisfied_flag column name is crucial. It gives the AI a clear binary output to target, which makes the entire classification logic sharper and the output more immediately usable by a data team.

STU: These prompts are really precise. Using dissatisfied as the anchor word makes every output focused.

INST: That is the lesson. In prompt engineering, adjectives like dissatisfied do not just describe — they define the data shape, the logic branch, and the naming conventions across your entire application.