Video Summary:

Prepare and collect source assets

• Decide on the precise visual style to reproduce (infographic, slide deck, social post, one-pager, etc.). Collect clear example files/images that show layout, typography, colour palette, iconography and other visual motifs. Save these examples locally and to cloud storage so they are available to upload into NotebookLM.
• Gather the content to be turned into the new marketing asset: the source document(s), copied text, notes, Google Drive files or uploaded documents. Verify the document you intend NotebookLM to synthesize is complete and final enough for conversion, or mark sections to be summarized or excluded before starting.
• Prefer high-quality, representative style examples: one-off odd variants reduce consistency. If targeting a brand look, include multiple on‑brand examples showing common structure and exceptions (cover, content pages, CTA, data visualization).

Create a dedicated notebook for styles

• Create a new NotebookLM notebook solely for style examples and name it clearly (examples: “Infographic Styles”, “Slide Styles — Blue Minimal”, “Brand Templates”).
• Upload every potential style example into that single notebook so NotebookLM can learn across the collection. More relevant examples improve pattern detection; include different pages/slides from the same deck where possible to capture slide-to-slide variation.
• Allow time for NotebookLM to process visual sources. Visual style ingestion can take appreciably longer than adding text-only files; wait until NotebookLM reports processing is complete before relying on style matches.

Manage files and make sources explicit

• Ensure NotebookLM can unambiguously identify the source content to be converted. If necessary, copy, re-save or rename the source file in Drive or inside the notebook so its filename is unique and obvious.
• Maintain consistent, explicit naming conventions for both content and style files. Use exact filenames in prompts (for example, “Use file Customer-Interview-Summary.docx as the source and Style Deck — Blue Minimal as the style”).
• If multiple versions of a document exist, delete or move duplicates out of the notebook or Drive folder to prevent NotebookLM using the wrong file.

Use plain-language chat prompts in NotebookLM

• Interact with NotebookLM through the chat interface using direct, plain-language instructions that state the source file and the target style by exact name. Example prompts:
  • “Synthesize Customer-Interview-Summary.docx into an infographic using Infographic Style 1.”
  • “Create a slide deck from Product-Strategy-Notes.pdf using Slide Style — Blue Minimal.”
• Always specify which example(s) from the style notebook to emulate. When several styles are present, name the chosen style by the notebook filename or the example title to remove ambiguity.
• When needed, add constraints in the same prompt: desired slide count, data points to highlight, required CTAs, preferred imagery or exclusions (for example, “Produce 6 slides, keep the cover image, use hex colour #0057B8 for headers”).

Infographic generation workflow in NotebookLM

• Select the source document and the style example(s) inside the notebook before prompting. Confirm the style examples have finished processing.
• Submit a plain-language instruction naming both the exact source and the exact style example. Be explicit about output format (single-page infographic, portrait/landscape, PNG/SVG) and content priorities (headline, stats, data visualization).
• NotebookLM can respond by:
  • Generating the infographic directly in the chat: it will produce the content, suggested layout and sometimes an embedded image output or layout description.
  • Providing a prompt or instruction for an internal infographic tool that you must open and paste into that tool.
• Review direct outputs within NotebookLM once processing finishes. For tool-based outputs, follow the supplied prompt into the tool environment.
• Let NotebookLM fully process visual styles before asking it to generate; rushing reduces visual fidelity. Infographics commonly produce faithful reproductions when given good style examples and explicit constraints.

Obtain a Gemini prompt for style replication

• Deselect or remove the source content from selection so NotebookLM concentrates on the generated visual rather than original text. That prevents the prompt from mixing content extraction with style analysis.
• Ask NotebookLM to analyze the generated graphic and produce a precise, Gemini-ready prompt describing layout, fonts, hierarchy, colour palette, spacing, imagery style and any constraints required to replicate the look. Request the prompt as an exact string if you plan to paste it verbatim into Gemini.
• Request additional detail in the prompt when necessary: hex colour codes, font names (or closest matches), proportions (margins, header height) and visual language (flat icons, gradient shadowing, photographic treatment).
• Copy the full prompt NotebookLM returns; it will serve as the basis for image-generation in Gemini and should include both aesthetic direction and functional layout instructions.

