AI initiative working group kick off meeting notes, 12 - 13 August 2026

Hello everyone,

We have relaunched the AI initiative, and last week we held our first async meeting in the #wg-ai channel on Slack to get things moving under clear ownership. Thank you to everyone who took part. Below are the notes and, more importantly, how you can get involved, because this is very much a shared effort.

If you are new to this, our direction is guided by Mautic’s AI Manifesto. The short version is that we want Mautic to be a home for privacy respecting, ethical AI in marketing automation, where your data stays under your control.

What we set out to do

Phase one focuses on a maintainable AI foundation plus our first two features, with data sovereignty protected throughout. The foundation would be a dedicated AI layer in core built around Symfony AI (GitHub repository here), so we stay aligned with the framework Mautic already runs on, with core owning the shared services and plugins delivering specific marketing use cases.

What we discussed

Why now. There was agreement that being open source makes privacy, sovereignty and governance especially important, and that building these in now will help the architecture scale as we add more AI later. A strong point raised was that features have to be genuinely useful to end users from the start, reducing the everyday friction of ‘how do I do this’, rather than being clever but over engineered.

The technical approach. Clear support for building on Symfony AI. Drawing on experience building the MCP bundle, the point was made that once there is a shared foundation and a clear abstraction, integrations become far easier, instead of every feature solving the same problems again. The emerging agreement is that core owns the foundation and shared services, and plugins deliver the specific use cases.

Privacy, security and governance. We want the foundation to enforce the safeguards, PII handling, local processing options, approved provider policies, prompt injection protection, output validation and permission controls, rather than leaving them to each plugin. This topic did not draw much discussion in the meeting, and since it is central to everything, we will open it again with concrete proposals. Your input here would be especially valuable.

The first features. We propose two, both starting in advisor mode where you stay in control. A content advisor inside the GrapesJS builder for emails and landing pages, and a reporting and analysis insight assistant that explains what happened, when and why. On the content side there was a clear preference for starting from a prompt that generates a full draft you can then refine by responding, rather than a limited inline text tool.

How we organise. We are running this as a small, focused, community led effort. We plan to develop a costed plan for accelerating the first phase, and organisations that rely on Mautic can help fund it through sponsorship alongside grant applications, so nobody has to carry it alone. We heard that people are keen but time is tight, which is exactly why we are pursuing funding to back a focused team rather than solely relying on the goodwill of volunteers.

Work we are building on

We are not starting from nothing. Several community members have already built valuable prototypes and tools, including AI Email Sections and mautic-cli from José Fernandes and Bloomidea, and a set of AI prototype bundles from Frederik Wouters at Dropsolid, including a reports bundle and a console bundle. There is also real experience in the group from building an MCP bundle and earlier developments of a ChatGPT integration in the builder managed by Ricardo at crafting.email. We will learn from all of this.

What happens next

Here is what we are doing now.

  • Creating and sharing the costed first phase plan and opening sponsorship conversations with organisations that rely on Mautic.

  • Reconvening the working group around the plan, with a short regular sync backed by async discussion on concrete decisions, and announcing the next meeting date.

  • Drafting an architecture outline for the AI foundation on Symfony AI, covering the shared services, the integration layer and the plugin extension points.

  • Writing up the two first features in more detail, and reopening the privacy and governance discussion with concrete proposals.

  • Setting up a public task board so anyone can see what is happening and pick something up. Link to follow here.

How you can get involved

We need builders, designers, testers, privacy reviewers and writers, and there is room for both paid and voluntary contribution. If you would like to help, please reply here and tell us where you would like to contribute and roughly how much time you could give.

If your organisation might sponsor the work or contribute people to it (both of which will be recognised formally), we would love to hear from you.

There is a place for you in this, whatever your skills. Thank you again to everyone who joined the first meeting, and let’s build the future of ethical, privacy respecting marketing automation together.

Next meetings

We’re planning to meet fortnightly, asynchronously on Slack. Join us there, get involved when you can. The meetings are open for 48 hours.

1 Like