AI in Drug Discovery May Be Having Its 'ChatGPT Moment': DIP Emerges as the OpenAI of Pharma, Reshaping Global R&D
In 2025, the field of new drug development is ushering in its own “ChatGPT moment.””。
Over 50% of the costs in the clinical development phase of new drug R&D are centered on the generation and dissemination of various types of documents.Generative AI has transcended its early limitations as a mere tool for text translation or simple writing, evolving into a native intelligent platform capable of deeply restructuring the entire drug discovery and development process. This counter-consensus approach, “starting from text,” is opening new frontiers in the digitalization of the pharmaceutical industry: by intelligently understanding and generating vast amounts of textual data, such as clinical study protocols and regulatory submission documents, AI is poised to drive leapfrog innovation in drug R&D, enabling breakthroughs from 0 to 1.
Deep Intelligent Pharma (DIP) is precisely the leader of this transformation.Since 2024, DIP, a China-based enterprise, has continued to “shake up” the overseas pharmaceutical industry. Widely regarded as the “OpenAI” of the pharma world, its generative AI platform is liberating new drug R&D from cumbersome documentation, achieving a qualitative leap in both efficiency and accuracy:Document generation time reduced by over 90%; regulatory submission speed increased by 75%.The innovation of DIP extends beyond traditional documents, encompassing databases, data analytics, and statistical programming within the scope of its “large document” intelligent processing.
At Microsoft Build 2025, DIP, as the sole invited speaker from the Asia-Pacific region in the field of AI-driven pharmaceuticals, delivered a groundbreaking presentation featuring real-world examples of AI-based drug discovery.”Nationwide,Shared a case study on the application of DIP’s generative AI in drug R&D—where AI automatically drafted clinical trial protocols, achieving “zero revisions” and securing first-time approval from Japan’s PMDA, thereby demonstrating significant potential for deep optimization of traditional R&D processes.
DIP’s Presentation at Microsoft Build 2025
In this exclusive interview, Ms. Li Xing, Founder and CEO of DIP, takes us behind the scenes of the “ChatGPT moment.”——How does DIP leverage its multi-agent architecture, a comprehensive product matrix covering the entire value chain, and “general contractor”-style one-stop services to build unique competitiveness in the AI drug discovery sector? From technological innovation to commercial implementation, what international perspective and long-term strategic thinking does DIP embody?
1Guided by “real-world needs” and leveraging the “capabilities and boundaries of self-developed models,” we build a product ecosystem that enables one-click generation and accelerates the entire value chain.
Pharmaceutical R&D is a field that relies heavily on textual information. From literature reviews for new drug targets and the design of clinical trial protocols to the preparation of regulatory submission dossiers, every stage of the R&D process generates a vast volume of specialized documents. In practice, these “intangible text assets,” which hold immense value, have often been overlooked in the past. This neglect has not only resulted in cognitive limitations but also led to a lack of systematic management and efficient application tools needed to break through the bottlenecks posed by massive, cumbersome, and lengthy documentation. For instance, a complete clinical trial protocol often requires teams to spend months repeatedly drafting and revising it. Furthermore, new drug registration submission materials must be prepared in multiple versions tailored to different regulatory requirements, making the process even more complex, tedious, and prone to errors.
The DIP team recognized that if AI could comprehend and directly generate professional texts, it would significantly enhance the R&D efficiency and regulatory compliance accuracy of pharmaceutical companies, while reducing the labor and time costs associated with critical processes such as registration filings.Li Xing stated, “Rather than jumping straight into molecular design or screening, we first transform the ‘paperwork’ in the R&D process into a stage where AI can truly shine. Empowered by generative AI, text intelligence not only automates document drafting but also generates innovative insights that even human experts might overlook—such as designing novel clinical endpoints for rare disease trials and automatically producing well-reasoned supporting materials for regulatory consultation meetings. Practice has proven this to be a highly powerful entry point, critical for improving both the success rate and speed of new drug development.”
