Skip to main content
Parley includes an import-data Skill for spreadsheet and legacy-system migrations. You do not need to clean your export, design a Parley-ready schema, or split it into perfect CSVs before giving it to the Agent. Start with the source you have: a CSV, TSV, Excel workbook, or provider export. Parley Agent inspects the complete source, identifies the entities and relationships it contains, finds inconsistent values and duplicate candidates, and proposes how the data should map into Parley. The Agent can handle the preparation work that happens before the underlying importer runs, including:
  • Profiling every relevant sheet and column
  • Normalizing headers, names, dates, labels, and field values
  • Separating combined source data into Accounts, Contacts, Projects, and notes
  • Identifying natural matching keys and duplicate candidates
  • Comparing the source against records already in Parley
  • Proposing new resource fields where the source contains structured data Parley does not yet represent
  • Resolving or flagging Account, Contact, and Project relationships
  • Producing the prepared CSVs used by the bulk loader
  • Maintaining mapping, routing, and progress ledgers for recovery and reconciliation
The distinction is between the workflow and the loader. The import-data Skill accepts the source export as it exists and prepares it collaboratively. The underlying CLI importer is intentionally strict: it loads the reviewed, normalized CSVs produced during that workflow. You do not need to prepare those CSVs yourself.

​
What can be migrated

The migration workflow supports:
  • Accounts for companies, employers, petitioners, agencies, and other organizations
  • Contacts for clients, beneficiaries, employees, signatories, attorneys, and other individuals
  • Projects for matters, cases, or other units of work
  • Resource fields on Accounts, Contacts, and Projects
  • Links between Projects and their Accounts and Contacts
  • Shared notes owned by an Account, Contact, or Project
If a legacy export contains data that does not map directly to a supported Parley record, the Agent identifies it rather than silently discarding it. You can decide whether to omit it, preserve it in a file or note, or map it to a supported concept where appropriate.

​
Review the migration plan

Before writing anything, Parley Agent presents a complete mapping and migration plan. The review covers:
  • The source files, sheets, and rows in scope
  • How every source column will be handled
  • Which records appear to match existing Parley data
  • Which records are expected to be created or updated
  • Proposed resource fields and value transformations
  • Duplicate candidates and conflicting records
  • Unresolved or ambiguous relationships
  • Data that will be skipped or preserved outside structured fields
Nothing is created, updated, or linked until you approve the scope and mapping. This checkpoint prevents a guessed mapping from being repeated across an entire data set.

​
Check a representative sample

For a migration of more than about 20 records, Parley Agent proposes a representative sample of approximately five records before the full run. The sample should include more than the first few rows: it can cover common records, edge cases, existing matches, new records, and relationship examples. Review the resulting Accounts, Contacts, Projects, fields, and links. Once the sample reflects the intended structure, approve the full import.

​
Run the import

Parley Agent converts the approved mapping into prepared CSVs and loads each primary resource through one long-lived import process. Accounts are normally loaded before Contacts, and Accounts and Contacts before Projects, so relationships can resolve against records that already exist. Each row follows a match-or-create model. Existing records can be matched and updated with differing nonblank values; new records are created. Blank cells are skipped rather than treated as instructions to erase existing data. Before the full write, the Agent runs a dry run against live Parley data. The dry run checks mappings, field types, matching keys, duplicate or ambiguous matches, and unresolved links without creating or changing records.

​
Recover safely after an interruption

Every import maintains an append-only progress log. If a run is interrupted or some rows fail, Parley Agent reconciles the log against the records in Parley and resumes only the rows that still require work. Do not start over with an unrelated batch. The approved mapping, row identities, and progress log are what make recovery deterministic and prevent successful rows from being duplicated.

​
Reconcile the result

After the import, Parley Agent reads the affected records back from Parley and reports what happened. The reconciliation distinguishes records that were:
  • Created
  • Updated
  • Matched with no changes required
  • Linked
  • Skipped by the approved plan
  • Left ambiguous
  • Failed and still require attention
It also compares populated values against the approved column mapping so that missing fields or relationships are visible rather than concealed inside a success total.

​
Example migration request

Upload the export and tell Parley Agent the outcome you want:
Migrate the organizations, contacts, matters, and notes in this workbook. Preserve our matter numbers and statuses, create any resource fields we need, link each matter to the correct company and people, and show me the proposed mapping and a five-record sample before importing the rest.
You can provide the source in its current form. Parley Agent will inspect it, propose the transformations, and ask you to resolve only the choices that require substantive judgment.
Growth and Enterprise plans include guided migration support. If you are moving from a legacy case-management system or need a phased migration, contact the Parley team before beginning the full run.