This AI has left us both Terrified and Impressed - Our 4 Genspark Experiments 🤯

EDITOR’S NOTE

Dear Future-Proof Humans,

What if your AI could think, plan, and act like a team of specialists rather than a single assistant?

This week, I've been testing Genspark.ai, a platform that uses a "Mixture-of-Agents" approach where multiple AI models work together under the control of an orchestrator.

After exploring connections to Gmail, WhatsApp, Figma, and other tools in previous newsletters, I wanted to see what happens when AI doesn't just respond to commands but actively plans and executes complex workflows.

I fed Genspark real B2B SaaS funnel data and watched it:

✔️ Analyze conversion patterns across different industries and channels
✔️ Create presentation-ready slides with data visualizations and speaker notes
✔️ Build a complete landing page optimized for increasing the number of requests made for the product demo
✔️ Generate comprehensive SEO educational content pages

In this edition of Nanobits:

⚡️ My candid experiments with each of Genspark's key tools
🔗 How it compares to alternatives like ChatGPT, Perplexity, and Gamma
⚠️ The strengths and limitations I discovered through hands-on testing

Let's see how this new breed of "agentic AI" might transform how we approach complex business challenges!

GENSPARK AI: THE NEXT GENERATION OF AI ASSISTANT

Genspark.ai is built by MainFunc Inc., whose founders come from Microsoft, Google, and Baidu. They've secured $100 million in Series A funding, pushing their valuation to $530 million.

What's interesting is how they've changed direction. They started as a search tool and gained 5 million users, but then pivoted to something they call an "AI Agentic Engine," which suggests a dynamic approach to market needs and a larger ambition.

Why the change? The AI search space got crowded fast. By switching to this "Super Agent" approach, Genspark is betting on something bigger: AI that doesn't just answer questions but takes action. Instead of waiting for you to pull information, their AI pushes forward by planning and executing tasks on its own. You tell it what you want done, and it figures out the steps without you needing to micromanage every detail.

MY EXPERIMENTS WITH GENSPARK AI

I decided to test Genspark.ai's capabilities with a practical business challenge: analyzing B2B SaaS funnel conversion data across different industries, client sizes, and acquisition channels. Could this "agentic" AI platform deliver meaningful insights and actionable recommendations?

I tried four key modules and compared them to other AI tools.

AI Sheets: Automated Data Analysis

My prompt and Genspark AI’s response:

In the first response, Genspark AI Sheets did a better job than Perplexity and ChatGPT. It immediately stood out from other tools. ChatGPT gave better outputs upon further prompting.

Here’s the complete ChatGPT prompt thread:

Here’s the complete Perplexity prompt thread:

This is the final submission, which was created with about 75% human input:

Since I was on a free tier, I could not prompt the tool more for better results, like I was able to do on ChatGPT. However, it was pretty good for the first cut of analysis. I was expecting some couple segmentation analysis, but AI Sheets didn’t do any of that.

If this interests you, here is the complete problem statement. Reach out to us if you are stuck!

I would also like to leave you folks with a thought that, no matter how good the tool is, it can't replace human thinking. The final submission and the AI-generated responses, although with one-shot prompting, is far more mature and comprehensive. 

AI Slides: Executive-Ready Presentations

Now, for my favourite part of the experiment.

I always hear people complain about existing AI PPT maker tools, like Gamma, etc. And, I agree. I tell you why shortly. One day I came across an AI influencer's review reel about Genspark AI. The quality of the PPT looked good in the review and the product explainer video [on the website]. So, I quickly tried it out and I was not disappointed. Here's why...

My prompt:

[I fed the entire response from AI sheets into the AI slides prompt window along with the following instruction] 

Help me create an executive summary deck based on the above response. 

Genspark's AI Slides produced a 12-slide deck that was genuinely presentation-ready. Unlike Gamma, which created visually appealing slides but with more generic content, Genspark:

• Created custom charts showing conversion rates by industry with clear color-coding

• Included an executive summary that highlighted the key findings

• The best part was that the entire deck is editable; you can also import the deck in PPT and PDF formats.

The presentation required some editing before sharing it with stakeholders. While visually better (slightly), Gamma's output needed significant content revisions to match the data-driven specificity of Genspark's deck.

Genspark AI’s response:

Gamma’s response:

Super Agent: Landing Page Creation

My prompt:

Research prove.com, the identity verification and management solution provider, and create a landing page for their fraud prevention solution. The objective of the landing page is to optimize for increase in form fills by the visitors on the page. The form fill denotes that visitors are interested in the fraud prevention product of Prove. 

