Version 16.0 - January
Version 16.0 Highlights
16.0 marks the launch of the new generation of data preparation — Turbo AI processes. It is a completely new no-code flow editor, with a faster engine for large volumes and a new reconciliation engine (analysis screen, manual reconciliation, justifications, and rule and relationship configuration). Throughout the version, database connectors (PostgreSQL, MySQL, Oracle, SQL Server, and SAP S/4 HANA) and SQL query with Databricks also arrived, expanding where data can come from. The other news and fixes are detailed below.
16.0.0 - December 15, 2025
New Features
New data preparation (Turbo AI processes)
Available when creating a process: in "New process", you now choose the Turbo AI type, which opens the new no-code data preparation editor. It is a completely new execution engine, designed for large data volumes and artificial intelligence features, without needing to write code to build the flow.
Build the flow by dragging steps in the editor, from loading to analysis.
Gain more speed and stability in processing, with the new data engine.
Continue relying on the Standard type when you prefer the standard preparation engine.
New reconciliation engine
Available in Turbo AI processes, through the Reconciliation Analysis step in the data preparation editor, with the analysis opening in the new Analytics screen. It is reconciliation rebuilt on the new engine, with automatic field discovery and automatic rule generation to speed up setup.
Configure relationships between the sides, reconciliation rules, and the review of displayed columns.
Manually reconcile and unreconcile records and register the justification directly on the analysis screen.
Monitor statistics, totals, and the detailing of each side of the reconciliation.
SQL Query step with Databricks connector
Available in the new data preparation (Turbo AI processes), in the SQL Query step when building the flow. You register a Databricks connector and bring large volumes of data directly to your process, with pagination for larger loads.
Register the Databricks connector once and reuse it in your flows.
Import large volumes of data with paginated search.
Transform SQL step in the new engine
Available in the new data preparation (Turbo AI processes), in the Transform SQL step within the flow. It now runs on the new data engine, with faster processing for SELECT, JOIN, WHERE, and GROUP BY — maintaining the same configuration method as always.
Write your queries as before and gain performance in execution.
Receive clearer error messages when the query has a problem.
Text File and Excel steps in the new engine
Available in the new data preparation (Turbo AI processes), in the data source steps when building the flow. The Text File and Excel sources now run on the new engine, with faster and more stable reading of input files.
Load text files (single structure) and Excel spreadsheets directly into the new flow.
Adjust advanced Excel source settings in a reorganized space within the step.
Automation base: API trigger and Source Directory
Available in Turbo AI processes, in the Process Settings, under Automation Rules. You now trigger execution through external systems via API and let the platform search for input files on its own, without depending on running everything manually.
Register an automation rule with API trigger, defining reference date and execution folder.
Connect a Source Directory (SFTP, S3, and similar) for the platform to read input files automatically.
Consult the status of each execution triggered by the integration.
Improvements
No improvements in this version.
Fixes
No fixes in this version.
Points of Attention (Breaking Changes)
When creating a process, the old Dattos AI option gave way to the new Turbo AI type in "New process" — artificial intelligence features now live in the new data preparation.
16.0.1 - December 19, 2025
New Features
Monitoring the execution phase
Available during the execution of a process, in both Turbo AI and Standard processes. While the process runs, the platform now shows the name of the step or source being processed at that moment — before, you only saw that it "is processing," without knowing at what point the execution was.
Track in real-time which step or source is being executed.
See the information update automatically as the execution progresses through the steps.
Reduce the feeling of slowness and the doubt about whether to cancel a long execution.
Improvements
No improvements in this version.
Fixes
No fixes in this version.
Points of Attention (Breaking Changes)
No points of attention in this version.
16.0.2 - December 26, 2025
New Features
No new features in this version.
Improvements
No improvements in this version.
Fixes
No fixes in this version.
Points of Attention (Breaking Changes)
No points of attention in this version.
16.0.3 - January 14, 2026
New Features
Multiple record structures in the Text File step
Available in the data preparation flow editor, in the Text File step, within Turbo AI processes. The step now reads text files that mix different column patterns on different lines — such as bank and CNAB files with header, detail, and trailer —, which previously required a single structure.
In Map Columns, choose "Multiple record structures" when the file does not follow a single column pattern in all lines.
Identify each type of record by position (fixed width) or by delimiter, with its own mapping and column names for each type.
Generate a separate output per record type from the same imported file.
Keep settings when re-importing the file or apply a new reading when needed.
Improvements
No improvements in this version.
Fixes
No fixes in this version.
Points of Attention (Breaking Changes)
No points of attention in this version.
16.0.4 - January 16, 2026
New Features
Chain processes as a data source
Available in the data preparation flow, in both Turbo AI and Standard processes, through the Integrate Process step (in the Data Source category). The result of an already executed process now feeds the flow of another, even when the two were built in different engines — before, each process was isolated.
Use the result of a Standard process as input for a Turbo AI process, and vice versa.
Choose the source type (Standard or Turbo AI) and locate the desired process directly in the list, with the analysis type label (Analytics or Analytics 2.0) to avoid confusing processes with the same name.
