AI examples

SPACE AI brings model understanding, data analysis, development, troubleshooting, and connected data into the same conversational workflow. You can start with a business question, a model object, an error, or a task and let the assistant use the relevant SPACE capabilities and connected systems.

The examples below show different ways SPACE AI can be used with IBM Planning Analytics. Exact results depend on your connections, permissions, model, and available data.

Understand an unfamiliar Planning Analytics model

Instead of manually searching through cubes, dimensions, rules, and processes, start with what you need to understand.

Explain how revenue is calculated in this model.

Which cubes, rules, and processes are involved in the forecast?

Find everything related to headcount planning and explain how the pieces fit together.

Where is the Product dimension used?

This is useful when onboarding to an unfamiliar model, reviewing inherited applications, investigating dependencies, or preparing to make a change.

See Models and logic.

Understand rules and feeders

Ask SPACE AI to inspect existing rule logic and explain it at the level you need.

Explain the Sales cube rules and identify the main calculations.

How is Gross Margin calculated?

Review the feeders for this cube and explain which calculations they support.

Explain this rule in business terms rather than TM1 syntax.

You can continue from explanation into investigation or development without starting a separate workflow.

Understand and review TurboIntegrator processes

SPACE AI can inspect TI processes, explain their logic, and follow their dependencies.

Explain the Load_GL_Actuals process and summarize what each section does.

Which cubes and dimensions does this process read from or update?

Review this process and identify anything I should consider before modifying it.

Find the process that loads actual revenue and explain the data flow.

This can be especially useful when working with complex or poorly documented TM1 applications.

Analyze TM1 data

Start with a business question rather than constructing an MDX query manually.

Compare Actual and Forecast revenue by region for this quarter and show the largest variances.

Which products contributed most to the decline in gross margin?

Show monthly operating expenses for the top five departments and summarize the trend.

Find the entities with the largest variance to budget and help me investigate them.

SPACE AI can inspect the model context, construct the required TM1 query, retrieve data in manageable result sets, and continue through larger results when needed.

You can also ask SPACE AI to visualize the results:

Show Actual and Forecast revenue by month as a chart.

Visualize the largest variances by department.

When an analysis produces a useful result, save it as a TM1 cube view instead of leaving it only in the conversation:

Save these results as a private view called Revenue Variance Analysis.

Save the last results as a public view called Monthly Forecast Review.

The saved view can then be reused in IBM Planning Analytics Workspace or through SPACE and SPACE AI. Views are private by default unless you request a public view.

See TM1 data.

Move from a variance to its model logic

A data question does not have to end with the numbers. Continue the investigation into the model itself.

Why is this forecast significantly different from Actual?

Which rules affect this value?

Is this value calculated, loaded by a process, or entered directly?

Find the logic responsible for this result and explain it.

This connects data analysis with model understanding and can help distinguish a business variance from a modeling or data-loading issue.

Investigate TM1 errors

Start with the symptom rather than manually deciding which log or object to inspect.

This TurboIntegrator process failed last night. Find out why.

Investigate this TM1 error and explain the likely cause.

Check the relevant server logs and process logic for this failure.

A value changed unexpectedly yesterday. Help me determine what changed it.

Depending on the problem, SPACE AI can use TM1 server, transaction, or audit information together with related model objects and logic.

See Investigating errors.

Investigate SPACE errors

SPACE AI can also help troubleshoot SPACE itself.

I received this error in the AI Assistant about 20 minutes ago. What happened?

This SPACE operation failed. Search the relevant logs and investigate it.

Find this error in the logs and follow the related trace.

Why did this scheduled job fail?

The assistant can use available SPACE application logs and Journal information, and can follow trace IDs when they are available to locate related application activity.

Create and modify TM1 objects

SPACE AI is not limited to explaining existing models. It can perform supported development tasks when the user has the required access.

Create a private view showing Actual and Budget revenue by month for the current year.

Create a subset containing the active cost centers.

Update this TI process to include the new scenario.

Modify the rule to calculate the new measure using the existing pattern.

SPACE supports creating and updating rules, TurboIntegrator processes, views, and subsets through the AI Assistant. Views and subsets are private by default unless you request a public object.

Review and test model changes using the normal development practices for your environment.

Work across Planning Analytics environments

When multiple TM1 connections are available, the same conversation can work with more than one environment.

Compare this rule between Development and Production.

Is this TI process the same in Test and Production? Explain the differences.

Compare this data between the two environments and identify the differences.

Check whether the objects related to this change are consistent between Development and Production.

For lifecycle workflows specifically designed to compare environments and manage changes, see Releases and deployments.

Combine Planning Analytics and BigQuery data

SPACE AI can work with native Google BigQuery connections alongside TM1.

Compare the forecast in TM1 with actual sales from BigQuery.

Use the detailed BigQuery data to investigate this variance in Planning Analytics.

Show the largest differences between the TM1 plan and the operational data in BigQuery.

Start with this TM1 result, then find the related transactions in BigQuery and summarize what explains the difference.

This allows planning and model context from IBM Planning Analytics to be analyzed together with supporting data from another connected source.

See Connected data.

Use your existing identity and security

IBM Planning Analytics and Google BigQuery connections can use SSO, allowing SPACE AI to work with connected data using the individual user's authentication and permissions.

This means different users can ask the same type of question while continuing to see only the data and objects they are authorized to access in the underlying system.

For example:

Show my available Planning Analytics cubes and help me find the one used for workforce planning.

Query the BigQuery data I have access to and compare it with this TM1 forecast.

Users do not need to provide their source-system credentials to SPACE when using supported SSO connections.

Turn technical logic into business explanations

SPACE AI can help bridge the gap between model developers and business users.

Explain how the forecast is calculated for a finance user.

Summarize what this allocation process does without using TI terminology.

Explain why this number changes when the scenario is switched from Budget to Forecast.

Create a concise explanation of this model logic that I can send to the FP&A team.

The same underlying model information can be explained differently depending on the audience and purpose.

Ask how to use SPACE

SPACE AI can also answer questions about SPACE itself using the available product documentation.

How can SPACE help me investigate a failed TI process?

How do I create a private subset with the AI Assistant?

How can I compare two Planning Analytics environments?

How do connections and SSO work in SPACE?

This lets users begin with the task they want to accomplish rather than first finding the correct documentation page or SPACE module.

Build on the conversation

Many of the most useful workflows combine several of these capabilities. Start with a broad question and refine the task as SPACE AI discovers relevant information.

For example:

Why is North America revenue below forecast?

Then continue:

Break the variance down by product.

Now check whether the largest variance is affected by a rule.

Explain that rule.

Check whether the rule is different in Development.

Then turn the useful result into something reusable:

Visualize the variance by product as a chart.

Save these results as a private cube view called North America Revenue Variance.

A single investigation can move from business data to visualization, model logic, environment comparison, troubleshooting, and development while preserving the context of the conversation. Useful analysis results can also become reusable Planning Analytics views rather than remaining only in the chat.