"here we connect and innovate"
Learning objectives
By the end of this lesson, you will be able to:
Every lesson so far has been building toward this moment: a simple, repeatable process for actually choosing a tool, instead of just knowing facts about tools. This is the part of the course you’ll come back to again and again — not for this course, but for every future AI tool decision in your career, including ones involving products that don’t exist yet.
Work through these four questions, in order, before you commit to — and especially before you pay for — any AI tool.
What’s your budget?
Most major LLMs offer a genuinely useful free tier — start there before paying for anything. Specialized tools, like transcription or enterprise analytics platforms, often have their own separate free or trial tiers that are worth testing before you commit to a paid plan. Budget isn’t just about affordability; it’s about not locking yourself into a subscription before you know whether the tool actually fits your workflow.
What’s the actual goal?
A one-off task — translate this one document, transcribe this one call — usually calls for a specialized tool used once or occasionally. An ongoing, varied need — I want a daily writing and thinking partner — calls for a general LLM you’ll return to repeatedly. Naming the actual goal precisely, the way you practiced in Lesson 2.1, is often the single most clarifying step in this whole framework.
What’s your industry?
Regulated industries — healthcare, finance, legal, and similar fields — should weigh data privacy and compliance features heavily, not just raw capability. A tool that produces excellent output but handles sensitive client or patient data in a way that violates industry regulations isn’t actually the right tool, no matter how impressive its features are. This question doesn’t apply equally to everyone, but when it applies, it can override every other factor.
Individual or business use?
Solo professionals can usually get by comfortably with consumer-tier tools. Teams processing high volumes of text — support tickets, survey responses, contracts — benefit from purpose-built enterprise NLP platforms that scale in ways a single chatbot subscription won’t. If you’re deciding for a team rather than just yourself, this question often points toward Category 7 (Enterprise NLP and Text Analytics) even when an individual in the same role might reasonably choose a general LLM instead.
Here’s how it actually plays out. Imagine you’re a solo bookkeeper (individual use) who occasionally needs to translate an invoice for an international client (a one-off, narrow goal), on a tight budget (start free), with no specific regulatory requirement tied to translation itself (industry question doesn’t override anything here). Walking through all four questions points you cleanly toward a free-tier translation tool for that specific task — not a paid LLM subscription, and not an enterprise analytics platform.
Now imagine a healthcare clinic’s front-office team (business use) that wants ongoing help drafting patient communication (an ongoing, ideally consistent need), operating under HIPAA and similar regulations (industry question is critical here), with an existing budget for enterprise software (budget allows for a properly vetted, compliant solution). That same framework, applied to a very different situation, correctly steers the decision away from a casual consumer chatbot subscription and toward a vetted, compliance-appropriate enterprise tool — even though the underlying task, drafting communication, sounds similar to plenty of general LLM use cases.
Same four questions, two completely different, both correct, answers. That’s the framework working as intended.
If you work through all four questions and you’re still genuinely unsure, the safe default is this: start with a free-tier LLM for general tasks, then add a specialized tool only once you notice a specific, repeated task that the LLM handles clumsily. This default protects you from overcommitting to tools you don’t yet know you need, while still letting you build toward a proper specialized toolkit as real, repeated needs reveal themselves.
Real-world examples
Analogy
The four-question framework works like a doctor’s intake questionnaire before deciding on treatment. The doctor doesn’t jump straight to a prescription; they ask about your situation, history, and constraints first, because the same symptom can call for different treatments depending on the full picture. In the same way, “I need help writing” isn’t a complete enough question to answer well on its own — budget, goal, industry, and scale of use all shape what the right answer actually is.
Best practices
Common mistakes
Important notes
This framework is deliberately simple by design. It’s meant to be usable in under a minute for routine decisions, and it holds up over time precisely because none of the four questions depend on which specific products exist right now. You’ll be able to apply this same framework to AI tools that haven’t been invented yet.
Lesson summary
The four-question framework — budget, actual goal, industry, and individual versus business use — gives you a fast, repeatable way to choose the right AI tool for any situation. Different combinations of answers correctly point toward different tools, even when the surface-level task looks similar. When you’re still unsure after working through all four questions, start with a free-tier LLM and add specialized tools only once a specific, repeated need becomes clear.
Hands-on exercise
Objective
Apply the full four-question framework to a new, unfamiliar scenario.
Instructions
Read this scenario: A three-person marketing agency wants ongoing help writing client-facing blog posts and social captions. They have a modest monthly software budget, work in the general marketing industry with no special regulatory requirements, and want this to be a consistent, ongoing part of their workflow, shared across all three team members. Walk through all four questions for this scenario and write down your recommended tool category and reasoning.
Expected outcome
A reasoned recommendation — most likely a paid-tier Large Language Model subscription, given the ongoing/varied goal, modest but real budget, no major regulatory constraint, and small-team (not solo, but not large-enterprise) scale of use, possibly combined with a Writing and Grammar Assistant for a final consistency pass across three different writers.
Budget: Modest but real, so a paid tier is reasonable if it clearly earns its cost.
Actual goal: Ongoing and varied (blog posts and social captions differ in tone and length), pointing toward a general LLM rather than a narrow specialist tool.
Industry: General marketing, no special regulatory weight to apply here.
Individual or business: Small team of three, sharing the workflow — worth considering a plan that supports multiple seats or consistent shared style, and potentially adding a grammar assistant to keep tone consistent across three different writers.
Recommendation: A paid-tier LLM (Claude or ChatGPT) as the core drafting tool, paired with Grammarly for a shared final consistency pass.
Reflection Question: Think of a tool decision you made in the past — for yourself or your team — without using a framework like this one. Would working through these four questions have changed your decision? Why or why not?
Key takeaways