Andrew Spiegel, Chair of the WPA, opened the final session of the three-part AI Training Webinar Series by welcoming participants from across the globe. He reiterated WPA’s role as the world’s largest umbrella patient organization, representing more than 700 member groups in 138 countries, and emphasized WPA’s continued commitment to empowering patients and patient organizations through accessible and practical digital tools.

Andrew highlighted the multilingual features available through Wordly, which once again provided real-time chat and audio interpretation in more than 50 languages. He stressed that accessibility remains central to WPA’s mission, ensuring that geography and language never limit participation. He also noted that this final session would build on the foundations laid in earlier webinars and guide participants toward adopting advanced, practical AI solutions within their own organizations.

Leading the session, Ravi Ruparel of Platform Worldwide explained why this final webinar was essential for patient organizations. While the first two sessions focused on foundations and practical tools, this session moved into the next stage: creating custom AI solutions, improving organizational readiness, and preparing for future changes in the digital landscape.

Ravi emphasized that patient advocacy groups continue to face growing demands with limited resources. AI offers a way to expand capacity, strengthen communication, and improve community engagement. The presenters highlighted that organizations do not need technical backgrounds to use AI effectively. Instead, small steps consistent experimentation, better prompt strategies, and adopting accessible tools can lead to significant impact.

The session began with a combined recap of key lessons from the earlier webinars, grounding the audience in the holistic progression of the series.

Webinar 1 (24 July 2025)
The first session introduced the foundations of AI, explaining how rapidly AI is shaping patient behaviour, expectations, and information-seeking patterns. Ravi stressed that patient organizations need to produce clear, structured, and accessible content so AI systems can reference it safely. Participants learned how different types of AI tools serve different purposes from chat assistants to translation and content generation and how clarity in prompting leads to higher-quality results.

Webinar 2 (24 September 2025)
The second session moved from theory to practice. It demonstrated how AI can support advocacy workflows through letter writing, podcast creation, translation, explainer videos, and social media content. The importance of building a prompt library and treating AI as a “thinking partner” was reinforced. Safety and ethical use were emphasized to ensure patient trust. The presenters also previewed that Webinar 3 would introduce custom tools and automation.

This recap set the stage for the advanced focus of the final session.

Click here to watch Webinar 1 recording.
Click here to watch Webinar 2 recording.

Ravi updated participants on major advances in AI over the last 60 days. He explained that new systems now include stronger safeguards around health-related topics. These updates help AI tools respond more responsibly when users ask about diagnosis, treatment, or urgent medical situations.

AI models now:

• Decline unsafe medical questions more reliably.
• Route users to safer language and general guidance.
• Emphasize “consult a clinician” in high-risk contexts.
• Provide more structured information.
• Avoid false certainty.

These changes make AI safer to use but also highlight why patient organizations must publish accurate, culturally inclusive, AI-readable content. If trustworthy information is not publicly available, AI systems cannot provide relevant or safe responses.

Ravi emphasized that AI learns from publicly available information. This means patient organizations have a direct influence on the safety, cultural relevance, and accuracy of AI-generated content. When patient organizations publish high-quality resources, they elevate the standard of information available to entire communities.

Ravi outlined several key actions patient organizations should prioritize:

• Updating websites to include structured headings, metadata, summaries, and alt text.
• Publishing information in multiple formats (HTML, PDF, audio).
• Simplifying complex medical explanations.
• Adding safety disclaimers such as “information, not medical advice.”
• Providing multilingual translations wherever possible.
• Keeping pages updated to avoid outdated or misleading information.
• Sharing lived-experience stories that reflect diverse cultures and communities.

These steps ensure that AI tools present helpful, safe, and culturally sensitive information drawn from reliable patient-driven sources.

A central part of the session was a live demonstration showing how quickly patient organizations can create practical AI tools without coding experience.

Ravi walked the participants through building a simple, real AI chatbot in just a few minutes. The demonstration showed how a chatbot can scan a website, read uploaded PDFs, and answer questions using only approved materials.

