How to Use Google Pics in Docs and Slides

You can now take an image already sitting in a Google Doc or Slide, isolate the object that is wrong, change or translate words inside the image, send the Pics file through team review, and put the approved version back without opening a separate design app. For an eligible Workspace customer, that turns many routine visual revisions into a marginal $0 software task inside a plan the team already pays for.

That does not make Google Docs a design studio. Google Pics is strongest in the last mile: one product needs a new background, one headline needs a new language, or one deck image needs to fit 16:9 before a meeting. Master artwork, exact vectors, layered production files, bulk variants, and print preflight still belong in a dedicated design tool.

What Google Pics in Docs and Slides actually is

Google Pics is a Gemini-powered image creation and editing app built on Nano Banana. It is both a standalone canvas and an editor that opens inside Docs and Slides. Think of it as a repair bay attached to the document: you roll an existing visual in, fix selected parts, and return it to the same parking spot.

The Docs and Slides integration began rolling out on September 1, 2026. It is not Google Photos, although you can import an image from Google Photos, Google Drive, or your computer into the standalone Pics app. Supported uploads from a computer or Drive are JPG, PNG, BMP, and TIFF.

The editor handles four distinct jobs:

• Describe a broad change in a prompt, such as cooling the shadows across a photo.
• Click a detected object and ask for a change to that object.
• Drag over a specific area and describe what should change there.
• Click detected image text to replace the words or change their styling. Google also says Pics can translate an individual text element while preserving its design and font.

You can stage up to five object, area, or text changes and apply them together. That is useful because a visual revision usually arrives as a bundle, not as one clean instruction: change the background, remove the cup, correct the date, translate the headline, and make the subhead bold.

How to use Google Pics in Docs and Slides

The same core loop works in both products. The example below follows one launch graphic through the full job.

Imagine a product marketer has a launch deck and a supporting brief. Both contain the same campaign image: a bottle on a studio background with the words “Available now.” The French sales team needs a localized version, the bottle cap is the wrong color, and the image needs a widescreen crop.

1. Check access before touching the asset

Open the document or presentation on a desktop browser. Google lists the two most recent versions of Chrome and Firefox, plus Edge on Windows. The feature requires an eligible plan and is still rolling out.

Business Standard, Business Plus, Enterprise Standard, Enterprise Plus, AI Expanded Access, Google AI Pro for Education, Google AI Pro, and Google AI Ultra are on the current eligibility list. If clicking Edit image opens the older Gemini bar in Docs or sidebar in Slides, your account probably has not received Pics yet.

There are two launch caveats. Pics in Docs is listed as English-only. Pics in Slides supports 29 languages. The app is desktop-only, so a phone or tablet is not a fallback route.

2. Open the image in place

In Docs, click the image, then click Edit image. Google Pics opens within the document. In Slides, click the image on the slide, then click Edit image to open the same embedded editor.

You can minimize the editor while a generation runs and keep working. Restore it when the result is ready. This small detail matters in a real deck: waiting for an edit no longer blocks work on the speaker notes, data, or surrounding copy.

3. Make the targeted changes

For the bottle cap, hover over the image and click the detected cap. Add a direct instruction such as “Change the cap to matte navy and keep the bottle unchanged.” If Pics does not isolate the right object, drag over the cap instead and apply the instruction to that region.

For the headline, click the detected text element and enter the approved French copy. You can also describe a type treatment such as bold or italic. Save the text edit, then add the crop as another change. Pics offers common aspect ratios including 16:9, 1:1, and 4:3, plus manual crop and rotation.

Queue the related changes before applying them. The limit is five simultaneous changes, including both image and text edits. If the revision list is longer, break it into deliberate rounds and check the asset between them.

4. Review the asset, not just the prompt

Share the Pics image with a teammate as a Viewer or Editor. A Viewer can inspect the result. An Editor can make changes to the same image. Keep the source Doc or Slide open so reviewers can judge the visual where it will actually appear, not on an empty canvas.

The review checklist should be concrete:

• Does the product still match the real product?
• Is every translated word approved by a fluent reviewer?
• Did the font, line breaks, spacing, and hierarchy survive the text change?
• Did the object edit disturb shadows, edges, reflections, or nearby elements?
• Does the crop still leave room for slide copy or document wrapping?
• Are legal claims, prices, dates, and names exact?

Generative editing can make a plausible image that is commercially wrong. Treat the prompt as an instruction and the rendered result as a draft that still needs approval.