Use Google Gemini (image generation) to replicate style

• Open Gemini’s image-generation mode and attach the NotebookLM-produced reference image or style example so the model can use it as a visual reference.
• Attach or paste the document text or content you want embedded into the visual output, or paste the textual content into Gemini’s prompt area. Provide any data or images that must appear on the graphic.
• If higher-fidelity options exist (Pro modes), consider selecting them while monitoring daily Pro query limits; Pro queries can produce more accurate matches but may be rate-limited.
• Paste the prompt copied from NotebookLM into Gemini. In settings, specify that the attached reference image should be used as a style guide. If Gemini supports direct “style transfer” or “reference image” toggles, enable them.
• Trigger generation. The referenced model (transcript name: “nanobanana”) or the selected Gemini imagery model will attempt to reproduce the style while embedding the provided content.
• Review the generated images and iterate: adjust wording, emphasize colour accuracy, refine layout constraints, or include more explicit spacing and typography details. Repeat until satisfied.

Slide/presentation workflow — preparing source slide styles

• Prefer exemplar presentations that export as PowerPoint (.pptx). .pptx preserves editable slide structure on import; PDF-only exports often import as flat images and lose editable elements.
• Download exemplar decks as .pptx and, if needed, import into Google Slides so NotebookLM can access them as editable slide sources.
• Upload imported presentations into the NotebookLM style notebook and rename files for clear reference (examples: “Style Slide Deck — Blue Minimal”, “Sales Pitch — Dark Theme”).
• Include multiple representative slides from the exemplar deck (title slide, section header, content slide, data visualization slide) so NotebookLM learns common component placement and transitions.

Synthesise a document into a slide deck with NotebookLM

• Add the document to be converted into the same NotebookLM notebook that contains the slide style examples.
• Use a plain-language prompt that names the exact document and the exact slide style source. Provide additional constraints: number of slides, required section headers, key messages per slide, or slide format (16:9/4:3).
• Expect NotebookLM to confirm whether you want it to generate the deck; reply affirmatively to proceed. Indicate if the output should be a downloadable presentation file or a set of slide-by-slide text+layout instructions.
• NotebookLM will produce a new presentation that attempts to match the style, slide structure and visual motifs of the example deck. Review the generated file for slide order, text overflow, image placement and consistent styling.
• Expect variable fidelity: layouts, fonts and elements may be approximated but not always transferred perfectly. Be prepared to manually refine certain slides (reposition elements, correct font substitutions or adjust colours).

Replicating slide styles in Gemini/Canvas

• After NotebookLM produces a slide deck, deselect the presentation and ask NotebookLM to generate a Gemini-ready prompt analyzing the slide style, similar to the infographic workflow.
• Copy NotebookLM’s style prompt verbatim. In Gemini, attach the style presentation image(s) or exported slide images and the document to be converted.
• Enable any Canvas or visual-generation modes required (for example, “turn on canvas” or “use image-to-image with reference”).
• Paste the NotebookLM prompt into Gemini and run generation. If Gemini can produce editable output or layers, export back to Google Slides or PowerPoint; otherwise export as high-resolution images and assemble manually.
• Expect partial fidelity: colors, fonts and overall layout can be approximated but some slides will require further prompt refinement or manual editing for precise spacing, alignment or font matching.

Practical tips, caveats and troubleshooting

• Allow ample time for NotebookLM and Gemini to process visual inputs. Visual ingestion and style learning are slower than text ingestion.
• Keep a single, dedicated style notebook to accumulate reusable templates and examples. Reuse the notebook across projects for consistent branding and quicker iterations.
• Be explicit and precise in prompts. Use exact file names, include hex colour codes, font names or preferred type styles, specify slide counts and element positions to reduce ambiguity.
• Attach both a style image and the content document when using Gemini so the model has visual and textual references. If only one is attached, results will skew toward the attached source.
• Monitor daily limits and costs for Pro modes in Gemini. Pro generation can improve fidelity but is subject to quotas.
• Prefer PowerPoint-exportable presentations for slide workflows to retain editable structure during imports. PDFs often import as flattened images that cannot be edited easily.
• When NotebookLM produces incomplete or imperfect slides, refine prompts with more granular instructions: exact slide counts, where to place headings, precise image cropping, hex codes, font substitutions or required icons.
• Expect infographics to be easier and more reliably replicated than complex multi-slide decks. Slides often require additional manual polishing, because templates, master slides and complex content flows are harder to reproduce automatically.
• If NotebookLM seems to ignore a style example, ensure the example finished ingesting, confirm it’s in the correct style notebook and reiterate the filename in the prompt.