Define how the data will be brought: by a reference date, by date range, or by the last available execution.
The step, previously called Integrate Preparation, was renamed to Integrate Process throughout the platform.
Step palette and formulas in the Turbo AI engine
Available in the data preparation flow editor of Turbo AI processes. The data transformation library of the new engine has been expanded, bringing to Turbo AI processes the same well-known preparation features — now ready for large data volumes.
Build transformations with the Transform, Filter, Consolidate, Sort, De-duplicate, Stack, Merge, Lookup, Transpose Columns, Transpose Rows, Split Text, and Fill steps.
Connect external sources via the Rest API step within the flow.
Use date and calendar formulas (such as business days, conversion, and operations between dates), as well as number and text conversion ones, when handling data.
Improvements
No improvements in this version.
Fixes
No fixes in this version.
Points of Attention (Breaking Changes)
The Import Identifier field was removed from the Integrate Process step (formerly Integrate Preparation): old processes that used it need to be opened and saved again without this field to return to executing normally.
16.0.5 - January 23, 2026
New Features
New database connectors in Turbo AI processes
Available in Turbo AI processes, in the Database step of the data preparation flow. Previously, Turbo AI processes only had the Databricks connector; now you connect directly to major corporate databases and bring the data to the flow with a SQL query, with controlled reading to run well even at high volumes.
Connect to PostgreSQL, MySQL, Oracle, SAP S/4 HANA, and SQL Server, choosing the connector already configured on the platform.
Write the SQL query in the step editor, with the number of records for sample, and validate the result before saving.
Import large volumes with reading optimized for the Turbo AI engine, maintaining stable execution until the end.
Improvements
No improvements in this version.
Fixes
We fixed an error where the chart of accounts for Account Substantiation invalidated the Reconciliation Map code used in legacy reconciliation.
Points of Attention (Breaking Changes)
No points of attention in this version.
16.0.6 - January 30, 2026
New Features
No new features in this version.
Improvements
No improvements in this version.
Fixes
We fixed an error when changing the month of a Closing Group in Account Substantiation.
Points of Attention (Breaking Changes)
No points of attention in this version.
16.0.7 - February 6, 2026
New Features
No new features in this version.
Improvements
No improvements in this version.
Fixes
We fixed an error in data import where the increment and replace buttons did not appear, preventing the choice of how to load the data.
We fixed an error that caused the process to start execution from a very old reference date.
We fixed an error in the detailing of reconciliation lines where columns appeared in a different order than in the analysis, leaving values misaligned — in the reconciliation used by legacy Reconciliation (Reconciliation menu) and by Standard processes.
We fixed an error when generating automatic reconciliation relationships, in the reconciliation used by legacy Reconciliation (Reconciliation menu) and by Standard processes.
Points of Attention (Breaking Changes)
No points of attention in this version.
16.0.9 - February 13, 2026
New Features
No new features in this version.
Improvements
No improvements in this version.
Fixes
We fixed an error that prevented reading Excel source files in the data preparation flow of Turbo AI processes.
Points of Attention (Breaking Changes)
No points of attention in this version.
16.0.10 - February 19, 2026
New Features
Smart Builder in the flow editor
Available in the data preparation flow editor, in both Standard and Turbo AI processes. Smart Builder is the Dattos AI chat assistant within the editor itself: you describe in natural language what you want to change in the flow and it applies the change in context, without needing to switch screens.
Open the chat panel via the floating icon on the side of the editor and keep the flow visible while chatting.
Request adjustments to the flow in Portuguese, English, or Spanish, according to your interface language.
Let Dattos AI build and modify flow steps based on what you describe.
Improvements
No improvements in this version.
Fixes
We fixed an error when loading the data sample when including a Dattos Formula in the data preparation flow editor.
We fixed an error that occurred when opening an analysis that had not yet been executed: the platform redirected to the data preparation configuration, and users with only execution permission received an error.
Points of Attention (Breaking Changes)
No points of attention in this version.
16.0.11 - February 23, 2026
New Features
No new features in this version.
Improvements
No improvements in this version.
Fixes
We fixed an error that prevented the import of multiple Excel files into the same data source when a column had mixed text and number values.
We fixed the display of the reconciliation status column in process integration import (integrate process), which had stopped appearing.
Fixed the incorrect name of automatically generated columns in the data preparation flow.
Points of Attention (Breaking Changes)
No points of attention in this version.
16.0.12 - February 26, 2026
New Features
No new features in this version.
Improvements
No improvements in this version.
Fixes
No fixes in this version.
Points of Attention (Breaking Changes)
We applied a package of security fixes identified in intrusion tests (pentest), reinforcing access controls and permission verification in active session management.
16.0.13 - March 4, 2026
New Features
No new features in this version.
Improvements
No improvements in this version.
Fixes
We fixed an error in process execution monitoring: when an execution failed, the failure was not signaled in real-time during processing.
Points of Attention (Breaking Changes)
No points of attention in this version.