You probably already use AI with PDFs, reports, research papers and evaluation documents.

But what about the books you use to understand evaluation methodology, qualitative research, systems thinking or international development?

Gemini Notebook can now use eligible books purchased through Google Play Books as sources in a notebook, alongside your own documents and research materials. It can then answer questions using those sources and provide citations back to the relevant material.

What You Will Learn

In this tutorial, you will learn how to:

• add an eligible book to Gemini Notebook;
• combine the book with your evaluation documents;
• ask questions grounded in the source;
• use the book to review an evaluation design;
• use it with a Theory of Change and qualitative research;
• check AI-generated citations; and
• use the workflow for evidence-grounded reporting.

Step 1 — Open Gemini Notebook

Start with a notebook dedicated to one research task or evaluation.

For this tutorial, use:

Youth Employment Evaluation — Methodology Review

A focused notebook is easier to work with than one giant notebook containing unrelated projects.

Step 2 — Add Your Evaluation Documents

First, bring in the documents you actually need for the evaluation.

For example:

• Evaluation TOR
• Theory of Change
• Baseline Report
• M&E Framework
• Relevant research papers

Wait until the documents appear in the notebook’s source list.

Step 3 — Add an Eligible Book

Now let’s add the professional book you want to use as a methodological reference.

Then return to the Sources panel and check that the book appears alongside your other sources.

This is the workflow described for the new book-source capability.

Your source list might now look like this

Now you have something more useful than a book sitting separately from your project material.

Step 4 — Check Whether the Book Qualifies

There is an important catch.

This feature does not mean that Gemini Notebook can automatically read every ebook you own.

The source explains that the specific book capability applies to eligible titles purchased through Google Play Books. The initial catalog was reported to include more than 100,000 titles.

If you have a PDF or EPUB yourself, you can still use it as a regular source where supported.

Step 5 — Test the Book Before Doing Serious Analysis

Don’t jump straight into a complicated evaluation question.

First, test whether the book is being used as a source.

Now look carefully at the answer.

Then:

Make this a habit.

A citation makes an AI response easier to verify. It does not automatically make the interpretation correct.

Step 6 — Use the Book to Review a Theory of Change

Now we can do something much more interesting.

Make sure your Theory of Change and your book are available as sources.

Notice the difference.

Instead of asking:

“Is my Theory of Change good?”

you are asking AI to apply a specific methodological lens to your Theory of Change.

Step 7 — Review an Evaluation Design

Now add your evaluation TOR.

Then ask:

This turns AI into a methodological review assistant, rather than just a writing assistant.

Step 8 — Ask AI to Challenge Your Thinking

This is one of the most useful ways to use the tool.

Don’t ask AI only to confirm your approach.

Ask it to challenge you.

This changes the relationship with AI.

Instead of:

“Help me write this.”

you are asking:

“Help me think more critically about this.”

Step 9 — Use It for Qualitative Research

Suppose you have an interview guide and a book about qualitative research.

Then ask:

Then follow up:

You now have a simple workflow:

Review → Challenge → Improve

Step 10 — Use It to Review Qualitative Findings

After data collection, add your qualitative findings or analysis.

Then ask:

For example

The point is not that the second sentence is always correct. The point is to make the strength of the statement match what the evidence actually supports.

Step 11 — Use the Sources for Reporting QA

Now imagine your evaluation report is almost finished.

Add the draft report to the notebook.

Then ask:

Then ask:

This can give you a useful AI-assisted quality-assurance pass before the report goes to a client or donor.

Step 12 — Verify Important Citations

This is the step I would not skip.

When AI gives you an important claim:

1. Find the citation.
2. Click the citation.
3. Read the original passage.
4. Check whether it really supports the claim.
5. Check whether AI interpreted the author’s argument correctly.
6. Decide whether the source actually applies to your evaluation context.

A Complete M&E Example

Imagine you are evaluating a youth employment program.

Your notebook contains:

• Evaluation TOR
• Theory of Change
• Baseline Report
• Interview Guide
• Evaluation findings
• Relevant academic papers
• A professional evaluation book

Now work through these questions in order:

You have now moved from asking isolated AI questions to building a source-grounded evaluation workflow.

A 15-Minute Exercise for EvalCommunity Users

Let’s make this practical.

1. Open Gemini Notebook.
2. Click New Notebook.
3. Name it M&E Methodology Test.
4. Click Add Sources.
5. Add one evaluation document.
6. Click Add Sources and look for Google Play Books.
7. Select an eligible professional book.
8. Check that both sources appear in the Sources panel.
9. Ask: “What methodological principles from this book are relevant to this evaluation document?”
10. Open at least three citations and check them yourself.

A Few Things to Keep in Mind

• Not every ebook is eligible for the book-source feature.
• The specific capability requires an eligible book purchased through Google Play Books.
• Your own uploaded PDF or EPUB can still be used as a regular source where supported.
• Feature availability and interface labels can change as Google updates the product.
• Always verify important citations and methodological claims.

The EvalCommunity Takeaway

The exciting part isn’t simply that AI can now work with books.

The bigger opportunity is bringing different types of knowledge into one research environment:

Published expertise
+
Research
+
Evaluation documents
+
Your evidence
+
AI-assisted analysis

That gives you a much better starting point for asking:

rather than simply asking:

“What does AI think?”

That’s the difference between using AI to generate research and using AI to support better research.