Golden Analytics: Data Deserves Better

Thirty years of business intelligence has failed the people it was created for and incumbents cannot fix it—the moment to build something new has finally arrived.

Every generation of analytics software has promised to put data in the hands of the people who need it. And every generation has found a new way to fall short. The first generation locked insight within static reports and behind IT approvals. The second generation tried to dissolve that bottleneck by democratizing data access to a meaningful degree. But the complexity that was removed from one place congregated elsewhere: features accumulated, dashboards multiplied, and "self-service" began requiring the same depth of expertise it was supposed to replace. The second generation had become a more sophisticated version of the first. Today, senior analysts at most enterprise companies still begin their Monday the same way they would have in 2005: pulling data from one place, cleaning it in another, building charts in a third, and assembling a narrative in a fourth. 

The business intelligence industry promised to fix this. Understanding why it hasn't, and why that is finally about to change, is the thesis behind our investment in Golden Analytics.

The Structural Problem That Incumbents Cannot Fix

Legacy business intelligence (BI) was built around a core assumption: that a human would always perform the critical translation work that makes a query return a meaningful answer rather than a technically correct but contextually wrong one. When you bolt a language model onto that architecture, which many major BI vendors have rushed to do, you inherit all of its assumptions. AI can generate SQL, but it queries data models it didn't build and can't precisely interpret. The incumbents aren't failing for lack of resources or intention—they're failing because the architecture that made them successful imposes a ceiling on what AI can do within it. The only way to build an analytics platform that AI can reason through reliably is to design it around AI from the ground up.

Introducing Golden Analytics

Golden Analytics is an AI-native analytics workspace where AI operates directly in the canvas as an embedded capability that works on data in place. It connects to existing data sources, surfaces insights automatically, and gives every user—regardless of technical background—a path from raw data to finished analysis in a single session.

With an architecture centered around AI, Golden makes several things possible. The product can serve the full range of users that enterprise analytics has historically failed to serve simultaneously. For the analyst, the blank canvas disappears. The hours currently consumed by pipeline preparation, data cleaning, and deliverable formatting can be re-allocated to critical thinking, pattern recognition, and unveiling insights. For the business leader, the dependency chain breaks. When a VP of Sales can connect data directly, ask a question in plain language, and receive an answer she can interrogate, she is no longer dependent on pre-configured charts created to answer pre-determined questions. She can make better decisions faster and with more confidence. Golden calls this the Slider of Autonomy, the ability to set—at any point—how much of the process the AI handles and how much the user directs. The dial belongs to the user and it moves in both directions. 

In addition to the already impressive and thoughtful list of product features, the real prize is a  unified workflow. One of the most persistent sources of friction in enterprise analytics is the handoff: raw data lives in one tool, cleaning happens in another, visualization in a third, narrative assembly in a fourth. Each handoff introduces latency, version divergence, and the possibility of error. Golden, in contrast, owns the full journey, from connected data source to finished, shareable story, without fragmentation or sacrificing advanced analytics for polished visualization. This effectively closes the gap between data and decision-making at every level of an organization. More so, rather than starting with data and trying to tease apart what questions it can answer, Golden inverts the process, enabling data analysts to investigate their data as if following the scientific method. Analysts can start with a question and work through their data research, augmenting their process with external data as needed, until they find an answer. In other words, Golden makes data analytics proactive, rather than reactive.

Golden Analytics was founded by François Ajenstat, one of the most credentialed product leaders in enterprise data: a 30-year veteran of the BI industry who came up through Cognos and Microsoft before helping build Tableau from startup to Salesforce acquisition and subsequently serving as CPO at Amplitude.We’ve also been fortunate to work closely with François through his role as an NEA Venture Advisor. Now, he is building Golden because he has watched two generations of analytics tools make the same promise and fall short in the same ways, and because he believes the technology now exists to finally execute. François knows this market better than almost anyone, where it has fallen short, why it has fallen short, and what it would take to fix it, and has assembled a world-class team of engineers from Snowflake, Tableau, Microsoft, Atlan, and Grammarly who share his conviction. Collectively, this team is among the most passionate, creative, and fast-moving we've seen in data.

A Practical Example

To demonstrate Golden in practice, we ran a quick analysis of YC's W26 batch entirely within the platform. Here's what it looks like to take a question and turn it into an answer.

Within seconds of uploading, Golden surfaces questions worth exploring and quick insights, unprompted.

We wanted to understand the thematic split across the batch. Using the company description column, Golden automatically classifies the entire dataset in two clicks.

From there, building a chart from the classification results is equally effortless.

What would have taken an analyst hours of cleaning, classifying, and charting across multiple tools took minutes and a single canvas.

NEA Leads Golden's Seed Round

The case for a new generation of analytics tooling is not new. What is new is that the technology to build it has finally arrived.

For decades, the constraint on analytics software was that it could retrieve data but not interpret it. The human had to bridge the space between raw data and meaning at every step. Generative AI removes that constraint. Now, software can participate in the analytical process itself: not just execute instructions, but explore, suggest, and surface connections across a dataset. That changes what is possible for the analyst. It also, crucially, changes what is possible for every business leader who has ever wanted a direct relationship with their data and been told, politely, to wait for the dashboard.  Vibe coding dashboards with Claude or Codex hints at what is possible, but analytics and BI demand precision, validation, and structured data compatibility that general-purpose AI wasn't built to own. The enterprise data market deserves an autonomous interface purpose-built for it.

At NEA, we have been investing in and helping to build data and enterprise infrastructure for decades, including leading investments into Tableau, MongoDB, Databricks, DataRobot, Elastic, Plaid, Cloudflare and more. In our view, the next phase of business intelligence requires an AI-native foundation, a workflow that spans from raw data to finished insight without fragmentation, and a product designed for every person in an organization who has a question.

Every generation of BI has had its defining platform. We believe Golden will define the third. We are proud to announce our seed investment in Golden and to partner with François and his team to create the future of data visualization.

***

Golden is selecting their first cohort of design partners now. If you lead an analytics team, run a business function that depends on data, or want to be part of building what comes next, reach out at goldenanalytics.com.