From this starting point, DIP focuses on building an AI-native multi-agent architecture in the pharmaceutical field. “In simple terms, instead of relying on a single large language model to solve all problems, we have developed a set of collaborative intelligent agents, with each Agent specializing in handling different tasks within the R&D process.”Li Xing mentioned that some agents are responsible for reading literature and generating hypotheses (Hypothesis Agent), others focus on drafting clinical protocols and research reports (Protocol Agent, Writing-X platform), some handle automatic translation (Translation-X platform), and still others are tasked with data validation and compliance checks (Validation Agent).
A team of AI agents with distinct functions collaborates in real time, leveraging a series of highly automated integrations to deliver standardized data outputs across the entire pharmaceutical R&D lifecycle. This approach significantly reduces manual intervention, effectively replacing the traditional “human patchwork” process characterized by fragmented, inefficient manual documentation and collaboration at various stages.
When R&D personnel need to draft a clinical trial protocol for a new drug, the DIP platform automatically invokes the corresponding agents: first, a language model interprets the trial background and objectives to outline the protocol framework; then, Writing-X automatically generates the text; next, the Validation Agent performs real-time proofreading to ensure compliance with formatting and regulatory requirements; finally, Translation-X can rapidly translate the protocol into multiple languages, including Chinese, English, and Japanese.
Schematic Diagram of DIP’s Proprietary AI Multi-Agent Platform Architecture
“In a sense, DIP is creating an ‘AI writer’ for the field of pharmaceutical R&D, enabling AI to think and write like an expert. This is why many partners describe us as building the ChatGPT or OpenAI of the pharmaceutical industry, hoping to leverage our technology to drive a paradigm shift in R&D,” emphasized Li Xing.Trained on numerous projects involving over 1,000 pharmaceutical companies and CROs worldwide, the DIP foundational large language model has progressively established robust capabilities in understanding and generating medical texts. This has significantly reduced error rates in generated content, with output quality and professionalism rivaling those of senior human experts. Meanwhile, the model demonstrates exceptional generation efficiency, requiring only minimal time to acquire knowledge in new domains.
Specifically, DIP addresses the text processing needs of pharmaceutical companies across various stages of research and development.Design a product matrix that covers the entire drug development process, truly achieving “one-click generation and full-chain collaboration”:
Writing-X Intelligent Automated Writing Platform
Most documents required by pharmaceutical companies from project initiation to market launch, including clinical trial protocols, Clinical Study Reports (CSRs), Common Technical Document (CTD) modules, and Periodic Safety Update Reports (PSURs), can be automatically drafted by Writing-X, with only minor human revisions and confirmation needed.
Translation-X Medical-Grade Machine Translation Engine
Tailored for the multilingual environments of pharmaceutical companies, DIP’s Translation-X features deep customization of medical terminology and stylistic conventions, delivering translation accuracy that far surpasses general-purpose tools. It also integrates seamlessly with Writing-X to enable an integrated workflow combining content creation and translation.
From a large-language-model perspective, DIP has also developed a suite of intelligent solutions built around two core platforms, creating a unified text matrix that covers all “text-related” scenarios across the entire drug R&D value chain:
It is for this reason that Li Xing vividly compares DIP’s full product matrix to“AI Agent Foreman”——“Just as the general contractor in a construction project must coordinate the collaboration of various subcontractors.”For pharmaceutical companies, this means they no longer need to outsource different stages to traditional CROs or SaaS-based CRO software providers; instead, they can benefit from a one-stop intelligent large-text production service.
2AI “Employees” Hit the Job Market, Launching a “No Satisfaction, No Payment!” Model That Shakes Up the Industry Landscape
In the AI era, Deep Intelligence Pharma (DIP) is spearheading innovation in pharmaceutical R&D service models, with its core focus on enabling rapid, customized delivery.Upon receipt of client requirements, DIP’s AI engineers can rapidly assemble project-specific agents from an extensive module library, achieving efficiency and flexibility akin to building with blocks. This “efficiency-first” philosophy permeates the entire value chain, manifesting not only in the AI technology itself but also integrated into every aspect of service delivery.
