GovComms.ai is my flagship AI solution, and in this episode, I’m pulling back the curtain on how it was built. From its first iteration as PolBot to a working domain-expert system, this is the story of combining 25 years of communications expertise with a governance-first approach to create trusted AI for public-sector communications across the EU.
From PolBot to GovComms.ai: The Origin Story
It didn’t start as GovComms.ai. The first version was called PolBot, and I developed it before and during my Harvard programme in artificial intelligence and public policy in 2025. It wasn’t the finished product — not even close. But it helped me validate something I had seen throughout my entire career: public-sector communications professionals need AI that actually understands the real work they do, the standards they operate under and the accountability that comes with it.
PolBot became the starting point. Nearly a year of research, specification, building, testing and refining later, GovComms.ai is now a working minimum viable product.
What GovComms.ai Actually Does
This is not a generic chatbot. It’s a domain-expert AI system designed around the actual functions and tasks of public-sector communications. There are six core functions and 18 essential tasks built into the system:
- Strategy and Planning — communications plans, stakeholder mapping, message development
- Media and Public Affairs — press releases, media lines, Q&As, opinion pieces
- Digital and Content — social posts, website copy, newsletters
- Governance and Accountability — FOI responses, parliamentary questions, ministerial briefings
- Public Engagement — consultations, awareness campaigns, community communications
- Analytics and Insight — measurement frameworks, insight reports, campaign evaluations
These aren’t random prompts thrown together. Every task is built around domain-specific templates, curated best practice and professional frameworks developed across 25 years of work in communications, journalism, public policy and education.
The Architecture: RAG, Retrieval and Organisational Context
GovComms.ai combines a domain-adapted model with retrieval-augmented generation — RAG. That means the system retrieves relevant material at the point of use, pulling selected source documents, frameworks and approved organisational context directly into the task. An organisation can layer in its own strategies, policies and messaging, stored in an authorised environment with access controls and scoped retrieval. That material stays within permitted tasks — it doesn’t feed into a generic public model.
The result is a clearer, more consistent drafting process than starting from a blank prompt in a general-purpose AI tool. And crucially, it supports better review — because in government communications, fluency isn’t enough. Drafting needs to be accurate, traceable and defensible.
Why Mistral and Why European?
The underlying model is from Mistral, a European AI company. That choice was deliberate. I wanted a European-developed model at the centre of a product designed for European public institutions, with greater flexibility over deployment and architecture. It reflects something I believe strongly: the model that performs best on a benchmark isn’t automatically the right model for a particular product. Model choice is also a choice about values, jurisdiction, data and dependency.
Governance-First by Design
Human oversight, source-grounding, documentation, traceability and responsible use are central to how GovComms.ai is built — not added on afterwards. The tagline is AI for Public Good. That’s not decorative. It’s the product strategy. Government communication affects trust, participation, access to services and the relationship between institutions and citizens. If AI is entering that space, it should strengthen the professionals responsible for it — not separate them from their responsibility.
If you want to build your AI literacy and skills, I can help
📚 Grab AI Playbook on Amazon, the perfect podcast companion.
📰 Subscribe to Human in the Loop, my LinkedIn Newsletter
🎓 Take the CPD Accredited AI Playbook Diploma
📞 Book a call at digitaltraining.ie
Key Statistics
– GovComms.ai is built around 6 core communications functions and 18 essential tasks.
– Nearly a year of research, specification, building, testing and refining went into the working MVP.
– The product draws on 25 years of domain expertise in communications, journalism, public policy and education.
– The underlying model is from Mistral, a European AI company, chosen for its European provenance and deployment flexibility.
– The first iteration, PolBot, was developed before and during Joanne’s Harvard programme in AI and public policy in 2025.
About the Podcast Host
AI and Communications Specialist | Author | Podcaster | International Speaker
Joanne Sweeney is a recognised leader in digital marketing and digital transformation, specialising in the practical application of AI to enhance communication strategies. The founder of Digital Training Institute, Joanne, empowers leaders and communications professionals to leverage AI responsibly and effectively, ensuring transparency, trust, and tangible results.
She completed the Harvard Kennedy School Executive Programme in AI and Public Policy this summer before launching her fourth book AI Playbook, in September. Joanne has also published two editions of Public Sector Marketing Pro, which were preceded by her first title, Social Media Under Investigation Law Enforcement and the Social Web.
Her books have become go-to resources for governments, public sector and policing agencies, as well as marketing and communications professionals, all navigating the challenges of the changing Internet.
Joanne is the host of the AI SIX Podcast, a daily six-minute podcast that provides actionable insights, expert interviews, and practical advice to help professionals navigate the new world of work.
