Gemini Spark is a dream AI assistant, except for one thing: Perplexity is better

I thought that day would be far in the future, but here we are. We already have AI assistants that can take instructions and perform multiple tasks without your supervision (not just at the same time)—almost like having a human assistant.

Google’s Gemini Spark is one such product. In fact, it’s one of the first to be easily and widely available to end users rather than enterprises. It’s so intuitive that when I need to do something in the Google ecosystem, I instinctively gravitate towards Spark. It’s supposed to work in the background, but I can’t help but wonder about an AI assistant that takes context and information from various sources, checks it, and performs any task like a thorough researcher. The novelty factor is what makes it so fun to use.

Apart from the Gemini Spark, I have used its direct rival extensively, A sophisticated computerit was much longer. My experience with both tools has helped me understand how different their approaches are, and there are some meaningful ways to improve by taking a page from Gemini Perplexity’s book.

Once you start looking at Spark’s shiny, automated features, you’ll notice that its greatest strength is the ecosystem it lives within. Other AI tools connect to Google Workspace apps like Gmail and Drive via APIs and MCP connectors, often requiring complex setup. Gemini Spark, on the other hand, is easy as part of the family. This means it doesn’t need any external connections to work within Docs or bring up any legacy context, and it works across supported Google services.

For example, I recently had to file a complaint with Amazon about a defective mattress they delivered. I can do all the digging by hand, spend at least 15 minutes on an unnecessary task, and break the workflow on Monday morning, or ask Spark to do it for me. And you know what I chose!

Gemini Spark extracted order details from Gmail with just a product name, opened up Amazon’s customer support email (a daunting task, as brands are increasingly burying direct contact methods under endless submenus to drive you to their AI chatbots rather than humans), and created a comprehensive email design that was all clearly displayed. I had to resend it once and send it.

This is just one example, of course, but you can do much more sophisticated things with Google services to eliminate everyday digital friction. But once you step outside the Google app ecosystem, you start to see how Perplexity Computer can handle any task.

Complexity doesn’t have its own large language model, so it uses third-party LLMs from OpenAI, Anthropic, Google, and others to answer your queries. This means that Perplexity Computer acts as the orchestrator of this powerful LLM bouquet, delegating internal tasks to the model that suits them best. In my case, it uses ChatGPT to generate high-quality images instead of Google’s own edge-deprived Nano Banana.

Likewise, Gemini Spark’s biggest strength — the closed Google ecosystem — turns out to be its weakest link, and that’s something I can’t expect Google to change. The company is stuck here because it can’t integrate competing models into its own products, so you’ll end up with any of the latest Gemini models. Spark currently runs on top of Gemini 3.7 Flash and I can’t find Google’s Flash class templates As reliable as Claude and ChatGPT understand and respond to my requests.

This is also reflected in the fact that Spark double-checks its work before sharing the final results. In my experience, it allowed for a few glitches in the results it returned, requiring some back-and-forth and cross-checking in another Google app. That’s not to say Perplexity’s computer handles everything perfectly, but its self-assessment is much more reliable, and something Gemini Spark should definitely learn from.

At the end of the day, there are so many invitations to family vacations and birthdays that I can ask Gemini Spark to plan for me. Its great potential continues to be better utilized in more professional environments every day. However, it is currently not available in Google’s personal business-focused Workspace subscriptions. Regulatory requirements for enterprises are likely keeping Google from making Spark widely available for businesses. But whatever the reason, the delay puts it behind the competition.

This is evident in how well Gemini Spark (and even vanilla Gemini) supports third-party tools and integrations. They are dwarfed by the extensive list that Perplexity has amassed over the past few months. It includes Salesforce, HubSpot, Microsoft 365, even Google Workspace and more. works with professional tools. It’s honestly the best tool for those who live their professional lives outside of the Google ecosystem—at least until Gemini lands.

You must have discovered that Perplexity Computer is more focused on complex work where you end up running for hours and hours to achieve the goal you set. He had a good start and it really paid off. The tool looks more mature and reliable than the Gemini Spark when it comes to getting your tasks done. Don’t forget that he can use multiple LLMs to his advantage — something that will remain unmatched for Google for a very long time. You won’t regret spending $200 a month on a Perplexity PC for the kind of practicality and performance it unlocks.

One subconscious reason I consider Perplexity Computer a professional tool is how serious and borderline creepy its interface is. Think Outlook versus Gmail. Like almost every Google product, Gemini Spark is more user-friendly, even if it’s your first day with artificial intelligence. Such availability also extends to its admission price.

Against Perplexity Computer’s $200 subscription, Google recently made Gemini Spark available to its Pro-level users as well, which means better value for money. A $20/month Google AI subscription gives you access to the most advanced version of AI ever created by Google. If any company can make such advanced tools available to billions of end users around the world at the click of a button, it’s Google.

The Gemini Spark has a few rough edges around the corners, but that’s what you’d expect from a new product. None of my concerns with Spark are anything Google can’t fix. Even the arrival of the new Gemini Pro model gives users more options and solves my biggest gripe with it. This is where I see the potential for Gemini Spark to become the default AI tool for the countless tasks we delegate through our devices.