Skip to main content

Report Onboarding

Alpha feature

Report Onboarding is an alpha feature. It is designed to get you about 80% of the way to systematic, data-driven reporting. For complex cases you will often need to tweak the result afterward - that is expected.

Report Onboarding automates the data-integration setup for a recurring report. Instead of manually creating collections, transformations, and workflow connections, you provide source workbooks and a sample report - Octigen's data assistant builds the pipeline and verifies the result against your sample.

Everything it creates stays transparent and editable: collections (warehouse ingestion), transformation steps, and workflow connections. When the automated result needs refinement, adjust those artifacts directly rather than starting from scratch — see Data Collections (Warehouse) and Transformations.

Creates all-new objects

Report Onboarding always creates new collections and transformations for the run. It does not reuse warehousing or transformation objects that already exist in the space. If a majority of the needed collections and transformations are already available, prefer the data bot on the Transformations or Warehouse pages to add only what is missing, then wire connections yourself — onboarding is the wrong tool for that incremental case.

When to use it​

Report Onboarding is ideal when:

  • You already have a configured template (or are integrating against an existing one).
  • You receive Excel or CSV workbooks with the figures for one concrete report run.
  • You have a sample PowerPoint report that shows exactly what the automated output should look like for that same data.
  • You want a strong first pass toward a fully deterministic, data-driven workflow you can rerun every period — and are ready to refine edge cases in the warehouse and transformations if needed.
  • The space does not already contain most of the collections and transformations you need (greenfield or near-greenfield data setup).

For a step-by-step manual walkthrough of the same end result, see Building a Recurring Report (Factsheets).

What you provide​

  1. Template - the PowerPoint layout to integrate against (charts, tables, and text fields should already be named and categorized).
  2. Source workbooks - .xlsx, .xls, or .csv files containing the data for one report run (one valuation date, one fund, and so on).
  3. Sample report - a .pptx file that contains exactly the data from those workbooks. This is the ground truth used to verify the integration.
  4. Naming prefix (optional, recommended) - a short prefix applied to every collection and transformation the run creates (for example mdn → mdn-holdings). Prefer always setting one: after a cancelled or experimental run you will want to find those objects quickly in Warehouse / Transformations.
Sample must match the workbooks

Do not upload data for one month and a sample report for another. The sample is scored against the generated output; mismatched inputs will fail verification.

Credits​

Report Onboarding bills per shape, and only for shapes that reach a sufficient quality match against your sample. The estimate shown before you start is an upper bound (shapes still needing connections × credits per shape) — what you would pay if every shape succeeds. Shapes that never reach a sufficient match are not charged. Cancelling mid-run stops further work; credits already incurred for shapes that had reached sufficient quality remain billed.

Available on Trial and Business plans. Trial spaces must have enough remaining credits for the full estimate before a run can start; Business may continue into overage.

What the assistant builds​

The onboarding assistant runs in the background and typically:

  1. Analyzes your workbooks and sample report to understand which shapes need which data.
  2. Creates data collections - one per source sheet where appropriate, with consistent parameters (for example valuation date and fund identifier).
  3. Defines transformations - filters, column selection with renames to presentation headers, sorting, and reshaping so each table matches the target shape format (including header labels from the sample).
  4. Creates a data-driven workflow - connects each chart, table, and text field to the right collection or transformation, with Data Driven Only mode enabled.
  5. Runs a verification report - generates one deck from the new workflow and scores it against your sample (improvement happens during the build cycles beforehand).

You can navigate away while onboarding runs; progress continues in the background and appears in the task manager.

After onboarding​

When onboarding completes successfully:

  • Collections and transformations appear under Data → Warehouse and Transformations.
  • A new data-driven workflow is attached to your template.
  • Producing the next period's report is the same as any other connected workflow: drop fresh workbooks in the Data Landing Zone, start the workflow, enter parameters, and generate.

Treat the result as a starting point, not a finished production setup — especially for complex layouts. Review and adjust:

  • Data Collections (Warehouse) — ingestion points, filename patterns, and parameters
  • Transformations — filters, column selection/renames, sorting, and reshaping steps
  • Data connections — which shapes wire to which collections or transformations

Those surfaces are the same tools you would use for a fully manual setup, so fixes stay in the open pipeline rather than in opaque AI output.

Monitoring progress​

While onboarding runs, the progress view shows:

  • Current phase (setup, verification).
  • A live data connections graph of collections, transformations, and shape wiring as they are created.
  • Scoring feedback from the verification run.

You can cancel an in-progress onboarding task from the progress screen.

Cancelling and cleanup​

Cancel stops the background loop. Automatic cleanup then removes internal run artifacts:

  • Verification report instances and their internal deck sessions (on cancel/fail; on a successful run these are kept until you use Clean up)
  • Staged source workbooks and the sample upload (on cancel/fail; kept after success until Clean up)
  • The temporary hidden sample template used for scoring
  • Superseded data-only workflows created during the run (the final kept workflow stays)
  • On a successful completion only: run-created collections/transformations that the final workflow does not use

Automatic cleanup does not delete the collections and transformations that feed the kept workflow — those are the deliverable. After a cancelled or unsatisfactory run, those objects also remain in the space.

To remove collections or transformations yourself:

  1. Open Data → Warehouse or Transformations.
  2. Switch to list view.
  3. Search (ideally by the naming prefix you set), select the objects, then Actions → Delete.

That is why a naming prefix is recommended on every run. Use Clean up on a completed run when you are done with the kept sample files and verification downloads — that action does not delete warehouse or transformation objects either.

Manual alternative​

Report Onboarding covers steps 2–4 of the factsheets tutorial in one automated pass. Choose the manual path (or the data bot under Warehouse / Transformations) when you need fine-grained control, or when most collections and transformations already exist and you only need a few additions.

ApproachBest for
Report Onboarding (alpha)Fast first setup (~80% there) in a greenfield or near-empty data space; standard factsheet-style pipelines you may refine
Data bot (Warehouse / Transformations)Most objects already exist; create or fix only the missing pieces
Manual setupUnusual data layouts, strict governance, or building each step yourself from the start

Prerequisites​

  • A template with quantitative slides and clearly named chart, table, and text shapes.
  • Data connections entitlement for your workspace (data-driven workflows require it).
  • Workbooks and sample report from the same reporting period.

Next steps​