500 Q&A Sessions with Zero Hallucinations: How Pharmato Builds an AI-Powered Intelligent Decision-Making Tool for the Pharmaceutical Industry

In the current pharmaceutical industry, a new data contradiction is emerging.

On one hand, pharmaceutical companies are accumulating unprecedented data assets. Global clinical trial databases, patent literature repositories, commercial pipeline databases, along with internally accumulated R&D data, clinical experience, and expert knowledge, collectively constitute a vast pharmaceutical information system.

However, on the other hand, the growth in data scale has not been accompanied by a corresponding improvement in decision-making efficiency. In new drug project initiation, competitive analysis, or clinical strategy formulation, researchers still need to switch between multiple databases, download, organize, and compare data, with a significant amount of time spent on data cleaning rather than value analysis.

“Pharmaceutical companies are not lacking in data; only by leveraging data effectively can its value be fully realized,” said Gao Fei, co-founder of Pharmato, in an interview. In his view, while traditional databases have addressed the issue of information access, they still fall short of supporting decision-making.

Meanwhile, the development of large AI models is bringing new possibilities to the pharmaceutical industry, but their implementation in professional scenarios still faces challenges.

Pharmaceutical R&D and commercial decision-making involve extensive specialized knowledge and demand high accuracy. AI must not only comprehend language but also grasp the complex interrelationships among diseases, drugs, clinical trials, and industry data.

“The more vertical and specialized the field, the more refined the technical requirements become,” said Ye Zhongkai, CTO and co-founder of Pharmato. In the pharmaceutical sector, users need more than just tools that can generate answers; they require intelligent systems capable of leveraging reliable data, understanding business processes, and supporting practical decision-making.

Based on this industry demand, Pharmato chooses to enter the market through intelligent pharmaceutical data, exploring how to leverage AI to connect data, knowledge, and business decisions, aiming to address the practical needs of pharmaceutical companies in effectively utilizing data.

Gao Fei and Ye Zhongkai’s connection began at a pharmaceutical big data company.

Profile of the Founder of Pharmato

At that time, the two jointly oversaw the research and development as well as commercialization of global big data products for the biopharmaceutical industry. Through their long-term service to pharmaceutical company clients, they gradually observed that industry demands were evolving.

In the past, pharmaceutical companies focused primarily on data acquisition and retrieval. However, with the rapid advancement of AI technology, they are now further exploring how to leverage AI to enhance R&D efficiency, drive internal digital transformation, and integrate data capabilities into the entire business workflow, from preclinical research to post-market stages.

“Many pharmaceutical companies already have mature R&D systems and professional talent, but how to leverage AI to improve daily work efficiency and connect different business segments remains a challenge,” said Ye Zhongkai.

It was precisely during this process that the two realized that the pharmaceutical industry needs not just data tools, but rather capabilities that can make Intelligent Solutions for Understanding Business Scenarios, Connecting Data Resources, and Supporting Decision-Making.

In 2025, the two decided to embark on a new journey by establishing Pharmato. This time, they aimed to integrate pharmaceutical data, artificial intelligence, and business scenarios to explore intelligent solutions more closely aligned with the actual workflows of pharmaceutical companies.

This strategic direction also necessitates that the team possess both a deep understanding of the pharmaceutical industry and robust technical development capabilities. Gao Fei brings a background in clinical pharmacology, complemented by interdisciplinary expertise in mathematics and computer science, and has been extensively involved in the construction of pharmaceutical databases. Ye Zhongkai is responsible for technical architecture and product development. Their collaborative model creates a synergistic complementarity: one partner comprehends the pharmaceutical business scenarios, while the other drives the practical implementation of technology.

In terms of brand naming, Pharmato aims to strike a balance between professional attributes and a youthful expression. “‘Yao’ (Phar) represents the pharmaceutical industry, while ‘fanqie’ (mato) is intended to convey a more open and vibrant brand perception,” introduced Ye Zhongkai. The English name, Pharmato, is derived from the combination of “Pharma” (pharmaceuticals) and “Tomato”.

