QueryLift, a Tokyo startup founded in May 2025, closed a JPY 100 million seed round this week—roughly $640,000 from JAFCO and Z Venture Capital, the investment arm of LY Corporation. The company's pitch is straightforward but reflects an emerging anxiety among businesses: it helps enterprises track whether ChatGPT, Gemini, Claude, and Google's AI Overviews are mentioning them at all, and if so, how often and in what context.
The funding announcement, released Tuesday, positions QueryLift at the forefront of what insiders are calling generative engine optimization, or GEO. Think of it as SEO's younger sibling, though the market remains small enough that calling it a sibling might be generous. Companies are scrambling to understand how their brands surface in AI-generated answers, and QueryLift offers one of the first systematic ways to measure it.
CEO and co-founder Souta Noguchi said the capital will go toward natural language processing research, platform enhancements, and building out an enterprise sales operation. Those are the usual spending priorities for a seed-stage software company, but the underlying problem QueryLift addresses is less usual: the black box of how large language models decide which companies to cite when answering user queries.
Monitoring the New Search Landscape
QueryLift's platform operates by monitoring AI search responses across prompts that clients define themselves. The system crawls answers from multiple generative models, tracking mention rates, citation frequency, and what the company calls "topic share." It identifies which sources the AI systems are pulling from, how brands stack up against competitors, and measures shifts over time.
In an interview with AWS Startup Blog published in mid-September, CTO and co-founder Shusuke Komatsu explained that enterprises feed the platform key questions relevant to their operations. QueryLift then retrieves AI responses on a recurring schedule. The startup is developing what it describes as a "GEO agent" intended to run analysis autonomously, recommend content adjustments, and verify results—though it's unclear how far along that work is.
Early customer results, at least according to the company, suggest the approach has merit. Japan's Liberal Democratic Party deployed QueryLift to optimize its official website for AI search engines. Over four months, the citation rate in generative AI answers climbed 2.7 times, according to a case study disclosed in the September 2026 funding announcement. That effort followed coverage in Nikkei earlier in the year highlighting the LDP's push to make official information more legible to AI systems.
Education startup Loohcs reported a sharper gain in June 2025: its citation rate on key prompts rose from 22 percent to 97 percent using QueryLift's platform, according to a customer testimonial on the product site. Whether those improvements hold as AI models evolve remains to be seen.
Academic Credentials and Technical Stack
Noguchi studied computer science at Keio University's Shonan Fujisawa Campus and later at the Technical University of Munich. Before founding QueryLift, he spent time as a research intern at Yahoo! JAPAN's Market Intelligence division analyzing search queries and big data, then moved into interest-rate derivatives trading at a hedge fund—an unusual path that suggests comfort with both technical systems and market dynamics.
Komatsu graduated from Keio SFC in 2024 and completed a master's degree at Nara Institute of Science and Technology in 2026. He maintains research affiliations with NAIST and RIKEN's Guardian Robot Project, focusing on natural language processing and human-robot interaction. Those academic ties give QueryLift a research depth uncommon for such a young company.
The technical infrastructure runs on AWS, using Terraform and Terragrunt for infrastructure-as-code, ECS on Fargate for compute, and Aurora PostgreSQL Serverless v2 for the database. Security runs through CloudFront paired with AWS WAF. Engineers Justin Wulf and Mitsuteru Oga told AWS the architecture has yet to produce an infrastructure-originated incident, though that's perhaps less remarkable for a company so early in its life.
Research Ambitions and Market Context
In August, QueryLift announced a full paper acceptance at CIKM 2026, an academic conference focused on information and knowledge management. The paper, titled "Assessing Attack Surfaces in Generative Search Engines through Publisher Attributes: A Case Study in Political Domains," examines vulnerabilities in how AI search engines surface and cite information—a subject with obvious commercial implications for QueryLift's business model.
The GEO market, such as it is, began taking shape in late 2023. A whitepaper from Comcast NBCUniversal's LIFT Labs published in early 2026 positioned GEO as a nascent discipline, naming Bluefish and Profound as other vendors working in the space. The report drew comparisons to the early days of SEO, though whether GEO becomes as central to digital marketing remains an open question.
Investor Perspective
"The GEO market is just emerging," said Shiori Taneichi, an associate at JAFCO, in the funding announcement. The investment rationale, Taneichi explained, centered on providing companies with mechanisms to deliver accurate corporate information to customers in the generative AI era—a polite way of saying businesses need help ensuring AI systems don't ignore or misrepresent them.
Masaki Yuda, a partner at Z Venture Capital, framed the opportunity slightly differently. Companies "need a process to observe how AI sees their organization and make improvements," Yuda said. That observation captures the core anxiety driving demand for services like QueryLift: AI-mediated search could reshape how customers discover products and services, but most companies lack visibility into how those systems work.
QueryLift previously raised a JPY 25 million pre-seed round in June and has scheduled a webinar with JAFCO for early October focused on enterprise GEO adoption. Whether the market grows large enough to support multiple vendors—or whether the major AI labs eventually offer similar monitoring tools themselves—will likely determine how much runway that seed capital buys.