DIP fully recognizes that efficiency is the lifeline of drug development. Therefore, it consistently positions itself as an enabler, focusing on refining AI model products without engaging in the actual creation of new drugs. To ensure timely responsiveness and in-depth support, DIP has established an operational system that combines global deployment with localized support. With teams based in China, Japan, Singapore, and other locations, DIP covers major pharmaceutical markets in Asia, North America, and Europe, enabling seamless 24/7 cross-time-zone and cross-language collaboration."Keeping Services Online"It is their promise: when the Japan team clocks out, the Beijing team takes over; when U.S. clients encounter issues, the Singapore team stands ready to ensure immediate response to customer needs.
This agile customization capability stems from DIP’s profound understanding of the distinct styles, therapeutic areas, and competitive landscapes of various pharmaceutical companies.At the outset of the engagement, the expert team delves into the client’s existing documentation framework and SOP standards to train and fine-tune AI models in a targeted manner. By flexibly combining agents according to specific needs, they deliver highly customized solutions. More importantly, this modular architecture enables the solution to adapt and evolve in real time as new client requirements emerge, with many new features even co-created alongside established clients.
The significant cost reduction brought by AI enables DIP to deliver service capacity and value that are nearly “UNLIMITED.”This is also reflected in its innovative business model: DIP was the first in the industry to introduce a “no satisfaction, no payment” commitment, directly addressing customers’ real needs and sharing risks with them. In addition, the company has implemented payment schemes that combine performance-based pricing with subscription models. For example, DIP charges fees only after customers successfully complete clinical trial protocol drafting on its platform; alternatively, customers may opt for an annual subscription with tiered pricing based on document output volume. This model closely aligns DIP’s revenue with customers’ regulatory approval success rates, truly achieving a win-win outcome for both parties and driving a customer repurchase rate of over 99%. This fully demonstratesEmpowered by AI, efficient, customized, large-scale, and truly customer-centric services have become possible.
By integrating modular and customized solutions, DIP has also unlocked the innovative potential of AIGC in “zero-to-one” development through practical application.Pharmaceutical innovation company Ayumo has developed a gait analysis software for assessing human health. As there were no precedent metrics for this type of gait analysis, the trial design required identifying an entirely new clinical endpoint and protocol. DIP’s reasoning model, by reviewing relevant literature and public databases,A complete Phase I clinical trial protocol and endpoint setting recommendations were automatically generated within three days.Among them, in terms of the innovative design of primary endpoints, AI proposesSeveral novel combinations of gait metrics, previously unconsidered by researchers, have been identified, demonstrating both scientific rigor and innovation.The experimental design based on this premise was promptly accepted by the PMDA, which specifically inquired about the rationale for the endpoint selection.
In nearly uncharted territory, by integrating and extrapolating from existing textual knowledge, AIGC has identified pathways that human experts might overlook, demonstrating the potential to surpass traditional approaches.—It represents not only a significant boost in efficiency and speed, but also a leap forward in innovation and precision. Li Xing believes, “AI is like a young genius sprinting at high speed, while human experts are like seasoned mentors. Therefore, human-AI collaboration can unleash more powerful innovative energy, combining strengths to get things done both quickly and steadily.”
3Adhering to an internationalized, tripartite strategy, we serve as a long-term companion to the pharmaceutical industry.
In the years that have seen it advance in tandem with the global pharmaceutical industry, Li Xing and her team have witnessed many changes, with an increasing number of AI-enabled application scenarios emerging. Over the past five years, Chinese innovative drug companies have increasingly expanded overseas, conducting regulatory filings and clinical trials abroad, and establishing pipeline collaborations and business development (BD) deals with international pharmaceutical giants. However, cross-language and cross-cultural communication barriers, along with vastly different regulatory requirements and standard frameworks, have often posed significant challenges for these enterprises.
Li Xing believes that AI will become a standard tool for Chinese pharmaceutical companies expanding overseas. On one hand, DIP has accumulated extensive expertise in cross-language and cross-regulatory support, based on its positioning to serve the global pharmaceutical industry.Writing-X and Translation-X are deeply optimized for Chinese, English, and Japanese multilingual support, capable of automatically translating Chinese clinical trial documentation into English or Japanese files that comply with the terminology standards of the FDA, EMA, or PMDA. Compared to traditional translation agencies or human translators, these tools are not only several orders of magnitude faster but also “medically savvy,” ensuring highly professional and consistent terminology and phrasing, thereby reducing the risk of misinterpretation. Furthermore, the platform has built-in compliance rule validation functions for major regulatory authorities, such as the U.S. FDA’s electronic Common Technical Document (eCTD) submission format requirements, European data privacy compliance requirements, and the Japanese PMDA’s list of common questions.