Following its establishment, Pharmato rapidly advanced the implementation of its products and business operations. In September 2025, the company’s official website was officially launched; in December of the same year, it released the first version of its SaaS product and completed the delivery of its inaugural consulting project. In February 2026, the company further expanded its product offerings to cover both ToC and ToB scenarios. Currently, our consulting services have served over 20 pharmaceutical companies, and the system (pharmato.cn) is used by nearly 600 enterprises.

From the perspective of Pharmato, the implementation of AI in the pharmaceutical industry is not merely about providing a software tool, but rather about helping pharmaceutical companies achieve a systematic upgrade spanning data management, process optimization, and business applications. This is precisely why it has adopted the “product + consulting + data” model.

Consultation serves as a key entry point for Pharmato to help enterprises initiate their AI transformation. The team conducts needs assessments to assist pharmaceutical companies in identifying application scenarios, planning data frameworks, and integrating AI capabilities with actual business processes.

Once requirements are clearly defined, standardized products become the vehicle for deploying AI capabilities, while data forms the foundation for continuous product optimization. “Only high-quality data can lead to superior products; when combined with more stable models and more efficient agents, this creates a complete application loop,” said Ye Zhongkai.

Pharmato Deeply Integrates Various Professional Data

In this model, consulting, products, and data do not simply follow a linear relationship; rather, they mutually reinforce one another: consulting identifies enterprise needs, driving product iteration and data infrastructure development; feedback from product applications further optimizes service capabilities; and long-term accumulated data insights, in turn, enhance consulting solutions.

Currently, Pharmato primarily serves pharmaceutical companies and healthcare professionals. For enterprise clients, it offers AI-powered products, customized deployment, and consulting services; for individual users, it addresses the information retrieval and efficiency enhancement needs of healthcare professionals through a subscription-based model.

Under the “Product + Consulting + Data” business model, Pharmato has built a vertical AI product suite centered on its Integras platform, tailored to the data characteristics and business needs of the pharmaceutical industry, in an effort to genuinely integrate AI into pharmaceutical companies’ R&D, medical affairs, and commercial decision-making processes.

The entire product line consists of three major modules: the Pharmaceutical Knowledge Base, Multi-Agent System, and AI Pipeline Integration Tool,  each addressing one of the three core pain points: knowledge accumulation for pharmaceutical companies, complex clinical analysis, and integration of global pipeline competitive intelligence.

The first module is the pharmaceutical knowledge base, primarily addressing the issues of fragmented and underutilized internal knowledge within pharmaceutical companies.

Pharmaceutical companies have accumulated a vast amount of R&D data over the years, including research papers, clinical reports, academic posters, and project presentation slides. However, this information often exists in unstructured formats, scattered across different departments and systems, making it difficult to quickly retrieve and reuse.

The Pharmato Medical Knowledge Base supports parsing of multiple formats, including PDF, PPT, images, web pages, and Markdown, and has been optimized for high-density information content such as clinical posters and project presentation slides. Leveraging multimodal parsing technology, the system can identify text, images, tables, and key entities such as drugs and targets, transforming previously scattered materials into knowledge assets that are understandable and accessible to AI.

Meanwhile, the platform emphasizes full-text traceability. When generating analytical results, the AI can link to corresponding sources such as literature, paragraphs, or charts, enabling R&D, medical, and regulatory affairs personnel to quickly verify the underlying evidence.

Pharmato Medical Knowledge Base: Operational Case Studies

Currently, the knowledge base covers application scenarios such as literature interpretation, patent analysis, regulatory inquiry, and clinical efficacy analysis, serving pharmaceutical companies across different stages including early-stage research, clinical development, and post-marketing.

If knowledge bases address the question of “where information is located”, then multi-agent systems solve the problem of “how to utilize information”.

Traditional pharmaceutical analysis often requires researchers to complete multiple steps, including literature search, data organization, drug comparison, and report writing, a process that is time-consuming and reliant on individual experience.

Pharmato leverages a multi-agent collaborative architecture to decompose complex tasks into distinct stages, including information collection, data analysis, and result synthesis, while enhancing result reliability through multi-round verification.

Pharmato High-Trust Multi-Agent Operation Case Study

To address the high accuracy demands of pharmaceutical scenarios, Pharmato has set a capability goal of “500 Q&A interactions with zero hallucinations”. Ye Zhongkai introduced that this capability is not reliant on a single model, but is achieved through the combined efforts of high-quality data sources, agent collaboration, and result verification mechanisms.

