Genspark vừa công bố việc tích hợp công nghệ Claude vào báo cáo tài chính theo yêu cầu. Điều này giúp người dùng có thể yêu cầu báo cáo tài chính một cách nhanh chóng và chính xác hơn. Đây là một bước tiến quan trọng trong việc cung cấp các dịch vụ tài chính hiệu quả và tiện lợi cho khách hàng. #Genspark #Claude #Báo_cáo_tài_chính

Back in June 2024 — an eternity in the fast-moving generative AI sector — a startup founded by Microsoft, Google, and Baidu alumni called MainFunc launched its first product, Genspark, an AI search engine.

Since then, the collision of generative AI, which can create new content on demand, and search, which traditionally retrieves it, has only intensified across the industry. Google recently added Search grounding to its Gemini AI Studio and of course, OpenAI just integrated its powerful realtime web SearchGPT directly into its signature chatbot product ChatGPT.

But now MainFunc — powered by a $60 million seed round led by Singapore-based Lanchi Ventures and supported by global angel investors — is hitting back, teaming up with AI model maker Anthropic to launch “Distill Web” for Genspark, a tool designed to make financial reports more understandable and accessible.

The tool launched earlier this week and powers a range of features on Genspark’s AI search engine, including Genspark Finance.

Overall, very simply, Distill Web gives users the ability to look up of 300,000-and-counting public companies and generate polished, readable, engaging financial reports on their earnings — complete with colorful graphics and charts — turning this complex financial data into visual, easy-to-use formats for a wide audience.

Whether you want to see how Apple is doing after the launch of the new iPhone lineup, or how Google is weathering the AI wars, Distill Web can generate reports that show off these and many other companies’ financial reports, automatically highlighting interesting outliers and trends. It’s like Yahoo or Google Finance on steroids.

“We think that in the AI era, search will become an underlying tool for agents,” said Eric Jing, co-founder and CEO of MainFunc, and the former Chief Product Manager of Search & Corporate VP at Baidu. “People won’t come to search just for a query or a list of links—they’ll come to complete tasks. By combining different tools, agents can do much more than search alone.”

With more than 1 million monthly users gained in just four months through word-of-mouth, Genspark.ai is already establishing itself as a significant player in AI-powered data accessibility. The new update underscores its broader vision to redefine how users interact with data.

Making financial information more accessible to those outside finance

“Our target audience isn’t financial professionals—it’s everyday users who want to understand financial data from public companies.” Jing told VentureBeat. “Eventually, we hope to help people with private company data, too.”

Distill Web’s flagship feature, Corporate Earnings Visual Reports, offers a new way to view financial information.

These AI-powered visualizations turn intricate company earnings into flowing diagrams, highlighting revenue streams, costs, and profit margins. The platform currently provides over 300,000 visual reports, with more added monthly.

To enhance accessibility further, Genspark also offers free Financial Data Packs. These downloadable PDFs provide visual analyses of income statements from over 100 major companies, enabling users to track revenue, expenses, and profits with ease.

Partnering for product integration

MainFunc claims Genspark is superior to other AI search efforts thanks to its efforts on high quality, accurate data — so it is not aiming to have any kind of the scandals observed with Google’s AI Overviews providing hallucinated and erroneous information, for example.

“What sets us apart from others is that we don’t just use AI to provide tools—we create data platforms,” Jing said. “Our approach combines AI-generated insights with traditional coding techniques to ensure the accuracy and trustworthiness of financial data.”

As part of that focus on accuracy, MainFunc evaluated which of the leading large language models (LLMs) would be best suited to comb through financial data and generate accurate charts and graphs, and discovered it was Anthropic’s Claude family — so the two partnered on this effort.

“We found that Claude, Anthropic’s model, is particularly good at handling numbers and complex calculations compared to others like OpenAI,” Jing explained. “That’s why we partnered with them for financial data analysis.”

To further build trust, Genspark implements rigorous validation measures. “One major barrier to building trust in AI is hallucination. To address this, we double-check numbers using both AI and traditional formula-based techniques. It’s critical that the data adds up and is reliable.”

Ask and ye shall receive

Distill Web offers more than just static reports through Genspark. The All-in-One Company Dashboard consolidates key financial metrics for over 70,000 companies, providing a comprehensive view of performance in one place.

For users seeking deeper insights, the AI-powered Financial Copilot answers customized questions, such as comparisons with competitors or identifying growth drivers.

This user-first approach reflects MainFunc broader mission.

“Normal users often don’t know what questions to ask when looking at financial data,” Jing said. “That’s why we present pre-generated, visually rich reports. Users can browse these and ask follow-up questions if needed, removing the initial barrier of crafting queries.”

More differentiated features coming

Looking ahead, MainFunc plans to continue expanding its offerings and introducing new features to Genspark search.

Jing told VentureBeat: “We’re launching a new data search agent soon. It will autonomously collect accurate data from various sources, even when users are offline, delivering results in minutes that would normally take hours or days.”

The company’s broader vision goes beyond tools to focus on transforming data accessibility. “We believe high-quality data is more valuable than the models themselves. Our mission is to build a platform that transforms the way people access and understand data, particularly for non-expert users,” Jing says.