Please note our content disclaimer in relation to blog posts.
Membership organisations invest considerable time in producing guidance, research, events, and learning resources. But having that expertise available does not guarantee that members will find it, recognise its relevance or put it to use.
In the third session of our AI in Membership series with Umbraco, we explored how connected content, AI-assisted workflows and behavioural insight could help close that gap.
Our first session examined governance, maturity, and organisational readiness. The second explored practical AI capabilities within the CMS. This time, the focus moved to how those capabilities can work together to create a better member experience.
The discussion brought together Phil Whittaker, Staff Engineer focusing on AI at Umbraco; Jon Whitter, Head of Development Services at Cantarus and Umbraco MVP; Nick David, Head of Growth at Cantarus; and Dani Barker, Marketing Services Manager at Cantarus.
Watch AI in Membership Part 3: The Future Member Experience for the full discussion, live demonstration in Umbraco CMS and answering audience questions on personalisation, permissions and implementation.
How can AI help make the value of membership easier to discover, easier to access, and more relevant to the individual? That question framed the session’s discussion of the future member experience.
An organisation may offer substantial value through its learning, guidance, and professional expertise. If members cannot find the right resource, or understand how it helps them, they may not recognise that value.
“Having content and someone experiencing value from that content are two very different things.” - Nick David
MemberWise’s Ultimate Guide to Member Value distinguishes between the value an organisation delivers, and the value members perceive. As Nick highlighted during the session, valuable expertise can remain underused if members cannot recognise its relevance or find a route into it.
As conversational tools become more familiar, members may increasingly arrive with a question or task. They want help achieving something, without first having to understand how an organisation has arranged its website.
This connects the digital experience with acquisition and retention: prospective members need to see why joining is worthwhile, while existing members need opportunities to experience that value throughout the year. Improved discovery can support those goals; its effect still needs to be measured.
One example discussed was an annual conference. An organisation might have hours of valuable recordings, alongside speaker biographies, session descriptions, research and related guidance. Once published, much of that material can gradually disappear into an archive.
Starting with a recording as the core asset, AI could help an editor draw on related, trusted sources to extract key themes and develop summaries, articles or learning resources that give members different routes into the subject.
The opportunity becomes more valuable when those resources connect to an organisation’s professional development framework. A session might support a particular competency, career stage, or learning objective. Making that relationship explicit can help members understand where it fits into their development.
Taxonomy matters here. Classifying a resource only as an article about leadership tells us relatively little. Connecting it to the skills it supports and the audience it serves creates a more useful foundation for discovery and recommendations.
The panel also discussed the possibility of packaging existing expertise into learning products. That is a potential commercial application, dependent on the organisation’s content, learning offer, and business model.
Personalisation only helps when it makes the experience more relevant. Addressing someone by name tells them little about whether the content is useful to their needs.
A more meaningful experience might connect someone’s career stage or learning interests with suitable resources. That requires a good understanding of both the member and the content available to them.
Dani introduced the AI-assisted conversational interface shown within Umbraco Engage designed to make insight more accessible to the team members planning and managing member communications. The discussion explored how access to content and behavioural data could help teams investigate why users drop out of a journey or fail to take the next step and ultimately, feed back into the content creation process.
That feedback loop becomes especially useful when it connects directly into the way content is created, structured and maintained. If behavioural insight can show where members need more support, automation can help teams respond by turning existing expertise into clearer, better connected resources.
Jon demonstrated how Umbraco AI and Umbraco Automate could work together using a podcast website example originally presented by Umbraco’s Matt Brailsford at Codegarden.
The example began with an episode page containing basic details and an uploaded MP3. Saving the page triggered an automated process that used an AI agent to transcribe the recording and generate content in the relevant CMS fields.
The agent’s instructions defined what it should do, which fields it should populate and when it should leave existing content alone. Its configuration also included tone-of-voice guidance.
This showed how a recording could become usable page content within the CMS, reducing the manual steps an editor would otherwise need to complete.