5. Replace in context, then export the delivery copy

Click Replace image in the top-right corner of the embedded editor. The edited version takes the original image's place in the Doc or Slide, where you can make the final contextual check.

For a separate delivery file, open the Pics image and download a JPEG at Original, 2K, or 4K. Workspace Experiments accounts currently get Original only. Higher-resolution downloads consume promotional limits more quickly, so choose 4K because the output needs it, not because it is the largest button.

This is a clean handoff rule: use Replace image for the working document or presentation, and use a downloaded JPEG for a website, print vendor, social scheduler, or other external destination.

The business math changes only for the right kind of work

The financial case is strongest when the team already pays for an eligible Workspace plan. The U.S. price displayed for Business Standard is $14 per user per month on an annual commitment. For that customer, Pics does not add a separate image-editor invoice. A personal user starting from scratch can get access through Google AI Pro, listed at $19.99 per month.

Compare that with a separate Canva Business seat at $250 per year or Photoshop at $22.99 per month on an annual plan. Replacing one seat only makes sense if the person's work is mostly targeted edits inside documents and decks. Canva's templates and brand controls and Photoshop's production precision are different jobs, so cancelling either tool without auditing the work would be false economy.

Outsourcing gives the contrast more weight. 24Slides currently lists pay-as-you-go rates of $16 per slide for a fix-up, $43 for a redesign, and $66 for a redraw. Across 20 slides, that is $320, $860, or $1,320. Pics will not redesign the deck. It can keep small image-only tickets, such as changing a background or correcting embedded copy, out of that queue so the paid designer handles the work that actually needs design judgment.

There is also a hard capacity number. Typical Workspace accounts get 30 Nano Banana image or infographic generations per month across Slides, Docs, Vids, and Pics used inside Slides and Docs. Google AI Pro has a 500-use monthly pool shared across Gemini features in Gmail, Docs, Sheets, Slides, Meet, Vids, and Pics. This is a convenience layer with quotas, not an unlimited production line.

Seven workflows ranked by who gains most

1. Global product marketing teams

A product marketer can take an approved English launch graphic from a go-to-market deck, replace the embedded headline with reviewed local copy, adjust one market-specific product detail, and return it to the slide. The payoff is not automatic translation. It is removing the export, upload, designer ticket, download, and reinsertion cycle around a small approved change. The language owner still signs off.

2. Sales and proposal teams

A sales lead can adapt a product mockup for a named account, remove an irrelevant prop, change the environmental context, crop to 16:9, and replace the visual in the proposal deck. That can move a late-stage visual correction from the design queue to the deal team while keeping the full slide narrative intact.

3. Ecommerce and retail operators

A merchandising operator could drop a product shot into a weekly trading deck, use the product as a reference for a new scene, correct a background object, and export the approved result as a 2K or 4K JPEG. This pays when the job is a concept or internal campaign mockup. Exact pack shape, labels, color, and regulated claims still require a close fidelity review.

4. Internal communications and people teams

An HR team can correct a date inside a recruiting poster, localize its headline, replace a distracting object, and place the revision back into an onboarding Doc. That keeps a routine update with the content owner instead of reopening a source file that may live with another department.

5. Event and field marketing teams

An event manager can revise a speaker card after a last-minute title change, crop a campaign visual for a presentation, and share the Pics source with a regional editor. The payoff is turnaround speed during the week when dates, rooms, speakers, and sponsor copy change fastest.

6. Trainers and educators

A learning designer can change labels inside an existing diagram, translate a visual for another cohort, and return it to a lesson plan or training deck. The surrounding curriculum stays in place. This workflow should use the same subject-matter review as any other instructional graphic because a polished diagram can still contain a factual mistake.

7. Agencies handling client revision rounds

An account team can keep a client-facing Doc or deck as the review surface, share the underlying Pics image with the assigned editor, and replace only the approved asset. The benefit is less file-version drift for small revisions. Larger campaign systems still need a controlled source file and a designer who owns the visual language.

For a broader comparison of when image models hold visual identity and when they drift, see the 2026 brand-consistency model guide. The earlier Google Pics rollout analysis covers why a fast editor and a complete design platform are not the same product.

Three products worth building around this workflow

1. A Workspace-native visual localization desk

This is the strongest opportunity. Build a productized service for global marketing and enablement teams that receives approved Docs and Slides, changes or translates embedded image text in Pics, routes the output to a language owner, and returns both the updated deck and delivery JPEGs.

