What AI Can Do
Capabilities & Limits
Mapping AI to Business Tasks
AI is not magic — it is a pattern-matching engine trained on enormous amounts of text. Understanding what it can and cannot do is the single most important skill for using it well in business.
What AI Excels At
Text generation — Drafting emails, reports, summaries, marketing copy, and internal memos. AI produces first drafts in seconds that would take a human 30-60 minutes. The output needs editing, but the time savings is real.
Analysis and pattern recognition — Given structured data (spreadsheets, tables, JSON), AI can spot trends, compare periods, identify outliers, and describe patterns in plain English. It does not "see" the data the way a human analyst does — it converts rows into text and reasons about them.
Summarization — Condensing 20-page reports into 5-bullet summaries. Extracting key findings from meeting transcripts. Turning dense financial filings into executive-friendly language.
Classification — Sorting customer feedback into categories (praise, complaint, feature request). Tagging support tickets by urgency. Labeling invoices by department.
Extraction — Pulling specific fields from unstructured text: names, dates, dollar amounts, product mentions, action items from meeting notes.
What AI Cannot Do
Make decisions — AI can list options, weigh pros and cons, and recommend. But it does not understand your business context, your risk appetite, or your stakeholders. Decisions require judgment that AI does not have.
Access real-time data — Unless connected to a live data source, AI only knows what you paste into the prompt. It cannot check today's stock price, query your CRM, or look up a customer record on its own.
Guarantee accuracy — AI can produce confident-sounding answers that are wrong. This is called hallucination. Always verify numbers, dates, and factual claims against the source data.
Replace domain expertise — AI is a general-purpose tool. It does not know your industry's regulations, your company's internal politics, or the nuances of your market. Use it to accelerate work, not replace thinking.
Setting Realistic Expectations
The right mental model: AI is a very fast, very knowledgeable junior analyst. It can do research, draft documents, and crunch numbers — but it needs clear instructions and quality checks.
| Task | AI Role | Human Role |
|---|---|---|
| Quarterly report | Draft the narrative | Verify numbers, add context |
| Competitor analysis | Summarize public info | Validate, add insider knowledge |
| Customer feedback | Categorize and summarize | Decide what to act on |
| Strategy memo | Structure options and tradeoffs | Make the actual decision |
What You Will Build
In this course, you will use AI to analyze real business data, summarize reports, draft decision memos, evaluate AI tools, and design your own AI workflow. No coding required — just clear prompts and critical thinking.
Glossary
| Term | Meaning |
|---|---|
| Prompt | The instruction you give to an AI model |
| Hallucination | When AI generates confident but incorrect information |
| Classification | Sorting items into predefined categories |
| Extraction | Pulling specific data points from unstructured text |
| Context window | The maximum amount of text an AI can process at once |
This is chapter 1 of AI for Business Decisions.
Get the full hands-on course — free during early access. Build the complete system. Your projects become your portfolio.
View course details