Your next BI tool shouldn't be a BI tool

Through all the trends and changes we’ve encountered in the world of data, Business Intelligence (BI) has been there through it all. We keep collecting data, and people want to use that data to inform themselves. It’s a powerful concept!

But how this is done has been malleable over the years. In 2005, it meant “delivering printed, paginated lists of numbers to business teams.” In 2015, it meant “visualization-based data discovery and dashboards.” Along the way, new requirements have entered the puzzle. Cloud-based. Embedded. Automated insight delivery.

At Hex, we’ve been a bit afraid of being designated “BI,” because it’s kind of a loaded term. It makes people think of drag-and-drop dashboard building tools and pixel-perfect chart controls. Over the years we've asked ourselves: "Is that really what we make?" We’ve always been motivated by making it easier for anyone to get insights from data, but have remained unsure if the current paradigm for it was right.

And then, about a year ago, something funny happened. Customers started telling us that Hex was their BI tool of choice.

Wait, Hex is a BI tool now?

Recently, we did a survey asking customers “what do you use for BI?” And a lot of people told us: “Hex!” We saw the same thing looking at peer review sites: Hex is a BI tool.

We were, admittedly, a little surprised when we saw this. Doesn’t BI mean complex, pivot-based visual systems you have to go to a week-long training to learn to use? Is AI making those things obsolete? What about data science? What about context?

And then we came to a realization. All of these customers who were using Hex for BI, weren’t merely using Hex for BI. They were redefining it, and doing more in the process.

What even is BI?

The problem that BI solves is: people want to know things.

More specifically, business people. “Business people” is an odd term because it represents an incredibly broad, diverse group of people across many functions and roles. When you’re delivering software for “business people,” you can’t assume any level of technical training or competency (note: some of them are, in fact, technical. But across such a broad audience, you have to think of the least common denominator, which is why BI is so hard).

And so we turn to dashboards, charts, reports, visual exploration systems. Historically, they have been the fastest, easiest way for us to deliver answers to people without a hard technical barrier to entry.

Of course, the objection that data teams, the guardians of an organizations data, will immediately interject is: what if they get something wrong? We can’t count on every business person to know how to ask a data question, to know when a query might be wrong, to know whether the customer data lives in prod_app_cust_rec or prod_app_cust_detail or whether a mean or a median is the right way to measure something.

And so BI tools include complex permissions-control systems, certified semantic models, folder structures for content and version-control systems for tracking changes.

Lastly, when someone — be it a business person or a seasoned data analyst —finds an insight that is truly interesting, how do they canonize it? How do they verify, validate, and share, and ensure that everyone who wants that answer knows where to find it?

And so BI tools have app-hosting platforms and directories where users can find answers and collaboration tools so people can iterate on a shared report.

So if you take those needs:

• Business users want to know things: they want self-serve access to facts from their data.
• Data teams want to guard against risk: they want to certify that the answers stakeholders use in business decisions are accurate.
• And everyone needs a way to share the insights: as dashboards, reports, and interactive apps that the whole organization can reference.

Is Hex a BI tool? You’re goddamn right it is.

More than BI

Today, though, BI is merely a part of the equation. We’re seeing leading companies building data practices that go beyond the above definition:

• Chime focused on making chatbots work with data by building a trusted foundation, investing in context and evals to measure whether agents consistently meet their accuracy goals.
• At LangChain, AI workflows were getting chaotic, with technical users using coding agents and non-technical users getting insights from BI. The LangChain team unified the experience by building a single system that could leverage context from everywhere in the organization. Now business teams and data teams work together, and—most importantly—reference each others’ work.
• EliseAI identified a process—sales QBRs—that was taking their field team away from customers, and automated it with an app that allowed reps to turn data into a customer-facing deck in 5 minutes.

All of these use cases have elements of BI: self-serve, governance, reporting. But if you approached the scenario in reverse, looking for a tool to automate QBRs or build a context layer for chatbots, you wouldn’t go shopping for a dashboard tool. That’s how the world is changing - new technology is forcing us to reevaluate these requirements. The impact of data can be bigger than ever. For teams that embrace it, there’s an opportunity to take BI from a cost-center that their companies begrudgingly pay for and turn it into a value driver for the whole organization.

We're delighted that customers are using Hex in so many different ways. Call it BI, or agentic analytics, or a context layer, or a secure, generative app platform, or even the best data experience for the leading, most data-driven companies.

To us, it’s just Hex.