AI Meets Data
Jump to section
Data is everywhere — AI is your new analyst
Every day you work with data. CRM exports, Google Analytics, sales reports, spreadsheets from the finance department. The problem is not that you lack data — the problem is that you cannot quickly get answers from it. Until now, you needed an analyst, SQL or Python. Now you just describe what you want to know, and AI answers.
AI tools like ChatGPT and Claude have improved dramatically over the past two years when it comes to working with data. They understand tables, generate SQL, visualize trends and write analyses — all based on natural language descriptions. You need zero programming knowledge.
This is not a course about data science or machine learning. It is a practical guide for people who work with data and want to be faster and more effective with AI. Analyst, marketer, product manager or CEO — if you look at tables and reports, this course is for you.
What AI can do with data
AI today handles a wide range of data tasks that previously required specialized knowledge or tools:
- Analyze CSV, Excel and other tabular data uploaded directly into the chat
- Generate SQL queries from natural language descriptions
- Clean data — deduplication, format correction, filling missing values
- Create charts and visualizations from your data
- Write formulas for Excel and Google Sheets
- Interpret A/B test results and statistical analyses
- Automate recurring reports and analyses
First steps: ask AI about your data
Start with a small fictional or suitable public dataset. Describe its columns, units, time period and your question. Check missing values and data quality before interpreting results. Real company exports require approved processing.
Prompt for first analysis: 'Here are our monthly sales for the past year: [paste data]. What is the overall trend? Which months are outliers? What would you recommend tracking next?'
Uploading CSV to ChatGPT and Claude
ChatGPT (with Code Interpreter) and Claude both allow uploading files directly into the conversation. Upload a CSV or Excel file and AI will automatically read it, understand the structure and let you ask questions about the data. This is by far the most effective way to work with data in AI.
When you upload a file, AI will typically say something like: 'I see a table with 1,247 rows and 8 columns. The columns are: date, product, region, revenue, cost...' And then it asks what you want to find out. Just answer in natural language.
# Example conversation after uploading a CSV
User: [uploads sales_2025.csv]
AI: I see a CSV file with 1,247 rows and 8 columns:
- date (transaction date)
- product (product name)
- region (area)
- revenue (revenue in USD)
- cost (costs)
- quantity (number of units)
- channel (sales channel)
- salesperson (sales rep)
What would you like to find out?
User: Which product has the highest margin and how
does it differ by region?
AI: [runs analysis, creates table and chart]How to formulate data questions
The quality of answers directly depends on the quality of the question. Good data questions are specific, measurable and contextual. Bad questions are vague and unspecific.
- Bad: 'Analyze this data' — AI does not know what you care about
- Good: 'Which 3 products generate the most revenue and how does their share change month over month?'
- Bad: 'Is the data okay?' — too vague
- Good: 'Are there duplicate records in the data? Does any column have more than 5% missing values?'
- Bad: 'Make a chart' — what chart? From what?
- Good: 'Create a bar chart comparing monthly revenue by region for Q1-Q4 2025'
AI tools for working with data
Each of the major AI tools has its strengths for data work:
- ChatGPT — file analysis and computation in supported modes; limits depend on the account
- Claude — file and visualization work through available tools; context depends on model and product
- Gemini in Google Sheets and Copilot in Excel — availability depends on licensing, organization and specific feature
- Coding assistants can draft SQL or Python; verify every suggestion
Data security
Use fictional or suitable public data without personal or confidential information for exercises. Real company data requires an approved tool and workflow. Replacing names with codes alone does not guarantee anonymity or permission to share.
What you will learn in this course
This course will walk you through the entire process of working with data using AI. We start with SQL queries, move to spreadsheets, visualizations, data cleaning, analysis patterns and finish with reporting automation. Every lesson is practical — you will work with real data and real scenarios.
Practice this scenario only with fictional or suitable public material without personal or confidential information. Do not upload actual messages, internal documents or client data; real use requires an approved tool and workflow. Create a fictional table or choose a suitable public dataset without personal information. In a supported AI tool, ask: 1) What trends do you see? 2) Where are the outliers? 3) What next step would you investigate? Check totals on a small sample and distinguish observations from hypotheses. Do not use a real CRM, payroll or client export.
Hint
If you do not have your own data on hand, use any public dataset — for example sales data from Kaggle or government statistics.
- AI tools today can analyze data, generate SQL, visualize and report
- Just upload a CSV or paste data and ask in natural language
- Quality of answers depends on quality of questions — be specific and measurable
- Each AI tool has its strengths: ChatGPT for ad-hoc analysis, Claude for reasoning, Gemini for the Google ecosystem
- Use fictional or suitable public data without personal or confidential information for exercises. Real company data requires an approved tool and workflow. Replacing names with codes alone does not guarantee anonymity or permission to share.