It also exposed a limitation: speaker names were misspelt. Jon explained how further integration with a CRM containing authoritative speaker records could provide better source information. That connection was discussed as an extension to the demonstration.
The example is a useful reminder that connecting sources improves the context available to AI, while editorial checking remains essential.
What this means for the member experience is not simply more content, produced faster. It is the chance to turn a single piece of expertise into richer, more useful resources that are easier to find, understand and act on. With the right taxonomy and source data, AI-assisted workflows can help make content more discoverable through intelligent tagging, clearer summaries and stronger links to related guidance, events or learning pathways. For members, that can mean less time searching and more relevant routes into the value the organisation already provides.
The previous example showed what is possible when AI is connected to a specific content workflow inside the CMS. Phil then widened that view, explaining two complementary approaches: AI in Umbraco and Umbraco in AI. The first brings AI capabilities into the CMS to support tasks such as creating content and asking questions about it. The second enables external AI assistants to work with Umbraco alongside other connected systems.
Model Context Protocol, or MCP, provides a way for those assistants to connect to tools and information. In the scenarios Phil described, an assistant could work with source documents and CMS content together, helping an editor assemble material without repeatedly moving between systems.
This extends the opportunity beyond generating a single field of text. It allows content, context, and actions to be connected across a wider working process. The available actions depend on the connectors, configuration and access granted.
Phil described Forms and Engage MCP integrations as being in beta at the time of the webinar. The capabilities available to a team depend on the tools and integrations in place.
Whilst an agent may assemble source material, generate a draft and populate fields, editors still need to be able to review and refine the result, with clear responsibility for approval. Defining those steps is part of designing the workflow, so time saved on production does not come at the expense of useful, trustworthy content.
Dani highlighted the role of Umbraco Workflow in managing approval stages within the back office. For marketing teams, those stages help maintain editorial standards as AI-assisted production increases, covering tone of voice, accuracy and any specialist review required before publication.
“In the context of AI assisted content production, it’s essential that you keep that human layer in.” - Dani Barker
An audience question brought an important distinction into focus: can AI-powered search find member-only content while respecting access permissions, and would it replace existing personalisation?
The panel explained that these are different considerations. Personalisation can change what is presented or recommended. Access controls determine what someone is entitled to see. A tailored content block should not be assumed to provide the same protection as enforced permissions.
Search within a membership website can be designed to respect a signed-in user’s access rights. That requires appropriate indexing and controls to prevent protected material appearing in results or answers for users without permission.
Public AI discovery is a separate context and should not be assumed to have a member’s authenticated access.
The speakers were also clear that the back-office AI agents shown in the session were designed for editorial work. Their access to content and editing capabilities means they should not simply be exposed as member-facing chatbots. A conversational service for members needs its own carefully defined access and behaviour.
That distinction helps organisations explore better discovery while protecting the expertise that forms part of their membership offer.
For many organisations, an existing content workflow offers a manageable place to begin. Choose a valuable recording or report that members currently struggle to find or use and identify the audience and task it should support. A shorter summary, clearer classification or connection to related guidance may provide a useful first improvement.
Before automating that process, establish the source material the AI can use, the tone and content standards it should follow, and the person responsible for checking the result. Map the steps already taken by your team so that the proposed workflow addresses a real bottleneck and fits the tools and capacity available.
Define success in terms of the intended improvement. That might include less editorial time spent preparing an approved resource, more members discovering relevant guidance, or better progression through a learning journey. Behavioural data can help assess use; member feedback can help explain whether it was useful.
The opportunity discussed in this session is to make an organisation’s expertise easier to access and apply. Start with a member need, connect the content and processes required to support it, and use what you learn to decide where to go next.
Watch AI in Membership Part 3: The Future Member Experience for the full discussion, live demonstration in Umbraco CMS and answering audience questions on personalisation, permissions and implementation.
Missed the previous articles and webinars in this series? Take a look on our content hub and make sure you're caught up.
