AI is showing up everywhere in government affairs. It’s incorporated into your tools, your team’s workflows, and increasingly in the bills and regulations you track. But the terminology can feel like a moving target, especially if your background is in policy, not tech.
This glossary highlights the AI terms you’re most likely to encounter in your work. Bookmark it, and we’ll keep it updated as the terminology (and the technology) evolves.
Core AI Concepts
Artificial Intelligence (AI): The broad field of building software that can perform tasks — like analyzing text, spotting patterns, or generating language — that would otherwise require human judgment. Everything else in this glossary is a subset or application of AI.
Machine Learning (ML): A method for building AI: instead of a programmer writing explicit rules, a system learns patterns from large amounts of data. Most modern AI tools, including the ones showing up in public affairs software, are built on machine learning.
Large Language Model (LLM): A type of AI trained on massive amounts of text so it can understand and generate human language. LLMs are the technology behind most of today’s AI writing and analysis tools, including AI assistants and agents built into platforms like Quorum.
Generative AI: AI that creates new content — text, summaries, images, or draft messages — rather than just analyzing or classifying existing information. When a tool drafts a stakeholder email or summarizes a hearing transcript for you, that’s generative AI at work.
Natural Language Processing (NLP): The branch of AI focused on understanding and working with human language — parsing a bill’s text, detecting sentiment in a press release, or answering a question typed in plain English rather than code or a rigid search syntax.
Prompt / Prompt Engineering: The instruction or question you give an AI tool. “Prompt engineering” refers to writing that instruction clearly enough to get a useful answer — for example, asking Quincy a specific question about a bill’s fiscal impact rather than a vague one.
Hallucination: When an AI generates information that sounds plausible but isn’t accurate — a fabricated citation, a misstated bill number, a summary that misses the point. It’s one of the biggest reasons AI tools built for public affairs are trained on verified legislative data rather than the open internet, and why final review by a policy professional still matters.
Training Data: The information an AI system learns from before it’s put to use. The quality, breadth, and relevance of training data directly shapes how trustworthy an AI tool’s output is — which is why an AI trained specifically on years of legislative and stakeholder data will generally outperform a general-purpose tool on public affairs questions.
Agentic & Applied AI
Agentic AI: AI that doesn’t just answer a question but takes multi-step action toward a goal — pulling data, applying logic, and producing a finished output with minimal back-and-forth. This is a step beyond a simple chatbot: think building an entire campaign or triaging a full bill list, not just answering one prompt at a time.
AI Agent: A specific AI tool built to handle one defined task end-to-end. Quorum’s Bill Tracker Agent, for example, applies your organization’s own priorities to every bill in your list and surfaces a recommended stance, priority level, and reasoning — with final sign-off always left to you.
Human-in-the-Loop: A design principle where AI produces a recommendation or draft, but a person makes the final call. It’s the difference between AI replacing judgment and AI supporting it. This is central to how AI should function in a field where the stakes of getting it wrong are high.
Sentiment Analysis: Using AI to gauge tone or position — positive, negative, mixed — across a large volume of text, like social media posts, testimony, or press coverage. Instead of reading hundreds of mentions of a bill one by one, sentiment analysis surfaces the overall trend for you.
Legislative Triage: The process of sorting and prioritizing bills based on relevance and urgency to your organization — traditionally a manual, time-consuming read-through. AI-powered triage tools apply your team’s stated priorities against incoming legislation automatically, so your team’s attention goes to what matters most first.
Institutional Knowledge (Search): Your team’s own accumulated history — meeting notes, staffer conversations, advocate stories, past strategy — made searchable alongside live public policy data. Rather than digging through old emails or asking around, an AI tool like Quincy lets you ask a plain-language question and get an answer sourced from your own team’s record.
Quorum’s AI Toolkit
The terms above describe AI in general. Here’s what they look like inside Quorum, specifically the named tools and features you’ll actually click on.
Quincy: Quorum’s AI assistant and the umbrella for everything below. Quincy connects to your Federal, State, Local, Grassroots, Stakeholder, and PAC data in one workspace, so you can ask a plain-language question and get an answer sourced from public policy data and your own team’s institutional knowledge.
Natural Language Search: Search that understands the concept behind your question, not just the keywords in it. Instead of guessing the right search syntax, you can ask Quincy a question the way you’d ask a colleague and get a precise answer back.
Policy Analysis: Ask a plain-language question and Quincy scans bills, hearings, and government documents across every jurisdiction you track, returning a fast, sourced answer instead of a pile of documents to read yourself.
Dialogue Analysis: Quincy’s read on lawmaker sentiment — pulling from social media, committee transcripts, and documents across all 50 states and Capitol Hill daily. It is sentiment analysis (see above) applied specifically to the conversations your policy team needs to track.
Stakeholder Notes Analysis: Surfaces insights from your team’s own past interactions as easily as asking about a public bill, bridging your private meeting notes and institutional knowledge with live public policy data.
Bill Tracker Agent: An AI Agent that automates legislative triage: it scans every bill, ranks it by relevance, and tags a suggested position grounded in your organization’s own stated priorities — with final sign-off left to you.
Meeting Prep Agent: Briefs you before every meeting on an official’s recent activity and statements, your team’s past engagement, advocacy history, and PAC history — so you walk in with full context instead of digging for it beforehand.
CRM Agent: Speak or type your notes the moment a meeting ends, and the Agent maps them into your data automatically. It turns a conversation into lasting institutional knowledge instead of a note that gets lost.
Policy Comms Agent: Turns a company announcement into ready-to-send legislative outreach automatically, so consistent storytelling with the Hill becomes the default rather than something your team has to build from scratch each time.
AI Governance & Risk
Responsible AI: An approach to building and using AI that accounts for accuracy, fairness, transparency, and accountability — not just capability. For government affairs teams, this term shows up both as a vendor evaluation criterion and, increasingly, as a policy area lawmakers are actively legislating.
Explainability: The degree to which an AI system’s output can be traced back to a clear reason. Why did the tool flag this bill as high priority, or draft this recommendation? Explainable AI shows its reasoning rather than delivering a black-box answer, which matters both for trusting the tool and for defending a recommendation to leadership.
AI Governance: The policies, oversight structures, and rules — internal to an organization or imposed by law — that govern how AI is built, deployed, and monitored. This is a fast-growing area of state and federal legislation, making it a term you’ll increasingly track as a policy issue, not just a product feature.
Data Privacy (in AI): How an AI system handles the information it’s given access to. Whether your organization’s internal notes and strategy stay within your own data environment, or are used to train models for other customers. Worth asking any AI vendor directly, since practices vary widely across the industry.
Have a term you’d like us to add? Let us know, and this glossary will grow as AI for your public affairs team does.