The process included:

• Pasting the organization’s website URL.
• Uploading PDFs or resources.
• Setting guardrails to prevent unsafe responses.
• Generating the chatbot.
• Copying an embed code for use on Wix, WordPress, or any website platform.

The aim was to show that powerful tools can be built quickly to improve patient access, reduce repetitive emails, support multilingual communication, and provide consistent, accurate responses.

Ravi reinforced that safety is non-negotiable. The chatbot demonstration included examples of guardrails that organizations should use:

• “This chatbot provides general information only and does not give individual medical advice.”
• “If unsure, respond: ‘I’m unable to answer this safely. Please contact a healthcare professional.’”
• “Always answer using the approved sources provided.”
• “Decline questions that fall outside validated information.”
• “Use inclusive, simple language suitable for low literacy.”

These measures help patient organizations offer accessible assistance while protecting patient trust and avoiding health misinformation.

The session showcased how AI can help organizations create compelling visual and educational content. Examples included:

• “What patients wish policymakers knew.”
• “Medication shortages: what to do.”
• “New research explained simply.”
• “My journey with [disease] | lived experience story.”
• “The truth vs myths about [topic].”
• “What to ask your doctor.”
• “Event announcement.”

Such content can improve outreach, raise awareness, and make health information easier to understand and share across communities.

Ravi outlined where patient organizations currently stand and what lies ahead.

Many organizations today are still at Stage 1 of AI adoption, using AI for chat and basic content creation. However, the next stages offer greater efficiency and impact:

Stage 2: Automation for surveys, volunteer management, reminders, newsletters, and fundraising workflows

Stage 3: Internal knowledge assistants that make policies, FAQs, and archives easily accessible to teams

Stage 4: AI agents that support multi-step tasks such as preparing reports, analyzing meeting notes, drafting policy briefs, and managing multilingual communication

He shared predictions for the next 12 months, noting that content automation will become standard, multilingual support will expand reach, website chatbots will become expected, and ethical frameworks will become essential for credibility.

Participants were encouraged to continue experimenting with AI after the series. Key suggestions included:

• Trying one new AI experiment each week.
• Building a prompt library.
• Updating websites for AI-readiness.
• Adopting small automations.
• Collaborating with neighbouring patient groups.
• Sharing knowledge openly.
• Placing safety, ethics, and inclusivity at the centre of all digital work.

The presenters reminded participants that confidence comes from steady, practical experimentation.

The interactive portion of the session was lively and highly engaging. Participants used the multilingual transcription and audio tools to join discussion from Europe, the Middle East, Africa, Asia, and the Americas. Questions focused on privacy, cultural inclusivity, guardrails, building custom chatbots, and adapting materials for communities with low digital literacy.

Attendees shared examples of how they planned to use AI in patient education, policy advocacy, and volunteer coordination. The presenters encouraged everyone to take one concrete action this week whether testing a chatbot, updating a webpage, or creating new AI-assisted visuals.

Hussain Jafri, CEO of the WPA, delivered the closing remarks for the final session. He thanked the speakers and participants for their sustained engagement across all three webinars and emphasized that patient empowerment is central to WPA’s mission. Hussain noted that the tools demonstrated throughout the series are practical, accessible, and ready to use today. He encouraged organizations to continue experimenting with AI to improve patient reach, communication, and advocacy impact, while always maintaining high ethical standards.

The three-part series concluded with a message of unity, confidence, and forward momentum.

The “Empowering Patient Advocacy with AI” series has equipped patient organizations with a practical roadmap for adopting AI responsibly. From foundational understanding to hands-on tools and advanced custom solutions, participants now have the skills to integrate AI into their advocacy, outreach, education, and organizational workflows.

The World Patients Alliance and Platform Worldwide thank all participants for joining this series. By embracing safe, inclusive, and ethical AI practices, patient organizations can strengthen their impact and expand support for communities worldwide.

The future of patient advocacy will be shaped by those who adopt new tools early while staying human-centred and committed to patient trust.