How losing our Tableau dashboard builder led us to build NousViz

In 2024, we made the decision for StatsDrone to use Tableau as our data viz of choice to visualize StatsDrone client data outside of the app. We believed that by building bespoke data viz for our clients that we could gain the experience and insights to eventually rebuild it back inside of our stats app.

When we hired our analytics and BI specialist to build these dashboards, they didn’t have experience pulling data via API and storing into database tools like AWS or BigQuery. What was required was another person on our team who had to build these data pipelines from scratch.

As we rolled out our analytics services to our customers, we felt the dashboards weren’t hitting the target as we had hoped. They were not using the dashboards daily and engagement with the dashboards varied significantly between customers. Everyone wanted these BI tools (Business Intelligence) but they needed more of our help.

Every customer that we had to onboard, we would have to do with a yearly subscription so if they were non committed to the service then we had to eat the cost. Commercial and workflow friction reinforced our feeling that we needed a more flexible approach. In many situations, a custom dashboard was being built for a customer and after a week of building, we would be presented with the dashboard in which we would not get what we were asking for. Sometimes the dashboards were overly complex and as management, we knew what we wanted but we didn’t have the time ourselves to do the work.

Slack plugin by NousViz showing sentiment analysis with our team.

Simply put, building and maintaining dashboards is a time consuming and costly endeavour. Management didn’t have the time needed to build the dashboards themselves and would only be able to review the dashboards at the end of the build. As a team, we were very capable of achieving perfection in our analytics but it simply required more stakeholders and this is the problem that most managers have. They have limited time to dedicate to projects and prefer to delegate than to do the work.

In the fall of 2025, our Tableau expert left our company in which we were without any data viz solutions for our clients. We had spent the last year and a half writing about data viz with Tableau so now we had to remove the mention in our marketing and social media posts. This felt like a painful step backwards.

AI dramatically accelerated our dashboard prototyping process

It was at this point that I took data viz into my own hands and started the work in using AI for building analytics. Personally, I had been working on this since April 2025 using ChatGPT and felt at the time it was a start but the work had a long way to go.

By October 2025, I was able to replace all our Tableau work and libraries we made in under 10 hours using AI. Don’t get me wrong, Tableau is still a fantastic tool but losing our dashboard builder was a big pain and I didn’t have the time get back into it. I was mainly using a mix of Claude Code and some Lovable for modelling of dashboards and the frontend code to render. The only thing missing was connecting databases.

The one thing I realized is that while we used Tableau, some of our clients were on other tools like Power BI and Looker Studio (now Google Data Studio) so people really didn’t want to change their workflows. They would only consider it if could make life better for the company.

Our solutions aimed to make it easier for even the managers to quickly build the dashboards they knew they wanted.

Custom Data Viz with AI

By late 2025 and early 2026, I had been building numerous models of all our data analytics needs for StatsDrone as a company as well as with some customers in working with their data.

StatsDrone, Google Search Console, Google Analytics, Firehose and Cloudflare.

My goal was to get customer data viz made for the collection of data sets for StatsDrone. One of the main driving forces of doing our advanced analytics had to do with automating the shaving detection work we do for affiliates.

I did also want to include other data sets as our customers have asked for this before. SEO affiliates wanted to blend StatsDrone data with Google Search Console, Google Analytics, Cloudflare, Ahrefs, Firehose (also by Ahrefs) and many more. Our paid media affiliates wanted us to give them a central dashboard that included their Facebook and Google ads and link tracking tools like Volume, Keitaro, RedTrack, Pretty Links and many more.

LLM traffic with the NousViz plugins or Google Analytics, thanks to Jabez Reuben for the increase in LLM traffic.

Then I started to use more of the tools we use as a company at StatsDrone such as Slack, Jira, Intercom, Calendly, Read.ai, Quickbooks, Stripe and many more.

Then I built all the other tools I dreamed about having that I didn’t think I’d be able to do in a scalable way in Tableau and this included annotations and alerts. You can do these things in Tableau but it requires time and now I can do this all myself. It is just a few hours to build out these features and the new idea was to make this scalable and repeatable so others could have these features on demand.

Armed with my API keys, I was able to build a series of dashboards and built my first MCP to export all of this data and have it retrievable with Claude Cowork.

Pattern recognition and the birth of NousViz

While building out all of these models of dashboards, I felt I was getting better and faster after every iteration. At NEXT.io NYC 2026, we drove to New York City and took our laptops with us to the conference so we could show people exactly what we were working on.

The tl;dr is that everyone was excited at what we had built but the product had no name and it was all custom for every situation.

I’ll add a preface that I made a commitment to focus on data viz while most people were going deeper on agents with the explosion of OpenClaw. I did experiment with OpenClaw and the open source framework was another inspiration for thinking about how the new product could exist.

On the drive back home from NYC, I felt I was starting to connect the dots of the problem and potential solution. The problem is very real but not everyone could articulate it well.

On the Friday night, I decided to rebuild the model in a brand new framework with the intention that everything could be repeatable and sharable. I have plenty of experience in data and working with open source tools like WordPress and also building custom CMS for affiliate sites. That Friday after NEXT NYC, I spent 5 hours on prompt planning and research before I made the first build in my VS Code app. That was a very long prompt but by the time I went to sleep, I was able to get the first model actually working. That is, built an app as a host and a plugin as the data tool that pulls data via API, stores in a database like Postgres and ClickHouse, then added data viz.

For the rest of the weekend, I kept building and iterating. I added the annotations and alerts that I wanted and proceeded to build plugins for a lot of the data sets I’ve done before. I had even built the website and a pitch deck plus a product roadmap of what we were building and why it needed to exist. I even built a site using NousViz for a custom CMS and launched complaints.wiki basically in a day.

Before sharing how we settled into the NousViz name, will share more discoveries of what we are building.

NousViz as a CMS

Since we needed a website for the app, I remember the things that frustrated me about the current StatsDrone website. If I wanted anything changed on it, I had to go through our tech team and it was a slow and tedious process. Simply put, I couldn’t do everything myself in the way I was doing it myself with our own analytics.

When you think about CMS, most people don’t associate these with data viz but the two go hand in hand.

CMS stands for Content Management System and the core concepts are almost the same: create data, store it and publish and visualize it. When it comes to a CMS, we feel like we are dealing with unstructured data but it is both structured and unstructured.

Most people might not think publishing content as visualization but the core concepts are for me identical. You’re taking data in a database and making it pretty, readable and understandable. Text rendering requires CSS and HTML and you’re choosing the colour, size, fonts, line spacing and the space it will occupy. You’re also optionally adding functionality to it too.

Why NousViz?

NousViz wins AffPapa Conference Award 2026 for Affiliate Intelligence of the Year.

It was difficult finding a unique name that had never been used before as core keywords are simply not available. Those domains themselves were very expensive and even if you could buy a good domain, you’d likely not have the core URLs or usernames in social media sites like Github, YouTube, LinkedIn, X and more.

Being based in Montreal, we thought about French in some of the name and we knew we wanted the name to be somewhat self explanatory and not some weird obscure brand. When we settled on NousViz, nobody had used it anywhere so we could get the domain and all the social media channels.

'Nous' in French means ‘we’ and in Greek it also means ‘mind’ and the viz hopefully is understood as visualization. Also we liked the word ‘new’ but ‘new’ + almost any keyword was not available. When you pronounce the brand NousViz, it sounds like ‘new viz’ for the French part of the ‘nous’ word.