30 Best Holistics Features Shipped in 2025

Roughly ten years ago, on the night of November 9, 2016, my two co-founders, Vincent, Thanh, and I were sitting frozen in a small Airbnb in San Francisco.

A few minutes earlier, an email had landed in our inboxes and completely redrawn the trajectory of our lives. We kept opening it, rereading it, closing it, opening it again, as if that could somehow make the words come out differently. Nobody had anything useful to say. The gooey silence clung to us and only let go when Vincent looked up from his phone and said the 2016 election results were coming in. So we talked about politics, and nothing about the email. The next morning, we packed our bags and got on a plane home.

That email was a rejection from Y Combinator. I’m leaving it here, in case you’re curious.

We came to San Francisco earlier that year with unwavering confidence.

We shipped fast. We knew our customers well. We were really passionate. And at the time, we already had real logos on the wall, namely Traveloka and Grab, the leading tech companies in Asia. (If folks at Traveloka and Grab are reading this: know that we’re forever grateful). We were convinced, in the way young founders often are, that making it was inevitable.

Yet, the email was right about our acquisition strategy, and especially, about our differentiation at the time.

Earlier this month, my team and I celebrated 10 years of Holistics.

It’s been a long ride. And sometimes, in a whimsical way (not regret), I do wonder what Holistics would look like if we’d been accepted—if we’d stayed in San Francisco, built there, raised there, lived inside that particular timeline. But most of our energy goes to a more useful question: what makes Holistics different?

Staying bootstrapped has let us focus obsessively on customer outcomes. It’s also let us stay opinionated about what we believe self-serve analytics should be, and now, what AI-native analytics should be. We’ve spent years building what we believe are the foundations, even when they weren’t fashionable: an intelligent semantic layer, an analytics-as-code workflow, and composable metrics, foundations that let us build what no one else can.

10 years go by in a blink. We’re older now, and more confident in how Holistics differentiates. You might see that in this blog, highlighting the best features we shipped in 2025. And if you don’t see it clearly enough, that’s on us, and we’re going to keep making the differentiation obvious, not just by saying it, but by shipping what we believe only Holistics can offer.

In the meantime, thank you for being with us in 2025. Thank you for every question, every piece of feedback, and every recommendation. Really excited to work with you in 2026.

Building AI with Semantic Intelligence

We launched Holistics AI earlier this year. Instead of writing code, dragging and dropping fields, or waiting for an analyst, you can now simply chat with your data using natural language to explore insights and build visualizations.

Built on top of our architecture, which includes a governed semantic layer and composable query language, Holistics AI is designed to help teams explore data through conversation while staying grounded in trusted, curated datasets. Since then, we've been adding more and more capabilities to our AI.

Here are some of the highlights:

🤖 Build charts and get insights with Holistics AI

Ask questions in natural language to explore metrics, run complex analyses, or get step‑by‑step instructions. Our Agentic AI automatically determines your intent and selects the appropriate knowledge source to deliver the right answers.

You can also inspect how ​Ask AI​ interprets your question, the analysis it performs, and the underlying formulas for each metric and dimension.

For more information, check out our docs here.

🤖 Organizational-level and Programmable Context for AI

Organization-level context allows you to define a global knowledge base and instructions for the AI. It ensures that AI responses follow your company’s business logic, terminology, and data standards.

Read more here.

🤖 Context Control for Smarter AI Answers

As you get more comfortable working with data, your questions for AI naturally get more ambitious, for example: Do accounts that adopt Feature X within 30 days have higher ACV and lower churn than those that don’t?

That question spans multiple datasets, Product, Sales, and Finance, so to answer that well, AI needs a little hint, and we’ve made it easier to tell AI where exactly it should look.

🤖 Holistics MCP Server

Holistics provides an MCP server that lets your own AI agents access and perform BI tasks on your Holistics workspace.

For more details, read our doc here.

Making It Easier to Find and Trust Data

As analytics spreads beyond a small group of power users, organization starts to matter as much as insight. In 2025, we introduced better tools to structure analytics content inside a Holistics workspace and gave everyone else clearer signals about where the source of truth lives. The result is a workspace that’s easier to navigate, easier to govern, and easier to trust, even as more people create, share, and explore data.

🏷️ Tagging System: Organize and Discover Analytics Content with Ease

As your workspace fills up with dashboards and reports, “just search for it” stops working. Important content gets buried, duplicates multiply, and business users fall back to whatever link they saved last, whether it’s correct or not.

Tagging is now available to fix that. It gives data teams a simple way to organize analytics content, and gives everyone else a faster path to the trusted version when they need an answer.

With this feature, you can:

• Quickly filter and find exactly what you need without navigating complex folder structures.
• Instantly understand content purpose, status, and relevance without reading lengthy descriptions
• Systematically categorize content with consistent, meaningful tags

For more information, visit our public doc here.

🗂️ Content Endorsement

Content Endorsement is another tool to help you systematically organize analytics content, ensure data quality, and provide clear trust signals that make reliable insights more accessible.

