Summary of panel discussion: Exploring the AI impacts on research creation, distribution, and discovery

Summary of panel discussion: Exploring the AI impacts on research creation, distribution, and discovery

At the June 2025 all-member meeting, we invited Elizabeth Attah (Bloomberg) and Rahul Ravikrishnan (Eidosmedia) to share their insights on the AI impacts on research creation, distribution, and discovery. Liz is a product manager for Bloomberg’s research & analyst product team, focusing on the Document Search DS platform. Rahul is currently the Head of Product – Research at Eidosmedia. Liz and Rahul have significant expertise in this area; see their bios at the bottom of this article.

A summary of this wonderful panel discussion is below.

Liz helped open the discussion by observing that end users – meaning both research creators and research consumers – have varying levels of comfort, experience, and sophistication regarding AI, and emphasized the importance of helping clients navigate the opportunities that AI can provide.

Since Eidos’ customer base includes those who use their tools for newspaper creation and those who use it for investment research creation, Rahul shared some of the similarities and differences he has seen between these two industries, observing that the investment research side has regulatory considerations that usually necessitate them moving slower.

Trust is an issue that we often hear about when discussing AI, and it was no different on this panel. In explaining ways that AI is affecting the process of searching for and gaining insights from investment research, Liz observed that the trust needs to exist first, and consumers need to feel that they can both trust the firms that are creating the tools and trust that the tools will provide a connection from an AI-created summary to the underlying content. Rahul noted that making it easy for users to access the source material from an AI-generated summary is now considered a baseline expectation, because people want AI tools to help them sift through the vast amount of information available to find that which they want to dig into deeper. The panelists also brought up a different aspect of trust: that both sellside firms and third-party vendors need to have trust in the security of the systems that are being built, otherwise content owners can’t afford to allow their content to be included.

AI is impacting the research creation process as well, from being used to improve the quality of content to helping vet new sources to facilitating the process of ensuring that compliance rules are being met. Rahul pointed out that many of these tools take metadata – such as RIXML tagging – into account, since this information has been added by the publisher and thus carries extra weight.

When weighing in on where they thought the biggest impact of AI would be felt for investment research, the panelists felt that creating a bespoke experience for each individual will have a big impact, across all parts of the research creation, distribution, aggregation, and consumption lifecycle. Another area will be making the difference between high-quality content, customer experience, and product offerings and low-quality ones more obvious, with some vendors taking full advantage of the benefits AI can allow them to provide while others will fall behind; likewise, the end users who find ways to take full advantage of AI tools will also take the lead. Both Rahul and Liz pointed out that individual analysts and individual firms will find different ways to make the most of AI tools, as they find ways that work best for their workflows, products, and systems.

This is just a high-level summary of the insights that Liz and Rahul shared during this panel discussion and the Q&A period that followed; we thank them for being so generous with their time, and thank all meeting participants for helping make the Q&A portion of the meeting so lively.

Elizabeth Attah
Product Manager – Research & Analyst Product Team, Bloomberg
Liz Attah is a product manager for Bloomberg’s research & analyst product team, focusing on the Document Search DS platform. She focuses on document analytics and knowledge extraction. Before joining the product team, Liz managed the global research contributions data group, whose responsibility is managing our global third party research relationships and ensuring their content is available on the terminal.

Rahul Ravikrishnan
Head of Product – Research, Eidosmedia
Rahul Ravikrishnan, with over 15 years of experience in investment research software, is currently the Head of Product – Research at Eidosmedia, a global technology provider powering research production and publication for leading sell-side Investment Banks, Ratings Agencies, Asset and Wealth managers. Throughout his career, Rahul has been dedicated to crafting innovative solutions that meet the evolving needs of investment professionals on both the buy and sell side. Passionate about the intersection of finance and technology, he believes that embedding next-generation technologies across the research value chain is key to driving future investor success and reshaping how the industry collaborates, curates, and consumes investment intelligence.

Artificial intelligence

3rd party data identification working group2

How do we identify components (parts) of research reports that are subject to different rules?

Download the overview document

The RIXML Third-Party Data Identification Working Group working group has been created to discuss how to address the critical issue of identifying third-party content contained within research reports and other research content. The goal of this group is to determine how best to identify third-party data that may require carveouts or other special handling when research is delivered via interactive platforms or fed into AI tools.

This group will focus on the business-side aspects of this critical issue; it will develop best practice guidance and may recommend specific functionality that would be helpful to add to the RIXML Research Standard to ensure that it meets the needs of RIXML member firms and the industry in general.

