What's New

Customize Ask AI Experience lets Ask AI be a native part of your product.

Previously, every embedded portal showed the same default Holistics Ask AI welcome page. Now you can customize:

• Logo - replace the Holistics logo with your own, with granular control over left panel, toolbar, and chat page icons
• Intro headline and body - set your own welcome message with full Markdown support
• Prompt placeholder - guide users with contextual example prompts
• Assistant name - rename "Ask AI" to match your product (shown in the sidebar and top-left corner)

Customization is defined at the embed definition level using AML, giving you version control and per-portal flexibility:

EmbedPortal ecommerce_portal {
  ai: {
    assistant_name: 'Ask Hooli'
    icon: 'https://cdn.example.com/logo.png'
    intro_headline: 'Hi, what would you like to know?'
    intro_body: @md
      ## 🤔 Use natural language to ask questions
      ## 💬 Ask follow-up questions to refine answers
    ;;
    prompt_placeholder: 'Ask me anything, eg: Top 5 products by revenue'
  }
}

Visual styling (fonts, colors, borders) will be handled in a future Theming update.

Reusable Library is coming to Reporting — browse and insert shared components directly in the dashboard editor, no need to switch to Development.

Previously, using library blocks required switching to Development, which added friction and completely blocked explorers. With this update, you can insert existing library components right where you work. Blocks built by one team member are instantly available to everyone.

Learn more: Reusable Components

Timeline​

Expected by end of March 2026.

In complex data models, users can create metric breakdowns that are technically valid but analytically wrong.

These queries run successfully and return plausible numbers, but the results are meaningless. This silent failure undermines trust and hurts self-service.

Now you can control which dimensions can be used with each metric using the filter_direction property on relationships.

Why this matters​

• Prevents silent data correctness issues before they reach end users
• Turns semantic correctness into a default behavior, not a best practice
• Keeps self-service powerful and safe, even in multi-fact (galaxy) schemas
• Reduces the need for manual user education or tribal knowledge

📚 Learn more​

• Controlling which dimensions can be used with a metric: Detailed walkthrough with examples
• Controlling filter and grouping paths: Reference documentation for filter_direction

As-Code Color Palettes lets you define and manage color palettes directly in AML, alongside your data models.

Previously, color palettes were managed through Admin Settings with no version control or easy way to reuse them consistently. Now you can:

• Define categorical, sequential, and diverging palettes in AML and reference them at the project, theme, or chart level — changes propagate automatically everywhere a palette is referenced
• Map specific data values to specific colors (e.g., "Vietnam" always yellow, "US" always red) and have every chart that displays those values automatically use the correct color, no per-chart configuration needed

Set workspace-wide defaults so new dashboards inherit your brand colors automatically.

Timeline​

This feature is currently in development and expected to be available in July 2026.

If you've used custom skills in AI assistants (like Claude, ChatGPT, or Copilot), you'll get the idea. AI Skills brings the same concept to Holistics AI: instead of cramming everything into a single context.aml file, Analysts can define multiple skill definitions right in AML, each scoped to specific teams with its own instructions, data references, and access controls.

With AI Skills, Analysts can:

• Create domain-specific skills: a finance_skill and an ecommerce_skill can coexist, each with its own instructions, terminology, and dataset references
• Control access per user: use user attribute expressions to restrict skills by department, role, or any attribute (e.g., disabled: user.department != 'Finance')
• Set activation modes: skills can auto-activate when relevant, require manual invocation via slash commands, or stay hidden as background context
• Reference contexts dynamically: inject datasets, contexts, or files with preloaded (${}) or lazy (@Context:, @Dataset:, @File:) references
• Import from existing docs: upload Claude or markdown skill files and convert them to AML definitions

End-users interact with skills through slash commands, UI toggles, or simply by chatting. The AI auto-selects relevant skills by default.

Timeline​

AI Skills is currently in development. We'll share availability details here when they're confirmed.

Filter Auto-apply is coming to Dashboard preferences. When enabled, your dashboard will re-run automatically whenever a filter or control value changes — no Apply button required.

Today, every filter change requires a manual Apply click to update the dashboard. This is easy to miss, especially for non-technical viewers who expect the dashboard to respond immediately. With auto-apply, you can remove that friction for dashboards where it makes sense.

How it will work​

1. Go to Dashboard preferences > General
2. Check Auto-apply filters and controls, then save
3. Publish your changes (Development mode) or reload the page (Reporting mode)

Note: Auto-apply triggers a query on every filter change. For dashboards with many charts or intensive queries, keep manual Apply to avoid unnecessary query costs.

Learn more: Enable filter auto-apply

Percentile calculations are coming to the Calculation Builder!

Create continuous and discrete percentiles directly in the GUI. Choose from common percentiles (P10, P25, P50, P75, P90, P95, P99) or enter a custom value.

Two types are supported:

• Continuous percentile - interpolates between adjacent values
• Discrete percentile - returns the closest actual value from the data

Example use cases​

• P10 customer lifetime value - identify your lowest-value segment to investigate churn risk
• P25 vs P75 salary range - show the interquartile range for compensation benchmarking
• P75 order value - understand where the bulk of revenue sits for pricing decisions
• P90 response time - track performance while filtering out outliers
• P95 session duration - identify power users by top 5% engagement time
• P99 query execution time - catch extreme outliers for capacity planning

Stay tuned - more Calculation Builder capabilities are on the way!

Arithmetic calculations (+, -, ×, ÷) are now available in the Calculation Builder!

You can now perform basic math operations on your fields directly in the GUI:

• Add (+) - e.g., Revenue + Tax
• Subtract (-) - e.g., Revenue - Cost = Gross Profit
• Multiply (×) - e.g., Price × Quantity = Total Order Value
• Divide (÷) - e.g., Converted Users ÷ Total Users = Conversion Rate

Like other GUI calculations, you can compose these with existing metrics, or convert to AQL for further customization.

Control who can use AI in your organization. This helps you roll out AI gradually and enable it for selected teams first.

Learn more: Holistics AI | User Access

Widget exports now reflect ad-hoc interactions you've applied on screen - so what you export is exactly what you see.

Previously, exporting a chart after applying ad-hoc changes would revert to the original unmodified data. To get an accurate export, users had to switch to Explore mode and manually reapply every interaction. Now the export captures the current on-screen state directly - no workaround needed.

This applies to all supported ad-hoc interactions on the widget:

• Block-level filters
• Date Drill
• Breakdown dimension changes

Both data exports (CSV/Excel) and image exports (PNG/JPEG/SVG/PDF) are covered. No UI changes - same export action, correct output.