Traditional data and research purchase arrangements may be holding back the buy-side from using artificial intelligence to take full advantage of new sources of market and other information.
Restrictions associated with broker and data licensing agreements were cited as the biggest barrier to adoption of direct data and research feeds by two-thirds of global asset managers surveyed in a recent study by Substantive Research and Aiera. The investigation also found that just over half of the 35 respondents said that compliance and entitlements were holding them back.The research comes as AI deployments within the buy side progress at pace, offering the opportunity for organisations to more efficiently ingest, process and deploy a broader range and volume of critical data. That argues for a swift resolution of broker research licensing arrangements, said Mike Carrodus, chief executive of Substantive Research.
“The buy side is moving much faster now on AI, but if it wants to ingest broker research at scale then it needs to understand how much more it is going to cost,” Carrodus told Data Management Insight. “For the most highly valued buy/sell side relationships, a research content feed may be included in what asset managers are already paying, but with anyone else we are still in the midst of negotiations.”
Tough Prospect
Management of data and research procurement channels is “incredibly difficult to manage and optimise”, wrote Substantive Research, because of the volume and complexity of the content obtained and the multiple sources from which it originates.
AI can make that easier by improving research discovery, surfacing insights quickly and enhancing analyst productivity. This can give organisations a competitive edge, but it may not come cheap.
“Research budgets are static in Europe and rising slightly in North America, so if there are significant new charges to be met, then something has to give,” Carrodus said. “That could simply be deprioritising certain brokers in order to pay more to others, which could increase research budget concentration from the existing high levels – the top 10 brokers already command 55.1% of investment research budgets.”
The need for a rethink on licensing is being driven by the buy-side’s need for good-quality data to feed into their portfolio management and risk engines. The difficulty is that current provisioning arrangements can mitigate against that, said Gavin Skinner, chief operating officer of Aiera, a content delivery and access platform.
“Buy-side firms overwhelmingly want broker research inside their AI workflows, but today’s licensing, entitlement, and compliance frameworks weren’t designed for machine-readable, AI-driven environments,” Skinner said in a statement.
“Modernising how premium content is governed and delivered is essential to unlocking AI’s full potential while protecting the intellectual property, transparency and commercial value that underpin the research ecosystem.”
Bracing for Change
Research provisioning has become a hot topic in the AI age and brokers and vendors are bracing for changes to their business and operating models as the technology reshapes the way data is generated, distributed and managed. Substantive Research has borne witness to this; it was acquired by Euronext in 2024 as the European stock market operator girded itself to meet the changing needs of market participants.
“Investment research is in a state of flux, with technology changing the focus and capacity of the sell side, and the consumption behaviour of buy-side investment functions also adapting to new formats and evolved search, aggregation and extraction capabilities,” Amrish Ganatra, chief executive of Euronext’s Commcise commission management subsidiary, said at the time.
The latest Substantive and Aiera study found that AI adoption was widespread within the buy-side and it is happening at speed.
More than three-quarters of organisations surveyed said they had adopted generative AI models such as Claude and GPT. Just a little less than a fifth said they had been able to approve, adopt and deploy the models within three months; a third between three and six months; while for a fifth, it took more than half a year.
“What this survey confirms is that in an AI-driven investment world, the sell side retains its core role in enabling the buy side to scale,” Carrodus said in a statement. “While research remains a relationship business with direct analyst access even more crucial now than ever before, written research still needs to be a key resource on portfolio managers’ desktops.”
Subscribe to our newsletter