Compared with general-purpose large language models, the core distinction of pharmaceutical AI agents lies in their integration of specialized data and industry-specific scenarios. By incorporating professional databases such as the U.S. FDA’s Adverse Event Reporting System (FAERS), the platform enables pharmaceutical companies to conduct drug safety analyses, comparative clinical trial assessments, and competitive intelligence research.

For example, in clinical comparison scenarios, the system can automatically extract baseline data, efficacy endpoints, and safety indicators from different trials to generate structured analysis reports, directly serving roles such as MSL (Medical Science Liaison) and MA (Medical Advisor).

In addition to knowledge management and analytical decision-making, pharmaceutical companies face another practical challenge during R&D project initiation and business development (BD): how to gain a clear understanding of the global competitive landscape.

Currently, pharmaceutical companies typically need to consult multiple domestic and international pipeline databases simultaneously; however, discrepancies exist among these databases in terms of drug nomenclature, target classification, and development stages. A Single Database Is Insufficient to Cover Complete Information.

According to Pharmato, in the small cell lung cancer pipeline analysis, after integrating multiple databases including Cortellis, GlobalData, Pharmaprojects, and PharmCube, a total of 1,341 pipelines were identified, with the highest coverage rate from any single database being only 57%.

To address this issue, Pharmato has launched an AI-powered pipeline consolidation tool, featuring its self-developed Pharmato X1 AI Pipeline Consolidation Engine. The tool supports data integration from multiple mainstream new drug databases, helping enterprises consolidate multi-source data through intelligent field matching, drug alias recognition, and standardization of development stages.

The entire toolkit enables full-spectrum integration in just four steps: document upload, intelligent field alignment, manual review and splitting, and preview and export. The system automatically recognizes drug aliases across different databases, standardizes target and development stage annotations, distinguishes original gray-source data from merged blue-standard entries, and supports manual editing and correction. The final comprehensive competitive landscape table can be exported and archived as a long-term corporate data asset.

For pharmaceutical companies, the value of this tool lies not only in reducing the time spent on manual data organization, but more importantly in helping enterprises establish a more comprehensive competitive intelligence system, providing data support for R&D project initiation, BD collaborations, and strategic decision-making.

Product capabilities must ultimately be validated in real-world business scenarios. Currently, Pharmato’s products, including its AI agents and knowledge bases, have been deployed across medical affairs, R&D, and business development (BD) teams within pharmaceutical companies. A user from a pharmaceutical company’s medical affairs team reported that these tools significantly reduced the time costs associated with evidence collection and analysis.

The current development of AI in the pharmaceutical industry is still in the stage of evolving from an efficiency tool to a business tool.

In the past, AI was primarily used to assist with information retrieval, literature organization, and data analysis; however, for pharmaceutical companies, the truly valuable applications lie not merely in accelerating data acquisition, but in empowering R&D, medical affairs, and business development (BD) teams to perform analytical assessments and make informed decisions based on data.

“We aim to develop next-generation multi-agent tools for drug intelligence and clinical decision support,” said Gao Fei. In his view, the key to future competition in pharmaceutical AI lies not merely in model capabilities, but in who can gain a deeper understanding of pharmaceutical business processes and effectively integrate data, domain expertise, and real-world scenarios.

Currently, Pharmato is leveraging its existing data and AI capabilities to explore expansion into more complex clinical scenarios, including Virtual Clinical Trials (VCT) and Virtual Control Arms (VCA), with the aim of further supporting clinical study design, patient recruitment, and post-marketing research.

Meanwhile, the company is also advancing its international expansion strategy, aiming to extend its accumulated pharmaceutical data capabilities and AI solutions to a broader base of overseas pharmaceutical clients.

For Pharmato, the exploration is no longer just about a simple efficiency tool, but rather a new way of collaborating on pharmaceutical information— enable sedimentary data to be understood, accessed, and ultimately serve R&D and clinical decision-making.

In an era where pharmaceutical innovation is increasingly reliant on data-driven approaches, how to truly unlock the value of data will also become a critical issue for AI-empowered advancement in pharmaceutical innovation.