AI Due Diligence Report Drafter for Securities



Johannes Gugl - CEO, VIA

“It’s great working with the Kreoh team, together we automated large parts of the report

writing involved in risk assessment for banking securities. Critical but often not given the

attention needed in people’s over communicated work days"


The project aimed to automate significant portions of the due diligence and risk

assessment process, thereby increasing efficiency, reducing human error, and streamlining the

generation of comprehensive due diligence reports.


  • Development of AI agents capable of understanding complex financial information and assisting in the creation of draft sections for due diligence reports.
  • Compatibility with a wide array of file types to ensure that various forms of data can be processed.
  • Implementation as a microservice API, enabling Via Markets to integrate the technology with their existing user interface (UI).

Platform Architecture:

  • Microservice API: The due diligence copilot was built as a microservice API, allowing for seamless integration with Via Markets' existing systems and UI, offering flexibility and scalability.
  • Support for Multiple File Types: The system was engineered to support a diverse range of file formats, accommodating the varied nature of data used in due diligence processes.

AI and Machine Learning:

  • Grounded AI Agents: The AI agents were designed to comprehend the uploaded documents and generate accurate draft reports. These agents are grounded in the specific domain of financial due diligence to ensure that their output adheres to industry standards.
  • Utilisation of Advanced Models: The system employed state-of-the-art language models, including GPT-4 and GPT-3.5-turbo, to provide nuanced and contextually relevant text generation and data analysis.
  • Distillation-Based Querying: A distillation-based querying agent was included to enhance the ability of users to interrogate the data room. This agent can infer connections and findings across different documents, aiding in the identification of risks and opportunities.

User Experience:

  • Integration with Client UI: By providing the AI capabilities through a microservice API, Via Markets maintained control over the user experience, ensuring that the copilot could be integrated into their existing workflows with minimal disruption.
  • By leveraging the latest in AI language models and distillation techniques, the Due Diligence Copilot developed for Via Markets represents a significant advancement in automating the due diligence process within the financial services industry. This solution not only enhances the accuracy and comprehensiveness of risk assessments but also provides Via Markets with a competitive edge in efficiency and technological innovation