We are pleased to announce several quick-win improvements, directly addressing feedback received from users. 

1. Enhanced dimension table filtering and sorting: 

You can now filter and sort the dataset value column in dimensions table. This feature significantly simplifies the management of unassigned dimensions, providing a more intuitive, Excel-like experience. 

2. Clearer version and scenario identification in the Manual Input Panel:

We've implemented a clearer display of the active 'Version' within the Manual Input Panel. When working with non-default scenarios, both the 'Version' and 'Scenario' will be clearly displayed. This ensures you always know which data context you are editing, minimizing potential errors and enhancing data integrity.

3. On-demand AI variable description generation:

You can now generate AI-powered variable descriptions at any time. This provides greater flexibility and allows you to efficiently document your variables as needed. Learn more about this feature here.

What’s new?

We're excited to announce a significant enhancement to our AI Forecasting feature, giving you even greater control and confidence in your projections. Now, you can preview the impact of AI-driven forecasting before applying any changes, allowing you to see exactly what the projected values will be.

Why This Matters

Forecasting is critical for informed decision-making, and AI can be a powerful tool in this process. However, understanding the precise impact of AI-generated forecasts is essential. This new preview capability empowers you to explore different scenarios and evaluate the potential outcomes before committing to changes, leading to more informed and strategic decisions.

Key Benefits

• Increased confidence:  Make forecasting decisions with a clear understanding of the projected outcomes.
• Improved accuracy: Refine your forecasting numbers by adjusting granularity and based on real-time previews, leading to more accurate projections.
• Enhanced control:  Maintain full control over your forecasting process, even when leveraging the power of AI.

What's Next?

We're continually working to enhance our AI Forecasting capabilities. Future updates will include:

• Integration of new forecasting methods, including deterministic and ML-driven models
• Visualization and selection of multiple scenarios for your forecast values
• More granular control over AI forecasting parameters, allowing for even greater customization

• Pages (Versions, Dimensions, Datasets, Models, Spaces, FX Rates, Activity, Integrations)
• Specific Resources (Variables, Spaces, Models, Datasets, Dimensions)

1. Press Cmd + K / Ctrl + K to open the command palette.
2. Type the name of the resource or page you want to access.
3. Use arrow keys to navigate and press Enter to open.
4. You can also click the new search icon in the top of the sidebar to open it.

• The palette suggests “Ask Abacum Academy,” “Abacum Academy,” and “Contact Support” for quick help.
• Your recently visited resources/pages appear under the “Recent” section for fast access.

What’s new?

We're thrilled to announce Abacum's groundbreaking AI Formula Assistant! Now, building complex formulas for your key metrics is as simple as a conversation. With just a click, you can generate formulas for common calculations, and refine or create entirely new logic simply by describing what you need in natural language.

Why does this matter?

For years, we've been accustomed to the complexities of formula syntax in tools like Excel. But now, Abacum offers a smarter, more intuitive approach. Leveraging the power of AI, we're ensuring in a new era of software interaction, moving beyond buttons and actions to the ease and flexibility of natural language prompts. Abacum's AI Formula Assistant is the first step towards a completely code-free experience for managing your actuals and forecasts, ensuring they work precisely as you intend.

How does it work?

The AI Formula Assistant is seamlessly integrated into both actuals and Forecast for variables without existing formulas:

• Actuals: Clicking the AI button automatically generates a formula based on the variable name. This is especially helpful for commonly used financial formulas and leverages your existing variables for a truly click-free experience. If the generated formula isn't quite right, simply describe the calculation you need, and the system will provide new formula suggestions.
• Forecast: When working with an empty formula, clicking the AI button opens a prompt where you can describe the forecasting logic you want to apply.

After generating a formula, the AI Formula Assistant provides a clear explanation of the underlying logic, ensuring transparency and giving you insight into how the system interpreted your request.

Key points to consider

• AI is a powerful tool, but it's still learning. Always review the suggested formulas to ensure they accurately reflect your specific business requirements.
• The AI Formula Assistant is a continuously evolving model. While this initial version doesn't yet support filters within formulas, you can easily add them afterward, just as you normally would.
• Rest assured that no confidential data is transmitted outside of your system or used for model training.

What’s next?

We're committed to continuously improving the AI Formula Assistant. Upcoming enhancements include:

• More accurate and refined responses for currently supported variables.
• Expanded support for more complex logic, including filtering by dimensions and/or time.
• The ability to refine existing formulas using natural language.

Great news for those who integrate their Microsoft Business Central data with Abacum! Our latest update enables you to import custom reports directly into Abacum, so that you can retrieve custom fields and data not available through the standard integration. This will unlock the full potential of your Business Central data, driving more informed financial decisions!

How it Works

1. Develop a custom API. You need to create a custom API to expose the desired data (via a certified Microsoft consultant).
2. Create dataset. Select the ‘MS Business Central Custom Report’ connection when creating your dataset.
3. Input relevant fields. Import the relevant fields that can be extracted from your custom API endpoint.

new We are excited to announce that our native integration with Rippling is now live! Effortlessly connect your Rippling data, including headcount, salaries, and department costs, for smarter workforce planning.

