Mockitt AI · Start with a question

AI Data Analysis Tool

Start with Excel or CSV.
Understand what changed—and what to do next.

Give your sales, channel, or operations spreadsheet a clear question to answer. Use the Mockitt AI desktop app to clean up, summarize, and compare local files. Define your metrics and calculations, then organize the results into charts or a report so every conclusion can be traced back to the data.

Excel and CSV tasks Metrics and definitions Charts and reports
Dark blue data analysis illustration with a spreadsheet, trend chart, and analysis cards

An AI data analysis tool helps turn a business question into a concrete analysis task. First, establish what the data can answer. Then decide whether to summarize it, compare groups, or investigate unusual records. The Mockitt AI desktop app starts with file-based tasks, so you can keep asking questions, checking results, and making changes around an Excel file, CSV file, or related materials.

Questions from your everyday work

Choose a question before you analyze the spreadsheet

01

Sales changes: Where is growth coming from?

Compare revenue and order counts by month, region, or product. Separate overall growth from changes in individual segments.

Before you begin Use consistent dates, product definitions, and amounts. Specify how to handle returns, canceled orders, and taxes.
02

Channel performance: What happens after a visit?

Start by comparing visits and leads, then consider conversions and costs to evaluate each channel. Traffic volume alone does not tell the whole story.

Before you begin Include the source channel, reporting period, and relevant metrics. Without cost and conversion data, do not infer return on investment.
03

Expense breakdown: Which costs changed the most?

Summarize expenses by department, category, or month to identify where spending changed most.

Before you begin Specify whether amounts include tax, how to treat one-time expenses, and how to identify duplicate records. Standardize category names first.
04

Project operations: Where are delays and backlogs building up?

Organize tasks by owner, status, and time period to see what remains unfinished and how quickly work is moving.

Before you begin Include task IDs and dates. A missing completion date must not be treated as “completed today.”
A table with blanks and duplicate rows is cleaned and standardized, with unusual records kept separately
Standardize fields and categories, and keep missing or unusual records separate. Cleaning the data should not change what it means.

First, get the data on the same terms

A well-organized spreadsheet gives the analysis a reliable starting point

Before analyzing anything, ask the app to identify the worksheets, fields, and records it has read, and confirm what each row represents. An order table and a line-item table cannot use the same method to count orders. Likewise, adding monthly totals to daily records will double-count the data.

  • Missing values: Distinguish an absent value from zero A blank amount may mean the value has not been entered; zero is a specific value. Identify where values are missing and which records are affected before deciding whether to exclude, retain, or fill them in.
  • Duplicates: Decide what uniquely identifies a record Two rows with the same amount are not necessarily duplicates. Choose an order ID, line-item ID, or a combination such as time and source based on what the data represents.
  • Dates and categories: Make comparisons consistent Standardize date formats, month ranges, units, and category names. When comparing this month with last month, remember that a month in progress is not directly equivalent to a complete month.

Start with one specific task

Where is growth coming from? Start with a channel spreadsheet

Walk through a channel comparison using six rows of sample data. Switch between the original data, month-over-month results, and findings and actions. You can check every number in the tables.

Your task for AI

Read the channel data file I specify. First, list the fields, record count, missing values, and duplicate records. Then summarize visits by month and channel. Do not change the original data. If you cannot read the file or perform a calculation, explain the specific reason before proceeding.

Refine the result Keep the original data and save the summary separately. Do not automatically treat blank values as zero; list the affected records first.

Sample channel visits Sample data
MonthChannelVisits
Last monthWebsite1,240
This monthWebsite1,480
Last monthPaid campaigns860
This monthPaid campaigns792
Last monthPartners310
This monthPartners356
Website1,480
Paid campaigns792
Partners356

This month's visits · Matches the original data

You can copy or save these sample materials. They illustrate the task instructions and result structure; they are not real customer data.

Match the method to the question

Summaries, comparisons, trends, and anomalies answer different questions

You do not need to start by asking for a “comprehensive analysis report.” Choose a suitable method and output first. That makes the calculation scope clearer and your follow-up questions more specific.

What you want to knowHow to analyze itWhat the result should include
Totals and compositionGroup and summarize by month, channel, or category.Values for each group, the total, units, and whether the groups cover every record.
Where values increased or decreasedCompare two consistently defined periods, calculating both the absolute change and the percentage change.The original values for both periods, the difference, and the denominator. Explain separately when the baseline is zero.
Whether a change is sustainedExamine consecutive periods to distinguish short-term fluctuations from a longer-term trend.The full timeline, any missing periods, and changes in definitions. Do not select only the interval that supports a preferred conclusion.
Which records need investigationCheck extreme values, missing data, duplicates, and inconsistent combinations against business rules.The location of each record, the rule it triggered, and what needs to be verified. Do not simply delete unusual records.

