AI Use Disclosure

1. OpenAI-assisted workflows

Vendinav currently uses the OpenAI Responses API in six workflows:

1. RFQ draft generation: turns a job description and existing RFQ lanes into draft Vendor-facing requests. Inputs may include the job title, buyer job reference, due date, description, RFQ title, category, status, and current Vendor-facing request.
2. RFQ package decomposition: proposes a structured package of category-specific RFQ drafts from job context, the available Vendor-category catalogue, and required category identifiers.
3. Single RFQ category drafting: drafts one structured scope using the job context, the selected category, and the names of other scopes being quoted separately.
4. Vendor category classification: recommends broad Vendor directory categories from uploaded Vendor fields, deterministic recommendations, active category options, and public website evidence.
5. Vendor spreadsheet normalization: maps uploaded spreadsheet content into normalized Vendor rows for administrator review. The submitted cells may include raw values and irrelevant uploaded columns, not only the fields ultimately kept as name, email, website, phone, location, or source category.
6. Request Access review assistance: extracts a structured operational summary and workflow signals from applicant answers. For newly disclosed eligible requests, Vendinav may also map the submitted company website and freshly retrieve up to six publicly available pages. Website-derived claims must cite captured source keys and are presented separately from applicant-reported facts. Application code applies the applicant-only deterministic scoring rubric; website evidence cannot alter that score, and the model cannot approve, decline, create an account, or activate Free. A person makes every access decision.

2. Firecrawl website retrieval

Vendinav may send public Vendor website URLs and an eligible Request Access company website URL to Firecrawl to retrieve publicly available page title and body evidence. Request Access retrieval is limited to the approved company host and up to six bounded pages. Firecrawl is not used to access authenticated pages, arbitrary subdomains, or a user's private website content.

3. Human review and limitations

• AI output may be inaccurate, incomplete, outdated, misleading, or non-unique.
• Users and administrators must review outputs and source material before relying on, editing, saving, or sharing them.
• Vendinav does not promise that an output is legally sufficient, commercially appropriate, or suitable for a particular project.
• These workflows do not autonomously send communications, select a Vendor, award work, form a contract, make a purchase, or initiate a payment.

4. Authority to submit data

A user must have authority to submit job, RFQ, Vendor, contact, website, spreadsheet, and other third-party data for this processing. Do not submit prohibited regulated or sensitive information. The customer remains responsible for notices, permissions, and lawful handling of third-party data.

5. OpenAI data controls

OpenAI states that API inputs and outputs are not used to train or improve its models unless Vendinav opts in to that use. Vendinav sets store: false on these Responses requests, which prevents Responses application-state storage for these requests.

Default abuse-monitoring logs may still contain prompts, outputs, or related content for up to 30 days unless approved provider retention controls apply; longer retention may also occur where required by law or reasonably necessary to protect services or others. Vendinav does not represent that special provider-approved retention controls apply to its account.

See OpenAI’s current API data controls documentation.

6. Firecrawl data controls

Vendinav sets storeInCache: false on the Firecrawl request. Firecrawl documents this setting as preventing the page from being stored in its index and cache. Firecrawl offers a separate provider-enabled no-retention option, and Vendinav does not represent that this separate option is enabled.

See Firecrawl’s current Scrape API documentation.

7. Questions and rights requests

Questions about Vendinav’s AI processing, correction requests, and privacy rights requests may be sent to privacy@vendinav.com.