Introducing Querri

It’s an exciting time to be building software. Large Language Models – and Generative AI more broadly – are opening doors in a way that no innovation has since the birth of the Internet. The pace of innovation is so rapid that it’s hard to keep track of just what is possible and what is just hype.

At Querri, we’re on a mission to help people do more with their data. We are focused on practical applications of these new technologies to help anyone work with and get answers from their data so they can make more informed decisions, work more efficiently, and ultimately get back to solving the problems they’ve set out to solve.

What is Querri and why are we building it?

Querri is a tool I’ve wanted to have for pretty much my entire career. I’ve always been fascinated by the promise and potential of data analysis to answer questions, predict the future, and generally make better decisions. I’ve had the privilege of working with phenomenal data scientists on interesting projects and have done my fair bit of work with data, first in building applications with Databases, then moving into data analysis with R and Python, building scripts to move data around and manipulate it in different ways, and integrating with and buying a variety of products that aim to make working with data easier.

The problem? It’s still always too hard. It’s not the fault of any of the software but rather the fact that working with most types of data is inherently messy. It comes in a lot of shapes and sizes, and there are so many ways that data can be entered and stored. This leads to variance in how people track anything from dates to addresses to product information to just about anything else. Oh, and it’s not all in one place, ever. So, different software systems track different parts of the data you’d like to use. One has customer information, another sales data, another tracks marketing information, and yet another financial data. And when I say “another,” I mean dozens or even hundreds of different systems in a larger organization and a whole lot of Excel spreadsheets mixed in as well to track bits and pieces of different parts of data that would be interesting to you.

What does Querri do?

In a nutshell, Querri helps you clean, move, and analyze your data. This means a lot of different things depending on who you are and what you do, but let me give you a few examples:


  • You have customer data with different columns holding address information. Sometimes you have a FULL ADDRESS with city and state in there, sometimes you have a STATE column with two-letter state codes, and sometimes you have a full word for STATE. You want to run a marketing campaign for all your customers in New York. How do you filter that? With Querri, it’s as simple as telling Querri something like “create a new column called STATE CODE, which sets the US state based on the address fields in each row.
  • You have customer survey data where people filled out their full name, but you want to load data into email software that has first and last name fields. Just ask Querri to “split the FULL NAME column into FIRST and LAST NAME.”
That can look something like this:



  • You have data in an Excel spreadsheet and want to load it into one of your SaaS tools. Let’s say an import into HubSpot. While HubSpot has a nice data importer, it still expects rows and columns of your data to follow particular formats and be in a particular “shape.” If your spreadsheet doesn’t match that format, you may need to do a variety of transformations first. In Querri, you can just work through those one at a time by joining multiple data sheets together from your single Excel file or even across several files and just asking for changes to be made until it looks right and your import works. Many SaaS tools don’t have a very strong importer, so getting this all exactly right can be even more important with other tools.
  • You’re migrating your product catalog from one e-commerce system to another. You can export all the data from System 1 as CSV files, but the column names and date formats don’t match up. The new system also has more available fields to track things like color and category. You can have Querri transform dates and columns.


  • You’ve loaded up some data and want to understand what’s wrong with it. This is often called “exploratory data analysis” or EDA. You can ask Querri things like “tell me about the ADDRESS column” or even “tell me about this dataset.” It will then give you some analysis of things like missing values, formatting errors, or other types of issues you might want to address. You can also have Querri draw charts and graphs or explain numerical distributions. You can even ask for text analysis on a column, and Querri can extract common keywords. These are all good places to start to understand what transformations your data may need for whatever your use case is.
  • You have a specific question you want your data to answer. In many cases, this might be the final step of your journey. You can start with exploratory data analysis and then iterate through all the cleaning and transformations mentioned above. When your data looks “right,” you can start asking questions about relationships between the different columns in your data. For example, you might ask about seasonal changes in sales, what factors most influence a particular customer segment to buy a product, or how labor hours and manufacturing costs affect output.

Here's a little preview of some of the analysis you can do in Querri:


How does Querri work?

You start with a Project. A “project” is just a container for the work you’re going to do. It’s probably something practical, like trying to answer a specific set of questions, clean up some data that you want to move from one place to another, or share some specific view of your data with some colleagues.

Whatever the project is, you just create one and then load as many data sources as you need. You can then start typing in the box at the bottom, or use the microphone icon if you prefer to talk, and give instructions or ask questions.

Behind the scenes, we use Large Language Models and a rich set of open-source data tools. Our first step is to pre-process your data and use a large language model to understand how your question can be turned into a plan of action. This is similar to a software engineer or data scientist breaking down your request into smaller, actionable steps. We then match that plan against a library of specific tools that are used to accomplish a variety of tasks. This can be things like joining multiple data sources, looping over each row to do data extraction, or triggering statistical analysis or visualization tools.

As we run through these various steps, you’ll see updates on the progress, followed by plain text outputs and graphs. You can ask Querri follow-up questions on the outputs or continue forward in your journey to gather additional insights.

Sometimes the answers or transformations Querri gives you aren’t exactly what you’re looking for. In those cases, simply hit the “undo” button at the bottom left corner and try again. This is a useful pattern to keep in mind when working with any AI-powered tool, since while they can save us a lot of time, they won't always handle our requests exactly as we ask. For that matter, a human may not either, so just think of the undo button as saying, "That's not quite right; let's try that again."

Pro tip: You can access your recent prompts without retyping by tapping the up arrow when in the prompt box.

Another nice little hidden feature is that you can see a list of the available "tools" that Querri uses to answer your questions by clicking on the tool selector by the prompt bar. If you select one of these tools, you'll see the icon change, and Querri will now try to use this specific tool to answer your request.

When you’re at a point where you’re ready to share insights or move your data elsewhere, look for the download and export buttons. You can export your data to an XLS spreadsheet in its cleaned and transformed format, or you can download your graphs for inclusion in presentations and reports.

Ready to get started?

We’re excited to see what you’re able to accomplish with Querri. If you want to just give things a try, you can sign up for our Early Access Program, or if you’d like to talk through your use case and explore ways we can work together, please reach out.

Similar posts