This is Hamamoto from TIMEWELL.
Most people have lost a full day to a single proposal deck. You structure the argument, choose the visuals, align the fonts and colors, then clean up the layout. No step is hard on its own, but stacked together they swallow an afternoon. Genspark is the tool that cut into that problem. You hand it a prompt, and a few minutes later you have slides with a title page, table of contents, body sections, and graphics. The first time I used it, I genuinely said out loud, "so it's come this far."
But over the past year, Genspark has outgrown the label of "document tool." In April 2026 it signed a strategic partnership with Microsoft and now runs directly inside PowerPoint and Excel.1 This article walks through how to produce documents efficiently with Genspark as of June 2026, along with the points where people tend to stumble.
How Genspark Changed: Then vs. Now
I first wrote this article in early 2026. Back then, the standard workflow was to generate an HTML document with Genspark and import it into Figma for editing. The slide-generation ability was already strong, but Genspark sat firmly in the category of a single-purpose slide generator.
That changed dramatically in six months. Today Genspark calls itself an "AI Workspace" — an integrated environment that handles not only slides but research, spreadsheets, document drafting, and even a phone agent, all from one screen. At its core is what Genspark calls the Super Agent. Rather than handing everything to one large language model, it uses an approach called Mixture-of-Agents (MoA), breaking an incoming task into smaller subtasks and routing each to the best-suited model or tool.2 Research goes to a research-strong model, slide formatting to a slides specialist — a division of labor.
The biggest shift came on April 29, 2026, with the Microsoft partnership. Genspark's Slides, Sheets, and Docs generation agents were embedded as native plugins inside PowerPoint, Excel, and Word.13 Not a sidebar extension, not a browser add-on — they live inside the Office apps themselves. With more than a billion Office users worldwide, the bet is to meet people where they already work rather than ask them to migrate to a new tool. For anyone who builds documents for a living, that change is far from trivial.
The instruction I wrote six months ago — "import into Figma to edit" — is now just one option among several. As the person who wrote it, I feel the pace of this field more acutely than anyone.
Looking for AI training and consulting?
Learn about WARP training programs and consulting services in our materials.
The Basic Generation Workflow
The actual steps for creating a document in Genspark have stayed simple. First, prepare the source material — the text or data the document will be built on. Whether it's a feature announcement or a market analysis, the more grounding material you provide, the more stable the output. Then hand that material to Genspark and tell it what kind of document you want. That's it.
The difference shows up in how you write the prompt. A sloppy prompt yields a mediocre document; a specific prompt yields consultant-grade slides. The more you use it, the more you feel that gap.
For instance, start with a base instruction like "Create a complete YouTube structure, script, and slides based on the following," then layer on conditions: "16:9 format," "consultant-grade design," "don't overcrowd the slides," "include practical use cases." Specifying a concrete title sharpens the direction of the whole deck. In my experience, information density creeps upward when you say nothing, so capping it — "maximum three key points per slide" — produces something far more readable.
What comes out is not text poured into a slide template. It includes a title page, a table of contents, logically sequenced sections, supporting icons and graphics, and, if requested, concrete prompt examples and use cases. The way a consultant organizes information — chapter it, then add visual hierarchy — is reflected, to a meaningful degree, in the structure.
Post-generation tweaks are also far easier than before. Text editing, font changes, layout adjustments, and element duplication or deletion all happen directly inside Genspark. Fix a specific line, delete a section that doesn't apply, reorder content — for that level of change, you finish in place without exporting anywhere. For many cases, this in-app editing is enough.
Office Integration and Export: More Exits Than Before
This is the single biggest leap of 2026. The old headache was how to get your Genspark document out into the world; now the exits have widened considerably.
Start with export. On a paid plan, you can write out to PowerPoint (.pptx), PDF, and Google Slides. PowerPoint export requires Plus or above, but the export itself consumes no credits. Once you're on a paid plan, you can export as many decks as you like at no extra charge.4 Earlier, exported PPTX files sometimes had broken fonts or missing images; that's been fixed, and decks now export with accurate typography and properly embedded visuals.4
What I've seen repeatedly in AI implementation work is the reality that simply installing a tool doesn't raise productivity. A tool like Genspark only pays off once you design who uses it, for which task, with what prompts. TIMEWELL's AI consulting service WARP walks alongside companies from tool selection through internal adoption.
The centerpiece is the Microsoft Office integration. In Genspark AI Workspace 4.0, you can run deep research inside PowerPoint and have the AI generate and edit slides without leaving the deck you're already in.1 In Excel, you can ask questions of your data in natural language and pull out analysis, charts, and insights without writing formulas. In Word, you get context-aware drafting and editing. Because it all happens inside the apps you already know, the hurdle of "learning a new tool" has nearly vanished.3 On top of that, the desktop "Genspark Claw" lets the agent reach into local files and everyday work apps.5
On the PowerPoint side, a feature added in late 2025 lets you import existing .pptx files into Genspark and save them as reusable templates.6 Teach it your house style once, and subsequent decks generate to that mold. It's a quiet feature, but it earns its keep in keeping a brand consistent.
To put this change in practical terms: previously, "build in Genspark" and "finish in PowerPoint" were separate jobs with a tedious export step bridging them. Now that bridge is gone — you keep your usual PowerPoint open and run research, generation, and editing through it. The cost of switching between tools drains more focus than people realize. Removing it alone changes the rhythm of the work noticeably. Among the people around me, since Office integration shipped, the flow of "draft in Genspark first, fine-tune in PowerPoint after" took hold almost overnight.
Finishing in Figma: When the Details Matter
The cases where in-app editing or Office integration is enough keep growing. Even so, when you need to follow a brand guideline strictly, position elements to the pixel, or align tone across multiple documents, Figma is the right fit.
