AI Fluency for Small Business – From Awareness to Action By the end of this lesson, students should be able to: Why first wins matter. The first AI application you try will shape whether your adoption continues. A first application that produces obvious, tangible value in the first week builds momentum. One that requires months of setup, fails to deliver, or produces problems builds resistance. Choose your first application deliberately using three criteria. The three criteria for a first AI win. Criterion 1 – Time-heavy today The task should currently consume significant owner or staff time. AI saves time on things that take time. A task that takes five minutes does not produce a visible win. A task that takes four hours per week produces an obvious one. Criterion 2 – Clearly language-intensive Writing, researching, summarizing, explaining. The task should be primarily about language, not physical action, physical judgment, or complex real-time data. Criterion 3 – Output is reviewable You (or a staff member) can evaluate whether the AI output is good before using it. This preserves quality control and builds comfort with AI's strengths and limitations on real tasks. The first-win test. Apply all three: if a task is time-heavy, language-intensive, and its output is reviewable – it is a first-win candidate. The task with the highest combined score across all three criteria is your starting point. Designing the first experiment. A first AI experiment is not a full rollout – it is a test. Run the test for two weeks: Defining success. Success for a first AI application means: Not every first application succeeds. If a first application does not meet these criteria after two weeks, it tells you something useful about that specific use case – not about AI in general. Try the next candidate. A pet grooming salon owner applies the three criteria. Her tasks: appointment reminders (5 minutes/day – not time-heavy enough), grooming instruction intake forms (done in seconds, not language-heavy), after-service care emails (30 minutes per client for premium packages, language-intensive, output is reviewable before sending). After-service care emails score highest on all three criteria. She spends two hours setting up a Claude Project with her service descriptions and common care instructions. After two weeks: time per after-service email drops from 30 to 8 minutes. Client responses are more positive – the emails are clearer and more detailed than her previous drafts. First win. The three-criteria framework works in both directions: it helps you find the right first application AND it helps you avoid bad first applications. If you are considering an AI application and it fails on criterion 3 (output is not reviewable before use), this is a warning. First applications should always include a human review step – both for quality and for learning what AI can actually do in your context. For first AI experiments, set a clear scope boundary: this is a test, not a replacement. Run AI and manual processes in parallel for the first one to two weeks where possible. This gives you a comparison baseline and a backup if the AI output is not usable. First experiments are low-risk only if the review step is protected. Apply the three criteria to five tasks in your business. Score each from 1-3 on each criterion. The task with the highest combined score is your first experiment candidate. Design the experiment: define the task, set up the Claude context, set the two-week test period, and define success criteria in advance (time savings, output quality threshold). You should be able to apply all three criteria to any candidate task, identify your own best first AI application, design a simple two-week test for it, and define success criteria in advance. Log in and enroll to access lesson quizzes.
Lesson 2: Finding Your First AI Win – Where to Start
Lesson Objectives
Lesson Content
Practical Example
Lesser-Known Tip
Safety Notes
Practice Task
Completion Check