Genspark builds the future of AI agents with Claude

Genspark uses Claude to power adaptive AI agents that transform how people research and create content, delivering unprecedented scale and efficiency for complex workflows.

Key results with Claude:

• Reached $36 million ARR within 45 days of launching Super Agent
• Serves over five million users with dynamic, adaptive AI workflows
• Achieved quality scores verified by large language model judges
• Saved users hours of research time with automated slide creation and multi-step reasoning

Moving beyond fixed search workflows

Traditional AI search products follow the same rigid process for every query: analyze keywords, retrieve web results, then summarize them into an answer. While this works for simple questions, it falls apart when users need complex research, detailed comparisons, or multi-step analysis.

Genspark initially built their search engine using this same approach, adding improvements like specialized data sources and verification agents. Despite reaching five million users, they hit a fundamental limitation. "We realized we were still shackled by a legacy design—a fixed, predefined workflow," said Kay Zhu, Co-founder and CTO at Genspark. "For truly adaptive, context-rich problem-solving, we had to break free entirely."

The team recognized that no amount of incremental improvements could overcome the core constraint of forcing every query through the same static sequence of steps, regardless of complexity or context.

Selecting Claude for adaptive intelligence

To build their Super Agent, Genspark needed an AI model that could handle complex reasoning and coordinate multiple specialized systems working together. After testing various options, they chose Claude as the foundation of their new architecture.

Claude's planning and reasoning capabilities made it perfect for orchestrating Genspark’s mixture-of-agents approach, where different AI models verify each other's outputs to reduce errors and hallucinations. "For the super agent, we're leveraging Claude's planning and reasoning capabilities to drive the entire agent flow," said Zhu.

Genspark discovered Claude excelled at specific tasks crucial to their product. For their AI slides feature, Claude's coding abilities were essential for generating interactive presentations. "Claude handles most of the heavy lifting when creating slides, and we've found it has excellent visual design sense," said Zhu. This combination of technical capability and design sensibility made Claude indispensable across multiple features.

Dynamic orchestration powered by Claude

Genspark’s Super Agent breaks free from the rigid workflows that constrain traditional search engines. Instead of following the same predetermined steps for every query, it adapts its approach based on what each individual question actually requires. Claude acts as the master coordinator, analyzing requests, planning steps, selecting the right tools, and adjusting strategy based on emerging information.

This adaptive system relies on three core innovations:

• Dynamic orchestration where Claude coordinates eight specialized AI models, ensuring quality through cross-verification
• A library of specialized tools and sub-agents that can handle everything from creating presentations to executing Python code to making phone calls
• Curated, high-quality datasets that verification agents continuously audit to maintain accuracy

The result is an intelligent system that scales its effort to match the task. Simple questions get quick, direct answers without unnecessary complexity. Complex research projects can iterate through multiple approaches, gather information from various sources, and refine results until they're comprehensive and accurate. Depending on user needs, the Super Agent can deliver anything from a straightforward answer to a full presentation, interactive webpage, or coordinate a phone call—all orchestrated seamlessly by Claude.

Transforming how users create and research

For Genspark users, the impact centers on time savings and enhanced capabilities. Tasks that required hours of manual research, copying information, and formatting now take minutes. One cross-check agent example showed five minutes of automated work equals three hours of manual effort. "We're enabling people to do complex research and gather information from the web and compile that into a good-looking slide," said Zhu.

Users can now tackle research projects of complexity and scope that would be impossible to handle manually. The AI slides feature automatically gathers information from multiple sources, analyzes it for relevance, and compiles everything into professional presentations ready for sharing and collaboration. This has changed how users approach complex information tasks, freeing them to focus on higher-level analysis and decision-making instead of manual data gathering.

The business impact has been equally dramatic. The rapid market adoption and substantial revenue growth validate Genspark’s bet that adaptive AI is the future of how people will work with information. Users consistently report that the combination of speed, accuracy, and creative capability has transformed their research workflows in ways they hadn't thought possible.

Building the future of adaptive AI

Genspark’s journey from rigid search to adaptive intelligence revealed a fundamental truth about AI development: the most powerful systems emerge when you give agents flexibility rather than constraints. Their transition from a successful search product to an even more capable Super Agent platform validated a counterintuitive principle.

"We discovered an essential principle: less control, more tools," the team explained. "Overly structured workflows limit creativity and depth, whereas allowing multiple specialized agents to tackle different aspects of a problem—and granting them the freedom to choose and switch tools—unlocks far greater capabilities."

This philosophy now drives Genspark’s roadmap as they prepare for their next feature launches. Each new capability represents another way for their AI to approach problems creatively. For Genspark, this represents the dawn of a new era where AI becomes a true thinking partner—one that doesn't just execute commands, but collaborates with human creativity to tackle problems we haven't even imagined yet.