The demand is visible: “translate text in image” averages about 5,400 U.S. Google searches a month, while “edit text in image” averages 2,900. The live question “Can AI translate text in an image?” describes the job plainly. The opportunity is not another generic translator. It is the review layer that makes localized artwork safe to publish.

The smallest sellable version is manual and disciplined: one intake form, a shared terminology sheet, a Pics editing checklist, Viewer and Editor roles, named language approval, and an export naming convention. Start with one asset class, such as sales-deck campaign graphics, and one language pair.

The catch is throughput. Google does not document a public Pics API or bulk export, and typical Workspace generation is limited. A human-operated service can start now, but software-scale automation depends on interfaces Google has not announced. Translation quality, text fit, and product fidelity also remain human approval tasks.

2. A deck asset repair queue for revenue teams

Build a narrow service that accepts a sales deck, identifies image-only revision tickets, edits those assets in Slides with Pics, and sends narrative, layout, chart, and master-template work to a presentation designer.

“Presentation design service” averages 210 U.S. searches a month with commercial intent and a $15.07 CPC. That small volume carries buying intent. The existing market also prices the work clearly: 24Slides starts its pay-as-you-go fix-up at $16 per slide and redesign at $43 per slide.

An MVP needs a Slides intake template, three ticket types, a before-and-after approval step, and a strict scope boundary. The catch is that customers may hand over a bad deck and expect Pics to fix the story. It cannot. The service wins by triaging cheap visual repairs away from expensive redesign, not by pretending those jobs are interchangeable.

3. A brand-safe Pics operating kit

Build a lightweight governance product for teams that already make their own visuals: approved reference images in Drive, prompt recipes for recurring asset types, a prohibited-change list, a reviewer checklist, and an asset log that records which source and language were approved.

The adjacent queries show a precise frustration. “Edit text in image with same font” gets about 320 U.S. searches a month, and “canva edit text in image” gets about 590. The latter showed a 680% yearly rise in this run's keyword data. People are not merely asking for image generation. They want controlled changes to something that already exists.

The MVP could be a Workspace template pack plus onboarding, sold to a small marketing team. Its moat is weak if it is only prompts. The durable part is the approval policy, brand reference library, and change log. Google has not documented a Pics brand kit, locked component system, or batch production layer, so those controls have to live around the tool.

What still requires a dedicated design tool

Google Pics is the right tool when the unit of work is a finished raster image with a short change list. It is the wrong center of gravity when the unit of work is a design system.

Google's published feature and export lists do not document vector path editing, layer-based compositing, master components, CMYK or bleed controls, or bulk asset export. Its documented download formats are Original, 2K JPEG, and 4K JPEG. That gives you a practical dividing line:

• Use Pics for an object swap, localized image copy, a crop, an aspect-ratio change, a visual concept, or a reviewed JPEG.
• Use Canva, Photoshop, Illustrator, Figma, or another dedicated tool for reusable brand templates, logo masters, exact masks, layered source files, large variant sets, precise typography systems, print production, and deterministic output.

The other limit is availability. Pics is desktop-only, personal accounts in Germany are excluded at launch, access is rolling out, and usage is capped. Standalone Pics has higher promotional access through at least February 28, 2027, but Google says limits may change afterward.

The honest take: Pics removes friction from the revision loop around an image. It does not remove the need for art direction, factual review, language review, brand governance, or specialist production.

The Monday move

Pick one live deck that already has an image revision ticket. Do not start with a new campaign. Ask the content owner to make no more than three bounded edits in Pics, assign one reviewer, replace the image in context, and export the approved JPEG with a clear version name.

Record four things: minutes from request to approval, number of generations used, defects caught in review, and whether the job still needed a designer. After five assets, you will know whether Pics removes a real queue in your business or merely adds another editor.

Can I put photos in Google Docs?

Yes. Add the image to a Doc, select it on desktop, and click Edit image. If your eligible account has received the rollout, Google Pics opens inside Docs and Replace image returns the edited version to the document.

Can I translate text from an image?

Yes. Google Pics detects text elements and lets you replace or translate them. Google says it can preserve the design and font, but a fluent reviewer should still check meaning, line breaks, names, numbers, and legal copy.

Can AI translate text in an image?

Yes, and Pics can do it inside a Doc or Slide. The valuable workflow is AI-assisted editing followed by human language and layout approval. A clean-looking render is not proof that the translation is correct.

How do I translate an image on my phone?

Not with Google Pics at launch. Google lists Pics as desktop-only. Use a desktop browser for the Docs or Slides workflow.

If you want this review and localization workflow fitted to your team's existing Google Workspace process, I build practical AI automation systems.