Data transformation enhancements for Intellistack Streamline workflows
Some data transformations need interpretation. Others need precision.
That difference becomes obvious when an Intellistack Streamline workflow has to do more than clean up text. Maybe you need to calculate a total, apply a discount, convert a score, reformat a date, or shape output using code or formula-like instructions. In those cases, a “close enough” answer is not enough. You want a result that is dependable and repeatable every time the workflow runs.
Transform Step, our prompt-based, AI-native data transformation feature, now does a better job with exactly those workflows. It delivers more consistent results for math-heavy transformations, and it better handles prompts that use code or code-like instructions to describe how data should be reshaped. The outcome is better results for the kinds of transformations that need exactness, not just interpretation.
Better support for calculation-heavy transformations
Math inside a workflow sounds simple until it is part of a real business process. Pricing, scoring, percentages, date differences, weighted averages, and final amounts all need to land on the expected answer without drift.
That is where this improvement matters. Transform Step is now better suited for transforms where the output depends on calculations across multiple inputs, especially when the logic needs to stay stable from one session to the next.
Examples include:
- calculating a final invoice amount from price, quantity, discount, and tax inputs
- turning component scores into a weighted final score
- computing age or elapsed time from date fields
- converting raw values into percentages, averages, or threshold-based outputs
Better handling for code-based transformation prompts
Not every builder writes a prompt the same way. Some describe the outcome in plain language. Others are more explicit and use formulas, code, or code-like instructions to define exactly how the output should be shaped.
Transform Step now handles those code-based prompts more reliably. That gives builders more flexibility in how they express a transformation while improving confidence that the output will match the intent behind the prompt, especially when the transformation depends on exact syntax, conditional logic, or repeatable rule-based output.
Practical examples you can use right away
BeforeAfterImpactTeams manually calculate totals from quantity, unit price, discounts, and tax before sending values downstream.Transform Step calculates the final amount automatically as part of the workflow.Less manual math, fewer handoff errors, and cleaner downstream data.Review teams combine multiple criteria by hand to create a final score.Transform Step generates weighted scores automatically from the mapped inputs.More consistent evaluations and less spreadsheet work.Follow-up workflows rely on someone figuring out how many days have passed since an event or injury date.Transform Step computes elapsed days automatically from the input dates.Faster follow-up workflows and more reliable timing-based logic.Document output depends on conditional, code-like instructions that are easy to get wrong or apply inconsistently.Transform Step handles code-based prompts more reliably and produces repeatable output.Better control over document-ready text and fewer formatting or logic errors.
Available now in Intellistack Streamline
These improvements to Transform Step are now available to any Intellistack Streamline customer. If you already use Transform Step for cleanup, classification, or formatting, you can now use it with more confidence for calculation-heavy and more code-based transformations too.