📁 Content Archiving

As analytics workspaces grow, outdated content clutters your environment and confuses users. Content Archiving lets you declutter by hiding outdated content without deleting it.

Making It Easier to Explore Data and Build Reports

In 2025, we continued to push hard on a simple goal: make exploration feel less like a workflow and more like a reflex. When a chart looks weird, you should be able to inspect the raw rows behind it. When you want to slice performance by Country, then Product, then Customer Type, you shouldn’t need 50 near-identical charts. When you need to tweak a filter, you shouldn’t have to jump into Edit mode and lose your place.

These updates bring more of the exploration experience into the dashboard itself, while also making metric creation and reuse less of a copy-paste sport.

Here are some of the highlights:

🔍 View Underlying Data

Ever seen a spike in your chart and wondered, "What's really going on here?" or "Why does this number look the way it does?"

Our upcoming View Underlying Data feature makes answering these questions easier. With just a click, you can instantly see what makes up any value at its most granular level.

🔄 Dynamic Dimension Breakdown

You’re trying to understand how your Sales perform across Countries. Then by Product. Then by Customer Type. So you build chart… after chart… after chart. 50 dimensions. 50 charts. And your once-clean dashboard is now bursting with clones, one for every angle.

Not anymore. With our Dynamic Dimension, you can now break down dimensions in visualizations with zero setup required.

⚡Widget-level Filter

Previously, users had to click "Edit" or "Explore" and jump into separate modals just to change basic filters like date ranges or categories, breaking their analytical flow and forcing them to work with one chart at a time.

Now you can apply filters to individual widgets, directly from the dashboard interface, maintaining full dashboard context while you explore. This is our first step toward bringing the complete exploration experience into the dashboard.

🧮 Calculation Builder: Create Powerful Metrics Without Code

AQL lets analysts build complex metrics much more easily, but learning it takes time. And we know not everyone has that luxury. That’s why we’re building Calculation Builder to let you tap into the full flexibility of AQL, through a point-and-click experience anyone can use.

🪄Turn Ad-Hoc Fields into Reusable Metrics

You create the perfect calculated field during analysis, and suddenly everyone wants to use it. But instead of easy reuse, you're stuck recreating the same logic across multiple dashboards, copying formulas, and watching inconsistent variations multiply across your organization.

Field Promotion eliminates this friction by transforming your proven ad-hoc calculations into permanent, reusable metrics that live in your Semantic Layer.

Launching Self-service Embedded Analytics Capabilities

"Embedding was built into your product from the start, whereas for most companies, it’s an afterthought”, said Vikram, CTO at Datacubed Health.

Too often, embedded analytics is something teams bolt on later, constrained by tools that were never designed to live inside a product. In 2025, we doubled down on the opposite approach. Our AI-powered, self-service analytics are designed to be embedded as a native part of the product experience, so your customers can explore data, ask questions, and get answers directly inside your application.

Here are some of the highlights:

🧱 Embed Portal: Add a Mini-BI Inside Your Product

Your customers expect more than static charts: they expect to explore, interact, and build. But delivering full BI capabilities inside your app usually means months of work: authentication flows, permissioning logic, version management, dashboard builders, and visualization libraries.

Basic dashboard embedding isn’t enough either. There’s no self-service report building or exploration. Users get frustrated the moment they want to filter data, build a custom view, or just want to ask anything slightly off-script.

That’s why we built Embed Portal.

Embed Portal lets developers add a full-featured BI experience directly into their product. Instead of hardcoding charts, you give users their own space to explore data, create dashboards, and self-serve, all without ever leaving your app.

🔗 Dynamic Schema & Data Sources

When implementing embedded analytics, you often have customers with their own separate databases or schemas. But traditionally, you're stuck with static connections that can't adapt to different users—forcing complex workarounds or separate dashboard instances for each customer.

That's why we built Dynamic Schema & Data Sources to let your embedded dashboards automatically connect to the appropriate database based on which user is viewing it.

🤖 Let your customers ask AI data questions inside your app

You already have a great product; now you can make it even stickier and unlock a new revenue stream by delighting your customers with a built-in, personalized AI assistant.

Earlier this year, we launched Embed Portal, which lets our customers add a mini BI experience directly into their products and give their users full drag-and-drop report-building capabilities.

We then took it a step further: you can give customers a simple, ChatGPT-style interface where they type a question and get going. Our Ask AI automatically determines the intent and selects the appropriate knowledge source to deliver the right answers.

For more details, check out our doc here.

Expanding Canvas Dashboard Capabilities

Canvas dashboards started as a way to break free from rigid grid layouts. In 2025, we pushed that idea further.

This year’s updates focus on control and reuse. You can theme dashboards consistently and beautifully, reuse components across reports, and evolve designs without rebuilding from scratch. With flexible layouts, dashboards defined as code, and a dedicated development environment, Canvas lets teams treat dashboards like real product surfaces. The result is data presentations that scale, stay consistent, and fit naturally into how your organization works.