How does it work?

To set up a Rippling connection, simply go to the Integrations section in Abacum and click Connect on the ‘Rippling’ integration card. You will be redirected to Rippling to complete the connection, and in just a few minutes, you will be able to create your first dataset! Learn more about this new connection here.

What’s new?

Scenarios are finally here! You can now create sub-versions of a version to explore different outcomes and tweak assumptions—no need to duplicate entire versions anymore.

Why does this matter?

Until now, running what-if analyses meant duplicating versions, which created disconnected models. Any updates to the original version had to be manually copied over. Scenarios change that. They let you build variations directly within a version, so you can easily test ideas like doubling your marketing budget or adjusting revenue growth assumptions—all without impacting the original version. It’s a smarter, more organized way to explore possibilities and collaborate with your team.

How does it work?

• Create scenarios: From the Versions page or Model Editor, you can create scenarios like Bull (optimistic), Bear (pessimistic), or Base (neutral).
• Edit values: Adjust manual inputs for individual scenarios. Once you change a value, it becomes independent from the default scenario.
• Compare outcomes: Use the Table Builder and version datasets to add scenarios for easy comparisons.

Key points to consider

• Scenarios share the same structure—variables, formulas, and dimensions—as their parent version.
• You can only detach manual inputs from the default scenario—formulas remain linked to it.
• Once a value is manually changed in a scenario, that row detaches from the default and won’t auto-update.
• Resetting is simple: remove the specific input to reconnect it to the default.
• On tables, for active budget and forecast versions, only the default scenario can be selected. To view non-default scenarios, use the specific version instead.

What’s next?

• Scenarios in KPI Cards (as Estimates and Comparisons).
• Scenario support in space version for version dataset blocks.
• Upload scenario manual inputs through version files.

Learn more in the Academy:

• Quick guide
• What's a scenario?
• Scenario manual inputs
• Scenarios in Tables & Charts
• Scenarios in version datasets

Scenarios will gradually be released in the upcoming days. If you want more information on the feature, please reach out to our Customer Success team.  

We've updated the Model Builder sidebar to better align with our platform’s UI, making navigation more intuitive. Plus, we’re introducing AI-powered variable descriptions to streamline your workflow.

New: AI-generated variable descriptions

When you create a new variable, an AI-generated description is automatically provided, offering a quick explanation based on the variable name. These descriptions can be edited manually in the variable info section on the right sidebar.

Learn more about variable creation here.

Key updates

• Variable descriptions are now available for all variables in your organization.
• AI-powered descriptions are automatically generated for new variables.
• Added icons to the left of the Model Builder sidebar for improved navigation.
• Relocated actions into the new sidebar for a more streamlined experience.

What’s next?

• We will be adapting the dependencies style to align with datasets for a consistent experience.
• Improved functionality and updated styling for dimensions are on the way!

We have made significant updates to our dimensions section with the purpose of making dimensions easier to use, ensure consistency across the full platform and provide greater visibility into any dependencies and errors.

🚀 New features

• Mapping unassigned items from the table– You can now map unassigned items directly from the dimensions table—no need to download files. A filtered view helps quickly find and assign missing items while providing current mapped items context.
• Redesigned interface: The dimensions page now has an updated look and feel, aligning with datasets and other platform tables for a more consistent user experience.
• Activity log: A new activity log captures all updates, errors, and warnings related to dimension file uploads, making it easier to track changes and troubleshoot issues.
• Row height adjustment: As in model builder or datasets, users now can adjust the row height of their dimensions for better readability.
• Dimension level dependencies insights: Expanded dimension-level dependency details provide clearer insights into where and how dimensions are used across the platform.
• Colorize items: Coming from makers project, you can now apply colorized chips at the dimension level, improving visual distinction and organization.

🔧 Improved functionalities

• Improved dimensions file upload:We’ve improved the dimensions file upload process to address these pain points by providing more detailed feedback and better troubleshooting tools:
  • The system now validates the entire file, identifying all errors in a single pass.
  • Users will receive clear, detailed descriptions of each issue, including links to the specific model or version causing the conflict.
  • These errors will also be stored in the dimensions activity log so that the user can access them at any point.
• Reordering Items: You can now reorder dimension-level items directly from the dimensions page, eliminating the need to switch to the model builder.

Get more information about all these features here.

What's coming next?

• Existing dimension items edition: You will soon be able to edit existing dimension items from the table. 
• Renaming more than one item: We will soon allow you do to edition of more than one item at the same time from the dimension level settings.

We’ve made it easier to work with actuals data in the model builder!

What’s new?

• A new “Actuals” option is now available in the version drop-down menu, allowing you to focus exclusively on actuals formulas and values.
• When "Actuals" version is selected, the month picker allows you to select any time frame up to the last actuals date, providing flexible access to historical data.
• When creating a new model, the “Actuals” version is pre-selected by default, with a default time frame of the last 12 months.