🔎 AI Library Knowledge Hub

What is the AI Library?

The AI Library is a conversational AI layer built on top of your existing Flinn Library. Instead of manually searching and filtering, you ask questions in natural language directly against the papers in your library, and get answers grounded in the actual study content — each with citations back to the source papers.

Think of it as a digital librarian for your literature: a knowledge hub with an overview of every paper your team has already evaluated, answering your questions from your own quality-controlled evidence — not the open internet.

What can I do with the AI Library?

• Conversational Q&A — ask in plain language (e.g. "What is the current state of the art for technology X?") and get an answer within seconds, grounded in your library’s papers, with citations.
• Research across your whole library — ask against everything your team has accumulated over time, not just one search's results.
• Query by labels and extracted data — ask things like "List all papers with the label XYZ" and get every matching paper, drawing on the labels and extracted data your team has already captured.
• Cross-team self-service — let R&D, marketing, and medical affairs tap the clinical library directly, without a literature manager running every search for them.
• One place instead of many tools — consolidate work people were doing in general AI chat tools or reference managers into the system that already holds your trusted papers.
  
  

How is this different from ChatGPT, Copilot, or NotebookLM?

The AI Library only works from your own curated, quality-controlled library — papers your team has already acquired, screened, and often annotated. Answers are grounded in those actual papers and cited, not generated from the open internet. It lives inside Flinn alongside your searches and extractions that are always linked to any citations.

Trust, accuracy and transparency

Can I trust the answers — are they cited and verifiable?

Yes. Answers include citations to the specific papers they draw from, with a relevance indicator showing how well each source matches your question. What did it search, what did it use, and what did it miss? Each answer also shows you what it searched, how many papers it found, and how many it cited — so instead of trusting a black box, you can check its work. As with any AI, verify cited sources before using output in a regulatory submission.

Why didn't it find all my papers? It missed papers I know we have.

For each question, the AI Library retrieves the most relevant subset of your library rather than every paper, so a specific paper may not appear in a given answer even though it's in your library. You can always check how many papers the AI has pulled and how many of those where cited for the output. If you notice relevant papers are missing, use the in-app feedback option to help us improve retrieval.

Is it reliable enough for regulatory / CER work where completeness matters?

The AI Library is excellent for exploring your evidence, drafting, and answering targeted questions quickly. For workflows where completeness is critical (CERs, systematic reviews), use it alongside Flinn's structured Library searches and filters rather than as the sole way to identify every relevant paper.

Is there a validation package — can I rely on it in a regulated submission yet?

The AI Library is currently in beta, so a formal validation package isn't available yet. Use it as an assistant for exploration and drafting, and verify outputs before using them in regulated work.

How it works: papers, data and scope

How many papers can the AI Library work with?

Your Flinn Library can store extensive collections — some customers hold tens of thousands of papers. Per question, the AI Library reads the most relevant subset of your library rather than the entire collection at once, which is what keeps answers fast and grounded.

What content and data does the AI Library use?

It works from the papers in your organization's Flinn Library, using the full text where available, plus your labels and Extraction template data. It only accesses your own library — never other customers' data or the open web. Papers held only as abstracts, or without an attached full-text PDF, may not yet be fully usable, so the more complete your library's full text, the better the results.

If I just added a paper, will the AI Library use it right away?

Adding a paper makes it available to the AI Library, but availability doesn't guarantee it will appear in the answer to every related question — each answer draws on the most relevant subset of your library for that specific question.

What languages does the AI Library support?

Ask in whatever language you prefer, and the AI Library answers in that language. It can also work with non-English papers in your library. As always, results depend on the quality of the text available for each paper.

Access, setup and availability

Is the AI Library generally available?

It's currently in beta: live in production and in daily use by selected customers, while we keep enhancing it based on their feedback. Speak to your Flinn contact about getting early access.

How do I get set up, and what do I need?

Your Flinn Customer Solutions contact enables the AI Library for your organization. Because the AI works from your library, you'll get the most value once you already have a meaningful set of papers — ideally with full text — in Flinn.

Can colleagues outside the core literature team use it?

Flinn offers read-only viewer roles so people outside the core literature team (e.g. marketing, R&D, medical affairs) can access the Library, therefore opening AI querying to viewer roles is an active area of work.

Data security and privacy

Is my data secure and compliant?

The AI Library runs on the Flinn platform, built for regulated MedTech use, and follows Flinn's existing platform security and compliance standards.

Does my content train Flinn's AI, or get shared with other customers?

No. The AI Library works only from your organization's own library. Your content is never shared with other customers or the open web.