Calibre has released the first artificial intelligence features within the Calibre Research Management System (CalibreRMS). CalibreRMS Intelligence is our name for features within the platform which use various forms of AI for investment research – such as Large Language Models – to provide new capabilities to our users.

Calibre is unique in the RMS market for its holistic investment process approach and delivering a platform which captures all stages of the process, and all activities within each stage. Research notes, company excel models, ESG engagements, scorecards, proxy voting, portfolio positions, pre-trade decision logs, third party data integration such as fundamentals, consensus estimates or carbon emissions. Integrations with Excel, Outlook and Word. Our belief is that the best investment teams have a process, and CalibreRMS can enhance all parts of the process.

We bring this unique perspective to how we integrate AI, enhancing and improving the investment process.

Most “AI for research” tools use public data and know nothing about your process. The result? If all investors use the same models with the same public filings, news, expert calls and transcripts, they all get the same AI-generated research.

The shift we’re making with Calibre Intelligence is simple: AI should live inside the research process, not next to it. When AI works where your notes, models, and portfolios already live, you can improve the quality, consistency and speed of investment decisions across the team to improve the alpha of your investment strategy.

Real world AI for Investment Research

The first new features enabled by CalibreRMS Intelligence are; Intelligent Notes, Auto-Tagging and AI Note Abstracts.

Intelligent Notes

Intelligent notes are built on your own templates using your custom prompts. 

Intelligent Notes Grounded with Documents and Audio files: Attach PDF files such as earnings releases, slide decks, annual and sustainability reports, broker notes, and proxy advisor research. Drop in an audio recording of a company conference call and CalibreRMS converts it to a transcript, complete with speaker diarisation and uses it alongside your files. Your selected AI models and CalibreRMS structured note ‘prompting’ then converts these raw inputs into insights, captured and stored as a structured note alongside all other research notes.

Intelligent Notes Grounded with the Web: Using Perplexity for near-realtime search, CalibreRMS Intelligent Notes can create up-to-date snapshots in a standardised format, fully cited. It can add Business descriptions which include details on product lines or regional segments. Management and Director profiles. Recent controversies. Regulatory changes. Industry news mentions. Again, all captured inside CalibreRMS.

Auto-Tagging

As an investment research system, CalibreRMS has a security master against which all content is tagged. This means notes live in their respective company dashboards. Portfolio views can surface research notes on all holdings. A second level of tagging exists to allow analysts to connect peer companies where the content is relevant. This peer tagging has previously been a manual process, relying on the analyst to tag the companies mentioned. Auto Tagging uses AI entity recognition to cross reference the CalibreRMS global security master database and apply tags to any companies mentioned in the content of the research note, conference call transcript, one-on-one meeting or documents analysed by AI.

Note Abstracts

CalibreRMS can create simple summaries using AI at the top of a note (whether the note is a standard note or an intelligent note). This can be useful for portfolio managers who may be skimming large amounts of research content.

AI Chat (our AI competitors) vs Process Integration (Calibre)

Scenario: post-result first take.

Our competitors: AI Chat tool centric approach: You drag earnings release and slide deck PDFs into a chat, ask for beats and misses, and get a chat response. Next, grab the audio from the conference call, send it to a speech-to-text service to extract the transcript. Copy the transcript and paste it into another chat service and ask for the key points. Maybe you have a structured prompt saved that you can use. Then you need to copy and paste all the key elements into a research note back in your RMS to write up the analyst conclusion.

Tomorrow, the source context is gone (or at best sitting in 3-4 different systems) and you start over. Each team member works in a silo and in 12 months time, there is no way to find or re-create the source components that went into the investment conclusion.

Calibre: Process-integrated approach: You drop the call audio and result pack into the ‘Earnings First-Read’ template. Calibre automatically converts audio to a full text transcript. The template uses your team’s shared/tuned custom prompts to extract headline numbers, guidance changes, management tone, broker alignment and tags any peers mentioned. The note auto-links to positions and flags any open checklist items. Prior research can be integrated. The analyst then annotates their own views directly in the same note and all the research and source material is saved into the RMS tagged to the reporting company and any mentioned peers alongside all the source materials used by CalibreRMS Intelligence. PMs open one view: what changed, where we diverge from consensus, and analyst view. 

CalibreRMS Intelligence: Frontier Models, Institutional Grade Security and Your Own Internal Research

Investment teams can have confidence that these features are built on the same institutional-grade infrastructure as their existing CalibreRMS platform. Decades of providing software to institutional buyside and sellside investors mean Calibre has a deep understanding how features can be both productive and highly secure. This is critical with AI.

Beyond the non-negotiable approach to security and data governance, there are three additional ways CalibreRMS Intelligence is unique in how we integrate AI into our platform:

Complete User Control: teams can control whether CalibreRMS Intelligence is enabled for their users and more importantly have full control over the security and governance of information sent to their own AI models with our Bring Your Own (BYO) API key feature. Teams can remain fully compliant with their internal IT policy by sending all LLM model requests to models hosted inside their own Microsoft Azure tenancy. They can control which provider and models can be made available to their team, and have full transparency on token costs. 

Frontier Models: Some AI vendors fine-tune open-source models with specific narrow tasks within their platform. CalibreRMS Intelligence takes a different approach, enabling user access to all of the cutting edge frontier models from tier-1 labs like OpenAI, Anthropic, Google, Perplexity, Deepgram and others. Your team chooses the model. As these models get stronger, faster and cheaper, the benefits are directly unlocked within the CalibreRMS platform.

Your own internal research: Most AI powered investment research tools unlock new ways of scouring public data sets for insights. CalibreRMS is unique in viewing your own proprietary investment philosophy, process and research as being a more durable source of long term investment edge. You can build highly secure and cutting edge AI directly integrated with your investment process.

CalibreRMS Intelligence features will continue to roll out across the entire investment process platform over the next 12-18 months. Intelligent notes, auto tagging and note summaries are just the first step.