Q&A: Inside Maryland DoIT’s “Demos, Not Memos” Approach to Artificial Intelligence

STATETECH: What’s your mandate as Maryland’s AI/ML program manager, and how do you help agencies get from idea to impact?

LAUREN MAFFEO: I work with CIOs across agencies to scope use cases and confirm whether they’re truly “AI-shaped” problems. If they are, I help them access DoIT’s sandbox environments, identify the right models and architecture, and secure the resources — like design support — to launch pilots. When pilots succeed in testing, our goal is to help teams scale them into production. I also run our statewide AI community of practice and weekly office hours so any Maryland civil servant can learn directly from peers who are already using AI in their daily workflows.

We want the people doing the work to show how they’re using AI in daily workflows — demos, not memos.

STATETECH: I like that — “demos, not memos.” Why does that matter for public sector AI?

MAFFEO: AI discussions can be nebulous and detached from day-to-day work. We flip that. We minimize our own airtime and elevate practitioners who can show practical examples. That builds confidence and clarity, shortens learning curves and helps people build with the tools they already have, rather than waiting on long procurement cycles.

STATETECH: How is Gemini available to state employees today, and how did you approach the rollout?

MAFFEO: Agencies can use Gemini through Google Workspace if their CIO has turned it on. We enabled that option in March after a cross-functional team — data, privacy, security and AI — vetted Gemini in Workspace. We didn’t mandate activation; we wanted CIOs to lead and for adoption to be consent-based. If an agency enables Gemini, the capabilities are available to their users right away.

STATETECH: What early use cases are proving most helpful?

MAFFEO: Transcribing meetings is a deceptively powerful one — freeing staff to listen and engage while creating searchable notes. I often upload those notes to Google NotebookLM and query them later to find tasks or decisions without sifting through it manually. We’ve also seen ad-hoc wireframing during meetings to accelerate feedback. Another agency is exploring a Gem [a custom Gemini assistant] for fleet management: on a state phone, staff could request real-time recommendations for oil changes or brake service nearby while on official business. And a colleague uses Gemini Deep Research for sentiment analysis — ingesting raw data to quantify dissatisfaction, then probing why to inform what tech debt to tackle first.

READ MORE: Chatbots top the list of government AI use cases.

STATETECH: What emerging or future tools are on your radar?

MAFFEO: We’re seeing more interest in Google application programming interfaces to connect Gemini with other tools, as well as requests around AI Studio, which isn’t included with Gemini — you procure it separately — so I’d call that aspirational. But growing interest suggests it’s worth formal vetting for a wider offering.

STATETECH: How does this align with DoIT’s broader modernization plans?

MAFFEO: DoIT recently released an IT master plan — the vision for more cost-effective, efficient technology over the next three to five years. You can’t do that without AI. We’re consolidating services and contracts and identifying where it makes sense to add infrastructure. The aim is to offer an AI-ready tech stack: Any agency can come to DoIT with needs, and our team will help craft a bespoke AI stack — ethics-first, security-vetted — that helps them do their jobs better.

Because we’ve already written a responsible AI policy and vetted privacy and security, agencies can have confidence in the AI we offer.

DIVE DEEPER: Governments must brace for AI’s transformational impact.

STATETECH: Governance and ethics often make or break trust. How are you addressing that?

MAFFEO: We started with ethics and responsible use, then built strategy and enablement on top of that. A four-team initiative at DoIT vetted the privacy and security of the AI tools we provide. That lets agencies move faster, knowing they’re building on a compliant, trustworthy foundation.

STATETECH: What role has leadership played in sustaining momentum?

MAFFEO: Strong top-down leadership is essential. Information Technology Secretary Katie Savage set a transformation vision that includes the hard, unglamorous work — legacy migrations, contract alignment — plus the enablers for AI. That clarity about DoIT’s value to agencies, and commitment to improving it, is why I joined the AI enablement team. I’m proud that Maryland started with ethics, invested in workforce capability and is now elevating the backbone services we provide to the next level.

STATETECH: Where will Maryland’s AI program be in 2026?

MAFFEO: AI will be part of the fabric of how Maryland delivers services. We’re evolving from an enablement team to offering a set of AI services that agencies can assemble into the stack they need — confident that the tech has been responsibly vetted and is ready to scale from pilot to production.