AI Ticket Chat is a support ticket board inside your Drupal administration: request, conversation, delivery plan and minutes spent, in one place. Site owners raise and approve tickets; your support team estimates, plans and delivers them from the same screen.

The day you install it, a Support entry appears in the administration menu with a board of five phases and fifteen editable sections. Connect the optional AI service and each ticket also gets a summary, a classification and a milestone plan with hour estimates your team approves before work starts.

What you get

A board that fits your process. Five phases (Waiting, Saved, Suggestions, In progress, Completed) hold sections such as Drafts, To estimate, Needs response, Queued backlog and In development; you rename, reorder and recolour them and edit the allowed transitions.
A ticket detail that keeps the whole story. Live comments, attachments, favourites, related-ticket search and the plan's milestones.
Client mentions that park the ticket. Mentioning the client moves the ticket to Needs response; their reply puts it back.
Fourteen notification types on the bell. From task assigned and task needs valuation to task valued, budget ready and the cost limit warnings.
A support team with departments. A soporte role, departments with colours, per-member working hours, cost and sale rate per hour, holidays and blocked days.
Milestone timers with a work mode. One active timer per person, per-step checklists and a per-member record of attendance, performance and profitability.
Task templates and automation. Save any ticket as a template; cron opens new ones on a schedule, when chosen content is created or as the next step of a chain.
Reports you can hand to accounting. KPIs, support time by department, AI usage, a month filter and export to CSV or XLSX, plus one-off importers from an XLSX or CSV file or a Monday.com board.

How it works

Everything hangs off /admin/aiticketchat. A ticket is a node of the tarea_de_soporte content type; a milestone is a support_subitem entity attached to it. Tickets move between sections by the transitions you allow; sections belong to phases, and the phases are the tabs.

Three backend modes decide where tickets go. Local, the default: this site manages its own tickets. Assisted: it pushes a copy of each new ticket to a named provider, which returns the estimated plan on cron. Centralized: this site is the hub, receiving its client sites' tickets and publishing the plans back.

Free, and what the optional service adds

Everything above works without connecting to any external service (the Monday.com importer, naturally, talks to Monday). In local mode the AI buttons are simply not there; the dashboard shows a notice linking to the provider setting.

The AI work needs a provider: summary and clarity check, classification as improvement, change or incident, the milestone plan with hour estimates, tag suggestions, replies to mentions, milestone chat, idea-to-ticket drafts and the learning summary sent when a ticket closes. So do the assisted and centralized modes, the AI usage meter and the ideas dialog.

The provider is a service outside your site, reached through the Central Connect AI SDK. By default it is the sponsor's Central Connect AI service, where the support assistant is a subscription (soporte_bot_ia) with metered, paid AI usage; any provider exposing the same API can take its place. You connect it explicitly in the settings; nothing is sent before you do. A feature you have not activated shows a notice with the activation link, never a broken screen; your own monthly cost limit in EUR stops AI calls when reached.

Use cases

An agency supporting several client sites. Each client raises tickets on their own site; your hub estimates them and tracks the hours you invoice.
An internal IT team. Colleagues ask, the team plans milestones, timers measure the effort and reports show where the hours went.
Recurring maintenance. A template opens the monthly review ticket each month, milestones included.

Getting started

Install the module; it creates the content type, comment type, vocabularies and soporte role and places Support in the administration menu.
Grant view aiticketchat tasks, create aiticketchat task, comment aiticketchat task, transition aiticketchat task and favourite aiticketchat task to the people who raise tickets; act as aiticketchat support and track aiticketchat milestone time to the team; view aiticketchat reports and administer aiticketchat to those who run it.
Open Administration » Support » Settings (/admin/aiticketchat/ajustes): choose the backend mode, fill the Departments and Team tabs and adjust the board in the Sections tab.
Create your first ticket at /admin/aiticketchat/tarea/crear or import your history. Templates live at /admin/aiticketchat/plantillas, reports at /admin/aiticketchat/informes.

Requirements

Drupal 11.1 or newer, and the PHP version it requires (8.3).
Core modules: User, System, Node, Taxonomy, File, Options, Text, Field and Comment.
Central Connect AI SDK and Multi-Role AI Chatbot, both required.
PhpSpreadsheet (via Composer) for XLSX export, import and template; CSV does not need it.
Install with composer require drupal/aiticketchat.

Privacy and security

Tickets, comments and attachments are nodes, comments and managed files; milestones are a content entity. Exported configuration carries only sections, phases, transitions and settings. Credentials and the repository token live in State (overridable from settings.php), as does the monthly cost limit.
Every board, detail and JSON route checks a module permission; in local and assisted mode a support member sees the tickets of their assigned clients, an owner those they created, and administrators everything. Tickets are published nodes, so core's content access applies too: anyone allowed to view published content can open a ticket node directly. Add your own node access rules if tickets must stay private.
All POST routes carry the CSRF request header token; the AI endpoints are flood-limited per user.
In local mode nothing leaves the site unless you run the Monday.com importer or point the assistant at a GitHub repository yourself. In assisted and centralized mode the ticket text, answers and milestone plan go to the provider you named over OAuth2; the learning summary leaves on cron only once credentials are complete.
The read-only database areas the assistant may consult are off by default and permission-gated. The Monday.com importer uses the token you type for that request only.

Part of the Central Connect AI suite

The suite is a set of free modules sharing one provider connection. This module builds on the Central Connect AI SDK for header, notifications, entitlements and the AI gateway, and on the Multi-Role AI Chatbot for the chat sessions that become tickets.

Similar projects and how this one differs

Ticketing Management System gives visitors a portal to submit tickets with configurable types and fields, without estimation, planning or timers.
Support Ticketing System is an email-integrated helpdesk for earlier Drupal versions, without milestone plans, timers or provider modes.
Drupal PM is a general project management suite on Field API and Views; this module is narrower, a support board with AI valuation and a hub mode for agencies.
Work time adds play and stop buttons on projects; here timers belong to milestones and steps and feed cost reports.
Kanban organises work around work-in-progress limits per step; this module adds the ticket lifecycle, comments, mentions and estimates on top of a board.
AI is the provider framework for Drupal; this module reaches its provider through the Central Connect AI SDK instead.

Development

Parts of this module were written with the assistance of AI-based coding tools, as required by the policy on the use of AI when contributing to Drupal. All the code has been read, corrected, tested and is maintained by the human maintainer listed below, who takes full responsibility for it.

Maintainers and sponsor

Maintained by pedroromán. Tangram Consulting sponsors development and operates the default Central Connect AI service. Issues, ideas and contributions are welcome in the issue queue.

Project Management with Ticketing AI Assistant