This was perhaps the most impressive module.

The Super Agent's response blew me away. Rather than just describing what a landing page might include, Genspark actually built a complete, functional page. The agent created a fraud prevention solution landing page with intelligent conversion optimization:

1. The hero section immediately communicated value with "Prevent Fraud Without Sacrificing User Experience" and included social proof

2. It included precisely quantified benefits.

3. The form strategy was particularly clever, it appeared at the top and bottom of the page, but with different messaging tailored to visitors at different stages of consideration.

4. Most impressively, the agent incorporated a visual explanation of the technology that made complex verification concepts accessible.

Genspark generated a fully functional responsive site with proper HTML/CSS/JS that I could immediately publish. It even provided a live URL (https://opbmqhtb.gensparkspace.com/) where I could see the page in action.

When comparing this to other AI tools I've tried for landing page creation, most generate design concepts or basic HTML templates requiring significant development work. Genspark's Super Agent delivered a polished, conversion-optimized page that directly addressed the specific business challenges identified in our funnel data - a level of integration I haven't seen elsewhere.

Spark Pages: Comprehensive Topic Coverage

My prompt:

Topic: Top 10 GaN Chargers for Travelers in India

Genspark generated a decent-looking page. Unlike Perplexity Spaces, which pulls information from various sources, Genspark's Spark Page:

• Referenced fewer articles

• Instead of 10 chargers, it wrote about 5 of them

• The content schema was not elaborate

• Media, YouTube videos, and tables can be inserted manually, but not by prompting

• It doesn’t save the chat history, so every time you generate something additional, you must insert it immediately into the page. Inserting additional responses also requires manual formatting.

The Verdict

For data analysis: Genspark AI > Perplexity > ChatGPT

For PPT creation: Genspark AI > Gamma

For Wikipedia type page creation: Perplexity Spaces > Genspark AI

WHAT ELSE IS IN THE GENSPARK AI’s MULTI-FACETED TOOLKIT

To provide a clearer overview of this diverse toolkit, the following table summarizes its key components:

GENSPARK AI’s VISION AND THE FUTURE OF AGENTIC AI

Several potential future developments for Genspark.ai can be inferred from its current trajectory and stated intentions:

• Enterprise Solutions: A clear direction is the development of specialized versions of its technology tailored for business applications, likely accompanied by customized pricing and support models.

• API Access: Providing robust API access for developers and businesses is a logical step, enabling them to integrate Genspark.ai's powerful agentic capabilities into their own applications and workflows. The availability of API documentation for the Slides tool suggests this is already underway for certain components.

• Specialized Vertical Applications: Creating industry-specific versions of its technology for sectors such as healthcare, legal research, or academic publishing could unlock significant value.

• Expanded Ecosystem: Future plans may include the integration of embedded AI on IoT devices, enhanced contextual personalization through continuous learning from user interactions, and potentially a marketplace for third-party tools and agents to expand the platform's capabilities.

The announcement of intentions to build "the First-Ever AI Search+Browser+Agents Super App" signals a powerful ambition to become an integrated digital hub where users conduct a wide array of online activities, moving beyond a collection of point solutions to a comprehensive digital assistant.

END NOTE

After evaluating Genspark alongside competitors, the verdict is in: it isn't perfect, but it's carving out an interesting position in the increasingly crowded AI assistant market.

In our AI face-off, Genspark outperformed both Perplexity and ChatGPT for data analysis with its actionable insights and beat Gamma in presentation creation despite the latter's slightly more visual appeal. Perplexity still creates better reference pages than Genspark's Sparkpages, though Genspark's interactive elements show promise.

While Genspark excels with its specialized agent system, real-world interactions, and generous free tier, it still struggles with bugs, quick credit depletion on complex tasks, and limited customization options.

Nevertheless, the platform offers tech enthusiasts, marketers, researchers, and businesses a glimpse of truly autonomous AI assistants. If Genspark can execute that ambitious vision, the new "Search+Browser+Agents Super App" could position Genspark as a comprehensive digital hub rather than just another AI tool.

Want to take Genspark for a spin? Try these three exercises:

1. The 60-Second Pitch Deck: Ask Genspark to create a 10-slide presentation for your product using only a brief description, and watch how it structures your key messages.

2. The Video Script Challenge: Have Genspark generate a one-minute explainer video script for a complex feature you offer and compare it to your current messaging.

3. The Automation Experiment: Identify a multi-step task you do manually (like researching competitors and industry news, creating thought-leadership content). Can Genspark's Super Agent handle it end-to-end?