The standard way to bring Genspark documents into Figma is the html.to.design plugin. Click "View & Export" at the top right of the Genspark editor to open a preview of your slides in a new tab, capture those slides with the html.to.design Chrome extension, then drag and drop the resulting .h2d file into Figma. Each element imports as something you can edit independently.7 Text blocks, icons, and layout all come apart so you can touch them individually.
One caveat: complex layouts may need adjustment after import. Simple documents convert cleanly; decks with intricate elements may require post-conversion cleanup. Still, basic adjustments aren't difficult, and you can tidy things up without breaking the original intent.
If you're not committed to Figma, Canva editing is an option too. Genspark itself officially documents continued editing in both Canva and Figma.8 Figma if your team has a designer, Canva if it doesn't — that's the practical split. Take it all the way into Figma and you can change fonts and sizes in bulk, adjust to brand colors, reposition elements precisely, and even add interactivity with prototyping. The final output exports as PDF, per-slide PNG, or prototype links.
Build the structure and content fast in Genspark, then polish the design in Figma. That combination captures both generation speed and the finish that client-facing documents demand — the most practical workflow available right now, in my view.
Putting It to Work, and How to Think About Pricing
A few places where Genspark really shines. Reference materials for video content are a natural fit — prompt it with "create a complete YouTube structure, script, and slides for the following," and you get the full content package. Meeting and proposal decks are another strength: "create a proposal deck for entering a ¥5 trillion AI market" returns market analysis, competitive landscape, entry strategy, and projected outcomes in a single generation. Internal training works the same way — "create an AI literacy training deck for new employees" stands up a structured curriculum from fundamentals through practical application in minutes.
A practical warning here. A generated document is sometimes ready to use and sometimes needs fact-checking and verification of the numbers. Always confirm AI-produced market sizes and statistics against primary sources. This is the cardinal rule of building documents with generative AI, not something specific to Genspark. The better a deck looks, the more tempting it is to wave the contents through, so this actually calls for more caution. Getting caught flat-footed when a prospect challenges a number is the one outcome to avoid.
Flip that around, and this is exactly where consultants and business professionals earn their keep. Genspark crushes humans on the speed of producing a draft, but judging whether the document will actually move the other party remains a human job. Hand the skeleton to AI, and redirect the freed-up time toward sharpening the story or swapping in a single slide tuned to the audience's concerns. The people who use it that way extract the most value from generative AI.
On pricing: as of June 2026, the Free plan offers 100 credits per day, Plus is $24.99/month (about $19.99/month billed annually) starting at 10,000 credits a month, and Pro is $249.99/month (about $199.99/month billed annually) starting at 125,000 credits a month.910 On paid plans, casual chat and basic research don't consume credits — the credit pool is reserved for heavy work like media generation and complex agent workflows.10 For individuals or small teams creating documents, Plus is plenty to run on.
Finally, if you're rolling this out across an organization, template your effective prompts. Keep purpose-built prompts on hand — weekly reports, proposal decks — and the only things you change are the topic and the source material. With the format instructions fixed, you can produce at volume. Effective prompts are a team asset. Once I started sharing the prompts that worked inside our company, the time spent on document creation dropped by more than half by my own reckoning. Aligning how a tool is used across the organization moves the productivity needle more than simply adopting the tool.
Tools move fast. After watching Genspark change this much in six months, I can't predict where it lands in the next six. That's precisely why it takes skill to design which tool goes into which task, and how. If that feels difficult, let's think it through together with TIMEWELL's AI consulting service WARP. We help map out AI adoption fit to your actual operations.
References
Footnotes
-
Genspark, "Introducing Genspark AI Workspace 4.0: Your AI Employee, Now Everywhere." https://www.genspark.ai/blog/genspark-ai-workspace-4 ↩ ↩2 ↩3
-
Lindy, "I tested Genspark AI's 2026 features: Here's what worked." https://www.lindy.ai/blog/genspark-ai-features ↩
-
Windows Forum, "Genspark and Microsoft Partner: AI Agents Embedded in Word, Excel and PowerPoint." https://windowsforum.com/threads/genspark-and-microsoft-partner-ai-agents-embedded-in-word-excel-and-powerpoint.415811/ ↩ ↩2
-
Genspark, "AI Slides FAQ." https://www.genspark.ai/docs/ai_slides_faq ↩ ↩2
-
Business Wire, "Genspark Launches Genspark Claw for Desktop and Microsoft Office Integration with AI Workspace 4.0." https://www.businesswire.com/news/home/20260408545044/en/Genspark-Launches-Genspark-Claw-for-Desktop-and-Microsoft-Office-Integration-with-AI-Workspace-4.0-Bringing-AI-to-Local-Files-and-Everyday-Work-Apps ↩
-
Genspark, "AI Slides Changelog." https://www.genspark.ai/docs/ai_slides_changelog ↩
-
html.to.design, "From Genspark AI Slides to Figma." https://html.to.design/blog/from-genspark-slides-to-figma/ ↩
-
Genspark, "Quick Guide: Editing Genspark AI Slides in Canva and Figma." https://www.genspark.ai/blog/genspark-ai-slides-edit ↩
-
Scribe, "Actual Genspark Pricing Plans 2026 Revealed - Free vs Plus vs Pro." https://scribehow.com/page/Actual_Genspark_Pricing_Plans_2026_Revealed__Free_vs_Plus_vs_Pro__bP3QgpUWR1KpqgUC9jbg1w ↩
-
Genspark, "Membership Plans | Genspark Help Center." https://www.genspark.ai/helpcenter/membership-plans ↩ ↩2
