Vision: To transform investment research with the world leading domain specific multi-agent, trustworthy and compliant AI.
Current Status
• Conducted market
survey of 100+ investment research analysts
• Raised £50k from angels
• Established partnership with FCA, AWS, Nvidia & Neo4j.
• Build a team of 10 Research Scientists, Data Engineers, Dev-Sec-Ops developers and Data Scientists
• Established a panel of 12 investment research experts each with 20+ years of experience from large asset management firms
• Created a community of 50+ alpha testers to assess our MVP
• Engaging with HSBC, Societe Generale, Franklin Templeton, BNP Paribas for design partnerships
Market
As per research published by McKinsey and Company, the financial knowledge market is expected to grow up to $91b by 2028. They estimate that approx 20% of this growth will directly come from AI based platforms for investment management.
Our platform would be useful to all Investment analysts from smaller boutique funds to mid-sized asset management firms to large investment banks. The platform can be used by both buy and sell side instituion in their equity research processes.
Problem or Opportunity
• Traditional research processes relies on individual expertise and fails to utilise the vast majority of the available data leading to inefficient trade ideas
• Latest Gen-AI based solutions often provide unreliable information and limited data security and AI regulatory compliance
• Due to information security risk enterprises face restrictions in leveraging their internal data with ChatGPT like AI platforms
• Key financial services data vendors also prohibit companies from using their data to train / fine-tune Gen-AI based models
Solution (product or service)
• F.InSight synthesises structured & unstructured data, from all available data sources using it’s IP on search, conversational AI & knowledge graph capabilities
• F.Insight securely and privately processes all the data an investor needs to review ensuring data, cloud and AI regulatory compliance
• F.Insight’s security by design architecture ensure that all enterprise data remains within their firewall
• Unlike other Gen-AI based applications F.Insight doesn’t need data for training / fine-tunning LLM+RAG* models instead uses its proprietary Knowledge Base
Competitors
As our platform sits at the cross road of latest AI tech platforms and leading financial services providers, we group our competitors in two buckets.
a) AI tech platforms 1) Open AI 2) Anthropic 3) Google Gemini 4) Mistral AI 5) Databrick's Mosaic 6) TII (Falcon LLMs) 7) Meta (LLaMA)
b) Finacial Services providers 1) Bloomberg GPT 2) Alpha Sense 3) S&P Global 4) 9 Fin 5) SentiMine 6) Sense Street 7) Cognitive-Credit 8) Kensho
Advantages or differentiators
Our platform uses a novel way of leveraging additional data from vendors without using it to fine-tune or train new large language models. This ensures (1) Data vendors are willing to partner as their data is not used in any model training, (2) It improves the overall accuracy of the model responses and (3) It opens up the opportunity to securely and privately create an enterprise Large Language Model without any litigation risks.
Also, we have a panel of 12 domain specialist advisors, each with over 20+ years of experience in the field of investment research from multiple asset classes/sectors. And we have a team of 10 research scientists, data scientists, data engineers and dev-sec-ops professionals from some of the leading research institutes globally.
Business model
Our platform will have 4 modules. Each of these modules will have a basic free tier for the users. We are planning to launch our platform through two commercial models.
(1) SaaS - User will access the application hosted on Fehmida.ai’s cloud instance. The application will use industry specific data curated by Fehmida AI with option for users to upload limited data.
(2) PaaS - The platform will be installed in client’s own infrastructure. This will require initial consultancy + annual maintenance fees.
Both these models will come with core features on data privacy, security, model explainability and personalised learning.