Here are some of the highlights:

🗂️ Tabs for Canvas Dashboard

Dashboard navigation shouldn't be a maze. When analysts create complex reporting solutions, end users often struggle with lengthy, scrolling dashboards that bury critical insights. That's exactly why we built Tabs for Canvas Dashboard.

Tabs let you combine multiple, connected dashboards into a single interface with easy-to-use tab navigation. This introduces better content organization, more intuitive navigation, and improved performance by loading content per tab.

🎨 ​Default Dashboard Templates

With Holistics Canvas Dashboard, you can design fully on-brand dashboards—but until now, you couldn’t reuse those designs. Every new dashboard meant redoing colors, fonts, spacing, and other style choices, making consistency hard to maintain.

Default Dashboard Template fixes that. Set your brand standards once, and every new dashboard follows them automatically.

🎨 Layout Assist in Canvas Dashboard

Canvas Dashboard introduces a flexible, freeform layout. But that freedom sometimes comes with a tax: more time on cosmetic fixes, overlap fixes, manual positioning, spacing adjustments, etc. Our Layout Assist solves this. With this feature, you can:

• Prioritize Non-overlap: Now you can set your blocks to avoid overlapping. To switch to overlay mode, just press and hold a block for two seconds.
• Trim whitespace to remove extra space around your elements.
• Shift multiple blocks at once for faster adjustments.
• Auto-expand the canvas as you add new blocks.

🧩 Reusable Block Library

Dashboard builders want speed and consistency, but they're stuck recreating the same charts, logos, and layouts from scratch every time. Reusable assets exist somewhere in your organization, but they're locked in code that casual users can't access.

That's why we built the Reusable Component Library to make every visualization a building block for future dashboards, accessible through a visual interface anyone can use.

Improving Developer Experience & DataOps

In 2025, we invested heavily in the parts of analytics most users never see but everyone depends on. This work Holistics faster to build with, safer to change, and easier to operate at scale. We expanded AQL with new functions, improved core capabilities like advanced period-over-period analysis, and made it possible to mix native SQL directly into AQL when you need full database power. We also strengthened the surrounding DataOps workflows, from real-time validation and in-editor references to first-class CI/CD, code search & replace, and Git-based change management.

The goal is simple: less friction for developers, more confidence in production. Here are some of the highlights:

🔧 New AQL Functions

Building complex analytics often means wrestling with limited function libraries, forcing you to write verbose workarounds or export data for processing elsewhere. Simple tasks like finding correlations, extracting text patterns, or getting values from related rows become unnecessarily complex.

That's why we expanded AQL with 30+ new functions across aggregation, text manipulation, time intelligence, and window operations—giving you the tools to handle sophisticated data transformations natively.

With these functions, you can…

• Calculate correlations instantly with corr(table, field1, field2) for statistical analysis
• Get related values efficiently using min_by and max_by to fetch data from rows with extreme values
• Concatenate grouped data with string_agg(expression, separator) for clean text aggregation
• Extract text patterns using regexp_extract, regexp_replace, and other regex functions
• Create statistical buckets with ntile(n) and percent_rank() for advanced analytics

Function categories include: Aggregation, Time Intelligence, Text Manipulation, Window Functions, and AI Functions (Snowflake & Databricks)

See all new functions and examples in our community post.

🔧 SQL Passthrough for AQL 4.0

You can now leverage native SQL functions directly in your dataset queries, handling JSON unnesting and any database-specific functionality not available in standard AQL, all without leaving your current workflow.

With this feature, you can:

• Extract fields from JSON directly in your dataset queries
• Apply database-specific functions (Postgres, BigQuery, MySQL, etc.)
• Maintain type safety while using native SQL capabilities
• Extend AQL's power without breaking your workflow or going back to the model.

For more details, check out this doc.

🔧 New and Improved Holistics CLI

Jumping between interfaces, waiting for app updates, and manually checking model validity, all just to start your analytics project? I hate it too!

With our improved CLI, you can manage your entire Holistics project from your terminal, validate dbt models with confidence, and seamlessly integrate with CI/CD workflows.

🔧 CSV & Google Sheets Import

Imagine needing quick insights from a customer survey. The usual delays with traditional BI include waiting for data engineers, navigating complex ETL setups, and dealing with access issues, resulting in hours or even days of downtime just for temporary data.

That's why we've launched a seamless CSV and Google Sheets import that loads your data directly into your existing warehouse for instant analysis.

170+ UX Improvements & Bug Fixes

Reflecting the meticulous effort of our designers and engineers, we’ve polished every corner of the app - from standardizing UI components, improving page layout, fixing tiny bugs, and refining interactions - to make your daily experience smoother and more reliable.

This year, we did a lot of unglamorous work on purpose.

We shipped over 170 UX improvements and bug fixes across Holistics, tightening up the product in all the places you feel every day but rarely call out. Designers and engineers standardized UI components so the same action behaves the same way everywhere. We improved page layouts to reduce visual clutter and make common workflows easier to scan. We refined interactions so the app feels more predictable, especially when you’re moving fast between exploring, editing, and sharing.