Business Analyst

Most business questions begin the same way. You have a spreadsheet export or some app data, and a question such as "why did sales dip last month?" Getting from the file to a clear answer usually means hours of filtering, copying and checking totals. The Business Analyst takes that work off your plate.

Think of it as a good colleague who sits between you and your numbers. It starts from the decision you need to make, then works back to the few figures that matter for it. Like a careful accountant, it checks totals against the source before it trusts them, and it shows its working so you can follow every step.

It is also honest about what it does not know. If data is missing or a sample is small, it tells you rather than presenting a guess as a fact. Like a chef tasting each dish before it leaves the kitchen, it checks its work first. You get the answer, the two or three numbers behind it, one chart that makes the point obvious, and a note on what to look at next.

Think of a librarian at a busy desk. You ask for a book, and the librarian finds it, checks it is the right edition, and writes down where it came from so you can trust it later. The Business Analyst treats your numbers the same way. It finds the figures, checks the edition, and names the source of every number it gives you.

• Analyse spreadsheets, CSV exports and data from your connected apps to answer business questions.
• Define your key numbers, such as revenue, customers or conversion rate, so everyone measures them the same way.
• Build dashboards and recurring reports around those numbers, so you can check them without asking again.
• Compare two periods or two datasets, such as sales before and after a price change.
• Research markets, competitors and customers on the web, and list the sources it used, so you can check each fact yourself.
• Read charts, screenshots and images you share with it, and check the numbers in them against your data, so a picture cannot slip past unchecked.
• Work with your connected apps once you connect them: Google Sheets, Google Analytics, HubSpot, Salesforce, Airtable and Gmail.
• Bring in the rest of the Analyst Team for bigger jobs, such as cleaning data or writing the report.

It works in three panels: the files panel for your spreadsheets, the browser panel for web research, and the artifact panel, where dashboards and reports open so you can share them. It never changes your source data. It asks before it reads from a connected app for the first time in a chat, and it asks before anything hard to undo, such as writing to a connected app.

The Business Analyst has four skills. Each one is a set of steps it follows for one kind of job. You do not need to pick one yourself. Ask in plain words, and it chooses the one that fits.

analyze runs the full Analyst Team pipeline on a data project. It profiles the data, cleans it, works out the statistics, and then builds the dashboard, charts, report and recommendations. Use it when you want the whole picture from a new export, such as "What does this data tell us?"

compare puts two data projects side by side. It shows how their structure, quality and key numbers have shifted, and writes a short report on the difference. Use it to see what changed between Q1 and Q2.

dashboard builds a standalone, interactive page with KPI cards, charts and a data table. You open it in your browser. Use it for a weekly view of sales by region that the team can check on their own.

report writes a readable analysis report: a summary, an overview of the data, the findings with their charts, and recommendations. Use it when you need a write-up for your manager.

Here is a common job, from question to answer.

1. You drop a sales export into the files panel and ask: "Which products are driving our growth this quarter?"
2. The Business Analyst starts from the decision behind the question. Here, that means finding which products deserve more attention.
3. It cleans a copy of the file. It checks totals against the source, fixes dates and numbers written in odd formats, and removes exact duplicate rows. Your original file stays exactly as it was.
4. It works out revenue and units for each product this quarter, and compares them with the quarter before.
5. Before it writes the answer, it lists its assumptions, such as how refunds were treated and which months it counted. You can correct any of them.
6. If the question needs a deeper test, it brings in the Data Scientist for that step.
7. It writes the answer: the top products by share of growth, the two or three numbers behind them, and one bar chart that makes the point clear.
8. You get a short note of what it could not see, for example, products with no price in one month. The answer, the chart and the note open in the artifact panel, ready to share.

At the end, you have three things: the answer with its two or three numbers, one chart you can drop into a slide, and a list of assumptions and gaps. Your original file has not changed, and nothing has been sent anywhere.

Call it when you have a question and a pile of files, and no afternoon to spend joining them up. Which products carried the quarter. Whether the new price changed what people buy. How many customers came back after their first order. Ask it when the question is about the business rather than about the maths, because it starts from what you want to know and only then reaches for the numbers.

Ask it also for the things you repeat. A weekly sales summary, a monthly metric sheet, a comparison of this month against last. It remembers your key metrics, how you define them and how often you want them, so the second report takes one line to ask for. When a question needs the whole team, say cleaning a messy export, a proper statistical test and a dashboard, it runs the Upfyn Analyst Team's pipeline and brings each specialist in for their step, then hands you one answer with its sources and charts.