On the other hand, “going global” is not merely about language translation; it also involves aligning data standards and shifting regulatory mindsets.In terms of communication, DIP also assists some clients in drafting emails and meeting materials for cross-border collaborations, providing real-time translation support akin to simultaneous interpretation. Li Xing stated, “The value of DIP lies in serving as a bridge, helping Chinese pharmaceutical companies transform local data and documentation into internationally recognized language and standards.” A business development head at a domestic pharmaceutical company once remarked with appreciation, “With DIP, we feel more confident when negotiating collaborations with foreign partners. In the past, they worried about suffering disadvantages due to inadequate communication; now, with AI assistance, both their materials and expressions are highly professional.”This is the change we aim to bring about—helping Chinese innovation integrate into the global landscape, while also introducing advanced global expertise into China.”
In Li Xing’s view, DIP is first and foremost an AI-driven technology company. Within the application scenario of the pharmaceutical industry, she defines DIP as“Platform-Based Service Provider Empowering the Entire Pharmaceutical R&D Process”. From grassroots technological innovation to building a proprietary AI-native multi-agent platform, and from identifying genuine market needs to implementing real-world scenarios and validating business models, DIP hasServedServing 1,000+ pharmaceutical companies worldwide, we have delivered over 20,000 regulatory submission documents, processed more than 5 billion words of medical and pharmaceutical text, and achieved a customer renewal and expansion rate exceeding 99%.In terms of capital, DeepGlow has garnered favor from top-tier investors such as Sequoia Capital and ZhenFund, having completed four rounds of financing.
In terms of technological innovation, DIP will continue to explore large medical models and introduce multimodal AI capabilities, such as image generation and molecular generation, to enrich platform modules and enhance intelligence.We have established in-depth R&D collaborations with Microsoft Research Asia, Google, and others to tackle the development of next-generation AI models for pharmaceuticals.
In terms of product and scenario expansion, DIP plans to build a series of AI-native platforms, evolving existing point-specific functionalities into comprehensive solutions to construct a highly intelligent, systematic, and sustainable overall operational platform. Meanwhile, it will expand its platform-based solutions in vertical fields such as pharmacovigilance, human genetic resource management, statistical programming, and clinical pharmacology.
In terms of global expansion, leveraging its business foundation in the Asian and U.S. markets, DIP will next focus its efforts on the European market. Aligning with the major trend of Chinese innovative drugs going global and international pharmaceutical collaboration, DIP will seize this momentum to comprehensively expand its service network, striving to become a standard platform for R&D teams of multinational pharmaceutical companies.
At the end of the interview, Li Xing shared her perspective on long-termism—DIP must always resonate in sync with the development of the pharmaceutical industry, continuously adjusting its direction and strategies based on actual needs.Pharmaceutical R&D is a protracted and serious endeavor, as well as a technology-driven and rapidly evolving innovative industry. It is for this reason that DIP needs to adopt flexible and pragmatic principles,Maintaining Vitality Amidst Change。
She said, “DIP hopes to become aBuilt to Last“...enterprises: on one hand, driving industry progress through technological innovation; on the other, achieving sustainable growth via a robust business model. As long as we continue to create value for customers, both commercial returns and social impact will naturally follow. We hope that in ten or twenty years, when we look back, DIP will have become deeply embedded in the foundational infrastructure of new drug R&D, much like how pharmaceutical companies today cannot operate without Office software.”
Starting from the unconventional and counter-consensus premise of “textual intelligence,” DIP has positioned itself at the forefront of the AI-driven transformation of the pharmaceutical industry. Yet, this is only the beginning—as AI technology matures and the digital wave advances, greater opportunities and leaps lie ahead. DIP will join forces with more like-minded partners to witness and shape this new era, in which AI is reshaping pharmaceutical R&D.