AUDIENCE TYPE: Account List CSV / CSV Upload - Accounts (platform stores `customAudienceType=FIRMOGRAPHIC_INCLUDE`; the main UI's audience-map renders this enum as "CSV Upload - Accounts").
            This tool is ONLY for company/account-level data (company names and websites).
            Do NOT use this tool for contact-level data (emails, phone numbers, individual people).
            For contacts, use upload_contact_list_csv_audience instead.

            TWO WAYS TO SUPPLY THE ACCOUNT LIST — provide EXACTLY ONE of:
              • `companies`: an inline JSON object map `{ "<companyname>": "<companywebsite>", ... }`. Use this for short ad-hoc lists you have already parsed into context.
              • `companies_source_csv_url`: a public URL of a CSV file with header EXACTLY `companyname,companywebsite` (case-insensitive, whitespace-trimmed). Use this WHENEVER THE USER ATTACHED A CSV to the chat — the URL is surfaced to you via `AudienceBrief.attached_file_urls`; pass it through verbatim. The MCP server downloads + validates + parses the CSV so a multi-megabyte file never has to travel through your context.

            If both are provided, or neither, the tool errors with a clear message — pick one.

            OPTIONAL CONTACTS CRITERIA: layer per-contact filters on top of the account list, matching the same shape `create_firmographic_audience` uses:
              • `location_country_ids` / `location_state_ids` — geography
              • `job_title_includes` / `job_title_excludes` — free-text title keywords
              • `job_function_include_ids` / `job_function_exclude_ids`
              • `seniority_include_ids` / `seniority_exclude_ids`
              • `contacts_per_company_limit`
            When set, the resulting audience matches only contacts inside the listed companies that ALSO satisfy these filters — e.g. "this 1,999-account list, but only VP/Director/Manager Engineering contacts in Canada + Saint Pierre and Miquelon + United States" becomes a single call. When all of these are omitted, every contact in every uploaded company is matched.

            WHEN TO USE:
            - "Upload these companies/accounts as an audience" / "Create an account list audience" / "Upload account list CSV"
            - "Create a CSV Upload - Accounts audience with this CSV and these contact filters"
            - The user attached a CSV of companies (website, with or without a name) and asked for a CSV-Upload audience

            WHEN NOT TO USE:
            - When the data is contact-level (emails, phone numbers, individual people) — use `upload_contact_list_csv_audience`.
            - For pure firmographic targeting without an attached account list — use `create_firmographic_audience`.

            INLINE-MAP DATA MAPPING RULES (when you choose the `companies` path):
              - Each KEY must be the **company name**; each VALUE must be the **company website URL**.
              - Format: { "<companyname>": "<companywebsite>", ... }
              - Example:
                  {
                    "Acme Corp": "https://acme.com",
                    "Globex International": "https://globex.com",
                    "Salesforce": "https://salesforce.com"
                  }
              - DO NOT pass column headers as keys; DO NOT reverse the mapping; DO NOT send raw file paths or bytes.

            CSV-URL HEADER RULES (when you choose the `companies_source_csv_url` path):
              - The first row of the CSV MUST be exactly `companyname,companywebsite` (case-insensitive — `Company Name,Website` is rejected; rename the columns first or fall back to the inline map path).
              - Only `companywebsite` must be filled: the platform matches accounts on the website domain and treats `companyname` as optional, so a file whose name column is empty on every row is a valid account list. Never ask the user to fill in company names.
              - Rows with an empty company website are dropped server-side.
              - The server enforces a 50 MB cap on the downloaded file.

            PARAMETERS:
            - audience_name (required)
            - companies (optional, mutually exclusive with companies_source_csv_url)
            - companies_source_csv_url (optional, mutually exclusive with companies)
            - location_country_ids / location_state_ids (optional contacts criteria)
            - job_title_includes / job_title_excludes (optional)
            - job_function_include_ids / job_function_exclude_ids (optional)
            - seniority_include_ids / seniority_exclude_ids (optional)
            - contacts_per_company_limit (optional)

            RETURNS:
            - success, id / audience_id, audience_name, audience_type (FIRMOGRAPHIC_INCLUDE), file_id, companies_count, upload_filename, expectedNumberOfCompanies, expectedNumberOfContacts, cappedContactCount
            - companies_count: the rows sent, each with a website. dropped_without_website names up to 25 rows left out for having none, which the platform cannot match; dropped_without_website_count counts them all.
            - companies_summary: status is matching, because the platform matches the list after this call returns, and submitted_companies is how many rows the platform took in. Read the final count later with get_deep_audience_details, and never report companies_count or expectedNumberOfCompanies as the audience's final size.

            IMPORTANT NOTES:
            - There is a small delay between upload and audience creation while the platform processes the file; the tool waits for that internally.
            - The companies map / source CSV must contain at least one row with a company website.
            - Company websites should be valid URLs (e.g., https://example.com).