Iteration and reuse

• Save successful NotebookLM prompts, Gemini prompts and any prompt variations. Maintain a prompt library keyed to style examples and project types to speed future work and ensure consistency.
• Treat the NotebookLM → Gemini → Google Slides pipeline as iterative: use the first outputs as a starting point, then refine prompts and perform manual edits to achieve production-ready results.
• Build a library of export-ready style images, master slide images and prompt templates so the same workflow can be applied quickly for different source documents and campaigns.
• Version and document the best-performing prompts and generation settings (model variants, Pro toggles, canvas settings) so others can reproduce the same visual fidelity.

Operational sequence checklist (concise)

• Identify target style and collect examples.
• Create and populate a dedicated NotebookLM style notebook; wait for ingestion.
• Upload and/or prepare source content; use clear, unique filenames.
• Use plain-language chat prompts naming exact source and style files.
• For infographics: allow NotebookLM to generate directly or provide tool prompts; review output.
• For Gemini replication: deselect content, ask NotebookLM for a Gemini-ready prompt, copy it and paste into Gemini with the style image and content attached.
• For slide decks: use .pptx exemplars where possible, import into the style notebook, ask NotebookLM to generate slides and expect variable fidelity; use Gemini for further visual replication and export back to editable slides if supported.
• Iterate prompts, tweak colour/typography/spacing details, and perform manual edits to reach final polish.

Expected outcomes and realistic expectations

• Stronger fidelity for single-image outputs and infographics when supplied with clear, consistent style examples.
• Variable fidelity for multi-slide presentations; more manual adjustment and iteration typically required to match exact templates, fonts and complex layouts.
• The pipeline centralizes visual references in one place, automates much of the initial design translation and accelerates production, but final brand-compliant deliverables usually need human review and light post-processing.

Key practical details to remember

• Use exact filenames and style names in prompts to avoid ambiguity.
• Wait for NotebookLM to finish visual ingestion before relying on a style example.
• Attach both style reference images and the content to Gemini for best results.
• Consider Pro options in Gemini for higher fidelity while watching daily limits.
• Prefer .pptx for slide workflows to retain editability; export and reimport as necessary.
• Save and reuse successful prompts and style notebooks to build a repeatable studio workflow.

End summary

• Centralise design references in NotebookLM, use explicit, plain-language prompts naming exact files/styles, obtain precise Gemini prompts from NotebookLM for image-generation, and iterate between NotebookLM, Gemini and slide tools to reproduce visual styles. Expect better results for infographics than for complex slide decks and plan for prompt iteration plus manual refinement to achieve production-ready marketing assets.

Video Transcript

Hello and welcome.

Now, in this video, we’re going to turn NotebookLM into our marketing design studio, assuming that you already know how to take your NotebookLM information and turn it into a marketing asset.

For example, you have already created an infographic, and depending on the infographic, you may like a specific style.

Once you’ve identified the style of infographic that you want, you’re going to download that infographic to your hard drive.

Now, for this process, you can create a new notebook.

Once you create the new notebook, you’re going to upload all of the infographic styles that you want to replicate into the new notebook.

Now, it will take time for the notebook to incorporate the visual style. So even though it takes longer than a typical source, you want to let this process complete.

While we wait for the new notebook, we’re going to give our notebook a name.

And because this is a new notebook, we can upload as many infographic styles as we want as long as we know that we’re going to want to use them in the future.

So there may be a style that we like in another notebook. We can then download that infographic as a new style.

And every style that we want to upload to our style notebook we’re going to upload.

Now, for the sake of this video, we’re only going to have two new styles. However, you can have as many styles as you want. We’re then going to want to get the source information.

That source information can be from your Google Drive. It could be copied text, it could be another uploaded file.

You just know that you want to turn this source into an infographic and you want to use NotebookLM to do it.

We’re going to get a random file from our Google Drive. We’re then going to add that file to our new notebook.

What we’re going to do in this case is we’re going to choose one of our infographic styles.

And what we’re going to do is we’re going to type in to the chat area prompt is going to be in plain language.

We’re just going to state that we want to synthesize our document into an infographic.

And what we want to do is we want to use infographic style number one in this case, and we’re going to go ahead and name that as the style for good measure.

We’re going to pinpoint the exact document.

