Pleom Thinks Business Intelligence Should Just Work, and It Might Be Right

The Macro: Business Intelligence Is a Full-Time Job That Nobody Signed Up For

I have watched companies spend six figures on Tableau licenses only to have three people in the entire organization who actually know how to use it. The rest of the company submits ticket requests to the “data team” and waits. Sometimes days. Sometimes weeks. By the time the dashboard gets built, the question it was supposed to answer is no longer relevant.

This is the dirty secret of business intelligence. The tools are powerful, but the learning curve is so steep that most employees never touch them. Looker requires SQL. Tableau requires its own visual query language that takes months to internalize. Power BI locks you into the Microsoft ecosystem and still expects you to know DAX formulas. Mode, Metabase, Sisense. They all assume you have a dedicated analyst sitting between the data and the people who actually need answers.

The result is that most business decisions get made on gut instinct dressed up as data-driven decision making. Someone pulls a number from a spreadsheet, drops it into a slide, and calls it analytics. The BI tool sits in the corner, underutilized and overpriced.

The market for business intelligence software is worth roughly $30 billion. A significant chunk of that spend produces dashboards that get looked at once, bookmarked, and forgotten. The opportunity is not to build another dashboard tool. It is to eliminate the gap between “I have a question about my business” and “here is the answer.”

That is what agentic BI promises. Instead of building queries, you ask questions. Instead of configuring dashboards, the system generates visualizations automatically. Instead of hiring an analyst, you connect your data and start talking.

The Micro: An 18-Year-Old ML Grad Who Already Had a YC Exit

Royce Arockiasamy is not your typical founder. He finished a Bachelor’s and Master’s in Machine Learning at Georgia Tech by the age of 18. Before Pleom, he co-founded Bits to Atoms, which went through Y Combinator’s Summer 2024 batch. So when he showed up at YC again with Pleom in Summer 2025, he was already a known quantity.

Pleom’s pitch is simple: connect your data sources in seconds and get automatic insights with zero learning curve. No SQL. No dashboard configuration. No waiting for the data team to get back to you. You plug in your databases, spreadsheets, or SaaS tools, and Pleom’s AI layer does the rest. It generates visualizations, runs workflows, applies machine learning models, and surfaces analytics without requiring you to understand what is happening under the hood.

The “agentic” part is key. This is not just a chatbot sitting on top of a database. The system actively monitors your data, identifies patterns, and proactively surfaces insights you did not ask for. Think of it as the difference between a search engine and a research assistant. A search engine waits for your query. A research assistant brings you things you need to know before you realize you need to know them.

The product supports voice interaction, screen sharing, and camera sharing for context. You can ask a question out loud, share your screen to reference something you are looking at, and get an instant answer. The interface is designed to feel conversational rather than analytical.

Where Pleom sits in the competitive landscape is interesting. ThoughtSpot pioneered the “search-driven analytics” category and raised over $600 million doing it. Narrator AI took a different approach, modeling everything as a single activity stream. Athena Intelligence is going after the investment research angle. But most of these still require meaningful setup and some technical fluency to get real value.

Pleom is betting that the right answer is radical simplicity. Zero learning curve is a bold claim, but if anyone has the technical chops to deliver on it, it is a kid who finished two ML degrees before he could legally drink.

The Verdict

I like the thesis. The BI market is bloated with expensive tools that most employees cannot use, and the companies paying for them know it. An agentic approach that actually delivers on “plug in and get answers” would be transformative for small and mid-size businesses that cannot afford dedicated data teams.

The risk is execution. “Zero learning curve” is one of those promises that sounds great in a demo and falls apart when someone connects a messy production database with inconsistent naming conventions and missing values. The quality of the insights is only as good as the data, and most company data is terrible.

Thirty days, I want to see real customer testimonials from non-technical users who connected their own data without hand-holding. Sixty days, I want retention numbers. Do people come back after the initial “wow” factor wears off, or do they revert to spreadsheets? Ninety days, the question is whether Pleom can handle the complexity of real enterprise data environments without losing the simplicity that makes it compelling. If it can, this is a category winner. If it cannot, it becomes another demo that looked great on stage.