Who should be involved?

  • sellside firms who wish to enter into arrangements with buyside firms, vendors, or others to allow use of their content in AI tools and/or who wish to ensure that the content they provide on their own websites adheres to their contracts with third-party vendors.
  • buyside firms and other investment research consumers who wish to enter into agreements with sellside firms to incorporate sellside firm research into proprietary NLP or other AI-powered tools or for other non-traditional purposes beyond that which third-party data owners currently permit.
  • the third-party data providers who own the proprietary data in question.
  • product vendors who are partnering with sellside firms to incorporate sellside firm research into proprietary NLP or other AI-powered tools, or for other non-traditional purposes outside of what third-party data owners currently permit.
  • aggregation vendors and others who need clarity on the allowed usage of the research items they receive (and the components contained within them).

Beyond identifying the ownership of third-party content, sellside firms may also want to indicate their own ownership of content that may become componentized, for a variety of reasons.

 Additional background, key tasks, and additional considerations can be found in the working group overview document.

-----

If a table within a research report is populated with third-party data, the publisher's license with the vendor may forbid that data from being fed into an NLP tool. As sellside firms begin to consider allowing trusted buyside and vendor partners to use their research in proprietary AI tools, how do firms flag certain components (parts) within their research that needs to be excluded or is subject to different rules?

 

All-member meeting slides and documents

Below are the slide decks and other meeting resources for our all-member meetings:

June 2025:            Slides

January 2025:       Slides           Key changes in RIXML v3.0 one-pager

September 2024:  Slides

June 2024:             Slides

April 2024:             Slides 

January 2024:       Slides

September 2023:  Slides

June 2023:             Slides

March 2023:           Slides

 

 

RIXML v3.0

Data Dictionary:

A draft version of the RIXML Research Suite Data Dictionary for RIXML Research Standard v3.0 can be found here.

 

Enumeration List Review Documents:

We encourage all member firms to review the enumeration lists below to ensure that the terms your firm needs are available. Each document refers to the relevant sections of the Draft Data Dictionary; it is important to review the indicated sections, as they provide the background and context needed to understand where and how each each enumeration list is used.

START HERE:  Introduction and overview: whether you are reviewing one enumeration list or all of them, please take a minute to review this document. 

Set 4: ESG - MEETING DATE: July 29, 10AM ET

Set 1: publishing content and framing the context - MEETING DATE: August 12, 10 AM ET

Set 2: people, groups, and organizations - MEETING DATE: August 26, 10AM ET

We are starting with ESG because it is a topic we have been discussing for quite some time, so people are familiar with it.

 

In the near future, the following documents will be updated and meetings will be announced:

Set 3: thematic tags including identifiers 

Set 5: subject and specialty

Set 6: asset class 

Set 7: ratings, weightings, estimates, and actions

Set 8: financials 

Set 9 - components, episodes, and related products 

Set 10 - events and interactions 

Set 11 - ISO Codes 

 

XSD (schema) files:

While the v3.0 standard is in development, the schema files are available only to member firms. Please contact the program office to request a copy.

 

 

 

Buyside Working Group

Introduction

The members of the RIXML Buyside Working Group have been meeting to discuss issues surrounding usage of their readership data and to develop a set of usage reporting standards that buyside and sellside firms can use to form their bilateral agreements. The members of this group worked with a number of sellside firms, including RIXML firms and others, to define a set of usage reporting standards that address the needs of both traditional static research content (PDFs, etc.) and the newer, more interactive research content. Significant progress has been made to understand the similarities and differences of the various delivery methods and to set standards that firms can choose to use as the basis of their agreements with their research partners.

Because the needs of all firms are unique, the goal of this initiative was not to create a boilerplate agreement; instead, the goal is to provide a set of standards that buyside and sellside firms can use as the basis for their discussions when developing bilateral agreements regarding how sellside firms can use buyside firms’ readership and content access data of research content. 

This document does not represent the views of any individual firm or of RIXML; instead, it is meant to identify the key usage reporting-related issues firms will need to discuss as they develop their bilateral agreements and to highlight areas where there seems to be broad consensus – and where there does not.

Final document

The final version of the RIXML Usage Reporting Standards document is now available.

You can access it here: RIXML Usage Reporting Standards v1.0

This document was finalized on July 2, 2024 and the content is unchanged from the version posted during our August 16 - September 15, 2024 comment period.

Learn more