For ratio metrics, recalculate the numerator and denominator. For example, the overall conversion rate is total conversions ÷ total visits; you generally cannot average the individual channel rates. Overall month-over-month change should also be calculated from the two period totals.

An original spreadsheet connects to three comparison bar charts and a report; the second group declines, showing why channels should be analyzed separately
Look separately at the total, differences between channels, and missing information. A single growth chart should not determine the conclusion.

Turn the numbers into a useful discussion

A 9.0% increase overall does not mean every channel grew

In the example above, visits rose from 2,410 to 2,628, a net increase of 218. Website visits increased by 240, paid campaign visits fell by 68, and partner visits increased by 46. Those three changes still add up to 218. The website's increase is larger than the overall net increase because the decline in paid campaigns offset some of that growth. The numbers are consistent.

  • What you can confirm: How the growth breaks down Show the values for both periods, the difference, and the percentage change for each channel. Total visits increased by about 9.0%, while paid campaign visits declined by about 7.9%. Both findings can be true at the same time.
  • What you cannot infer yet: What caused the change A change in visits alone does not establish ad quality, content effectiveness, or a change in revenue. You need to examine campaign activity, changes in traffic sources, conversions, and reporting definitions.
  • Your next question: What data is still missing?“Add costs, lead counts, and conversion counts for each channel. Define the denominator for conversion rate, then calculate conversion rate and cost per lead. If the definitions are not consistent, list the differences before ranking performance.”
Start analyzing a spreadsheet ↗

Download the app and bring your data

Complete your first analysis in four steps

Define the input and the result you need, then refine the work based on the output.

01

Prepare the file and explain the data

Choose an Excel or CSV file. Identify the worksheet, time range, and what each row represents. Keep the original file and save the results separately.

02

Inspect the data before calculating

Ask for the fields, record count, missing values, and duplicates. Confirm the file and definitions are correct before grouping or comparing data.

03

Break the question into answerable steps

Summarize first, compare next, and then interpret. For example, total each channel's visits for both months before calculating the change. Ask separately for the results table, chart, and summary.

04

Check the results and follow up

Spot-check original rows, totals, and ratios, and confirm the chart matches the table. If information is missing, add the data instead of asking AI to fill in a business conclusion.

Explore the desktop app guide ↗

More to know before you begin

AI data analysis FAQs

How is this different from a dedicated BI platform?

This workflow starts with a file-based task in the desktop app and helps you define and carry out an analysis. Direct database connections, ongoing synchronization, enterprise metric systems, and online dashboards are separate capabilities. Support for local files does not imply support for those features.

How can I check the growth rate in the channel example?

Add the three channels for each month: 2,410 last month and 2,628 this month, a difference of 218. Then calculate 218 ÷ 2,410, which is about 9.0%. Each channel's change uses its own previous-month value as the denominator; do not simply average the percentages. The page uses sample data, and you can recalculate every value.

What if an Excel or CSV file cannot be read?

First, confirm that the file opens on your computer, identify the worksheet containing the data, and check the character encoding. Provide the file location, column headers, and field definitions. If the task cannot read or calculate the data, resolve that issue before generating a report. Do not substitute estimated numbers.

Can I use AI analysis results to make a decision right away?

First, verify the sources, handling of missing values, aggregation definitions, and calculations. Correlated changes do not establish causation. Decisions involving costs, returns, or forecasts also require the relevant data and business judgment.

Can I start an analysis report before I have the data?

You can outline the questions, required fields, and report structure first. Do not present invented data as real findings. If you need an example, label the sample data clearly and replace it with verified data before making an actual decision.

Can I analyze data with AI without programming experience?

You can start with a question in everyday language, but you still need to explain the fields and business definitions. Begin with a small, clear task—such as a monthly summary or a comparison of two channels—and gradually add conditions. Understanding the metrics matters more than using advanced terminology.

How should I choose a chart for AI-generated results?

Start with a bar chart for category comparisons or a line chart for changes over consecutive time periods. Keep a table when readers need exact values. State the metric, units, and time period in the title. A chart does not replace the original data or the calculation definitions.

Start your next analysis with a good question.

Open the Mockitt AI desktop app and describe the work you need to do.

Download Mockitt AI ↗