Live Workflow

The current product is a guided data onboarding flow, not a loose upload form

The important part is not just accepting a spreadsheet. The important part is giving the team a controlled path from upload, to review, to ready handoff, without losing visibility halfway through.

1. Start with a spreadsheet upload

Begin in Clean a file with the spreadsheet the customer or operations team already has. Extrakt parses headers, prepares the run, and keeps the work inside a guided workflow instead of a loose file exchange.

  • Upload CSV or Excel without reformatting the whole sheet first.
  • Confirm the intended dataset early so the run is grounded in the right workspace.
  • Use the parsed headers preview to catch obvious setup mistakes before you process.

2. Review field matches and sample values

Before records move forward, Extrakt suggests field matches and shows sample values so the operator can validate what is about to happen. This is where bad assumptions get caught early.

  • Check suggested mappings, especially for similar-looking columns.
  • Review sample values to make sure the right fields are receiving the right data.
  • Use the review step to confirm that the run belongs in the correct dataset.

3. Work through review and invalid rows

After processing, rows are grouped by status so the team can focus on what needs attention instead of redoing the entire import. The goal is not perfect spreadsheets. The goal is a controlled cleanup flow.

  • Use `Needs review` for rows that require a decision or a closer look.
  • Use `Invalid` for rows with issues that block a clean handoff, such as broken dates or malformed emails.
  • Revalidate after fixes so the run reflects the latest status before you finalize it.

4. Mark the run ready and continue in the dataset workspace

When the output is acceptable for downstream use, mark the run ready. That action turns the run into a clear handoff point so the team can continue from the dataset workspace instead of getting stuck in the import table.

  • A ready run means the dataset can move forward even if some review context remains visible.
  • The next destination is the dataset workspace, where clean output remains accessible.
  • Ready imports should be treated as an operational checkpoint, not as the end of visibility.

Operator Playbook

A simple checklist for teams running real business data

The live workflow is designed for onboarding, operations, implementation, and support teams. The job is to make confident decisions with messy data, not to rebuild the sheet from scratch.

Before processing

  • Confirm the right dataset and verify the parsed headers.
  • Review a few sample values before trusting the suggested mappings.
  • Catch naming mismatches early rather than fixing them row by row later.

During review

  • Triage invalid rows first because they are the clearest blockers.
  • Use `Needs review` for human judgment calls rather than treating everything as an error.
  • Revalidate after meaningful edits so the run status stays current.

Before marking ready

  • Decide whether the clean output is good enough for the next operational step.
  • Keep unresolved rows visible as context, but do not let them hide the next action.
  • Move to the dataset workspace once the run is ready so the team continues from the right place.

Where Work Lives

A quick run should not disappear once the upload is done

One of the biggest workflow risks is losing context after processing. Extrakt should keep the same work reachable from the place you started and from the place you continue.

Clean a file

Use Clean a file to start new spreadsheet runs and reopen recent work without hunting for the underlying project path.

Dataset workspace

Once a run is ready, the dataset workspace becomes the main place to continue working with the clean output and the dataset-level workflow.

Projects

The same work also remains tied to the underlying project structure, so the shared file-cleaning workspace is not a dead end or a separate silo.

Developer rollout

The deeper DX layer is the next release wave

The public product focus today is the business workflow. The next release wave will add the developer-facing setup and delivery guidance back on top of that live foundation.

  • Embeddable setup guidance and framework-specific integration patterns
  • API references and implementation docs for the deeper DX layer
  • Webhook and delivery controls for product-facing handoff workflows