Skip to main content

Pelcro AI Co-Pilot - Import tool

Pelcro Product

Problem Statement

As a publisher admin, I have no way to quickly query my subscription data, analyze trends, or act on insights without manually running reports or contacting Pelcro support, which results in slow decision-making and dependency on external help for routine data questions.

User Story

As a publisher admin, I want an AI assistant embedded in my dashboard that understands my live Pelcro data and can analyze uploaded files, so that I can get instant answers, surface insights, and execute guided actions without leaving the dashboard.

Definition of Done

A feature is done when:

  1. Given I am on the publisher dashboard, when I open the AI Copilot panel and ask a question about my subscriptions or revenue, then the assistant responds in real-time by calling live Pelcro tools and streaming the answer token-by-token with tables, charts, or action buttons as appropriate.

  2. Given I upload a CSV, Excel, or PDF file, when the upload completes, then the copilot can reference that file in my next message and analyze its contents alongside my live account data. Previously uploaded files remain downloadable from past conversations.

  3. Given I navigate away and return to a past conversation, when I load it from the history sidebar, then all message content is fully restored including tool call steps, artifact tables and charts, file upload pills with download links, and action button states.

  4. This change will impact the UI (new Copilot panel, message partials, history sidebar), the API (new routes under /admin/copilot-ai/ for chat, conversation history, file upload, and file download), and the Database (new agent_conversations table). SDK and Integrations: no changes.

  5. This solution will include the following limitations: conversations are scoped per publisher account; file uploads support CSV, Excel, and PDF only; the agent performs up to 25 reasoning steps per message; S3 storage is required for file uploads and downloads; action buttons from past conversations are shown as completed badges and are not re-executable.

Posted by Abdallah Shapsough ·
Status: Available5 comments

Log in to comment and vote

Comments5

  • Michael Ghattas

    Team•

    Apr 30

    https://youtu.be/hx5nHH5PetE?si=Mf3ZYmfvm2-esp9N

  • Michael Ghattas

    Team•

    Apr 14

    We just released the new import tool. Please check out the related docs here:

    https://docs.pelcro.com/docs/ai-agent

    To enable the free trial, please contact support@pelcro.com - this is a beta feature that is opt in on a free trial basis.

  • Michael Ghattas

    Team•

    Apr 2

    Sneak peak of the AI agent that will be available for import, mass updates and much more.

  • Asif

    •

    Apr 2

    •

    Merged request

    •

    1 vote

    Bulk CSV Customer update - Feature Request

    🔍 Problem Statement

    There is no way to bulk-update customer records via CSV upload. When client data changes in bulk (e.g. address corrections, company assignments, metadata), admins must update each record individually — which is impractical for large datasets.

    💡 User Story

    As a Pelcro admin, I want to upload a CSV to bulk-update customer records, so that I can apply mass data changes (addresses, metadata, company) efficiently without editing records one by one.

    🎯 Definition of Done (DoD)

    A feature is done when:

    ✔️ Given a CSV file with customer updates (keyed by customer ID or email), when an admin uploads it, then matching customer records are updated with the new values — and a summary report shows successes, failures, and skipped rows.

    • Rana Haleem

      Team•

      Sep 18, 2025

      Thanks so much for sharing this feedback — it’s really helpful to hear how you’re using the platform and where the pain points are.

      To help us better understand your use case, could you share a bit more detail on:

      • What types of actions or updates you’d ideally like to perform through the UI (e.g., address changes, bulk field updates, deletions)?

      • How frequently do you usually need to run these large updates (quarterly, monthly, or more often)?

      This context would be super valuable as we review your suggestion and evaluate potential improvements.

      Thanks again for taking the time to share your experience!