We can go to the rename tab and copy the document, resave it, and then we’re going to make sure that the AI knows that this is the document that we want to use.

Once we have our chat, we’re then going to Click Enter and NotebookLM is going to tell us the rest of the process.

In some cases, we may need to go into the infographic tool and actually enter a prompt.

In other cases, what NotebookLM will do, as you can see here on your screen, is NoteBookLM will then start to generate the infographic.

It will also tell us exactly how it’s going to lay out the infographic in the chat area.

So in this case, we did not have to go into the infographic tool. We did everything from the chat area. I’m going to stop the video now.

We’re going to pick it up from the creation of our infographic and our new infographic is now complete. And as you can see, NotebookLM will then match the style.

We are then going to deselect our document.

We’re then going to go to NotebookLM and we’re going to ask NotebookLM for a prompt to put into Gemini in order to replicate the style of this infographic.

We’re going to ask NotebookLM for a specific prompt and we’re going to ask for it in plain language and we’re going to click Enter.

Of course we have the infographic style selected when we ask this question here in the chat area and NotebookLM is going to give us the prompt.

So we are going to copy this entire prompt and we’re going to head to our Google Gemini area.

Since we have the image, we’re going to go ahead and upload the image. We’re now going to select the information that we want to synthesize into the infographic.

We now have our document. If you have enough Pro queries left in your account for the day, you can select the Pro version.

We’re now going to paste in the prompt. So we’ve attached the image, we’ve attached the document we’ve placed in our prompt.

We’ve also asked for Gemini to use the attached image as a reference for style. We’re then going to select Create Image. We’re then going to select Enter.

What should happen is nanobanana should then load and Gemini then replicates the style of the infographic.

As an experiment, we can attempt to do the same thing with a slide presentation.

If we go to the slide presentation and we attempt to download, we want to use a presentation where we can Download as a PowerPoint presentation.

If you have been creating presentations in the past and you’ve been creating them into PDF, you may not have this option available on all of your presentations.

Newer presentations which have the PowerPoint export, you will be able to do this with. We’re going to download this as a PowerPoint presentation.

We will then import the presentation into our Google Slide account. And so all we’ll need to do is to bring the presentation in and import it in this presentation.

We’re now going to create slide styles for our new slide styles. We will go to our sources, we’ll then go to our Google Drive. We will then look for our presentation.

We will then add in our presentation as a new slide style. We are going to rename this presentation as a new slide style. So we now have our new slide style notebook.

To go through the rest of the experiment, we need to pick up a document that we want to synthesize into a slide presentation.

We now have our new document. We have our slide style.

What we’re now going to do is we’re going to request that NotebookLM synthesize this document into the slide presentation exact style.

So again, we’re going to use a plain language prompt, but we’re going to use the exact file names that we want to use.

In this case, we’re going to ask that NotebookLM replicate this in a new slide presentation.

We are now going to click Enter and we want to be able to determine if NotebookLM will start creating the slide presentation just based on the chat that we have placed in here in this area, without us having to go into the individual slide presentation to go through the process.

So NotebookLM is asking us if we would like for it to generate a slide deck.

We’re going to go ahead and ask for the slide deck and once we do that, we’re then going to Click Enter and NotebookLM will then begin to generate the slide deck.

We’re going to stop the video now and we’re going to pick it up from the creation of the slide deck. And our new experimental slide deck is now complete and you’ll see the style replicated.

There are definitely elements of replication. So in this case, we’re going to deselect the presentation.

We’re going to ask for a prompt to replicate this slide presentation in Canvas in Gemini.

We’re going to ask for NotebookLM to analyze the style and to give us the exact prompt that we’re going to use in Gemini and we’re going to try to get closer on the colors in this part of the experiment, we’re going to copy this prompt.

We know we’re going to have to get rid of this information. What we’re then going to do is we’re going to add in the presentation from our Google Drive.

We’re then going to add in the document that we want to synthesize into a presentation. We are going to turn on the canvas, and we are then going to paste in our prompt.

We will now export the document to Google Slides.

Now, the final presentation is not as seamless of a process, and as you can see here, some of the slides did not come over just with our prompt.

So we may need to be more specific in our prompt and more specific in the way that we need to do the prompt. It is possible to get close to the original design with minimal prompting.

However, process is not nearly as seamless as we did with the infographic styles.

Okay, so that thanks and I will see you in another video.