‍Tariffs, sanctions and policy shifts now move markets as much as earnings. This panel explores whether geoeconomic risk can be measured, modelled and built into the investment process. How are investors turning news, text and expert analysis into usable signals? What can past shocks teach us about forecasts versus outcomes? Where is AI genuinely helping? And as policy uncertainty shapes the outlook for 2027, where should investors be looking next?

This panel will address the drivers behind rising demand for nowcasting products; how accurate these models have proven relative to the official prints they aim to anticipate; the practical criteria funds use to select between competing providers and methodologies (coverage, latency, transparency of inputs, track record through regime changes); and the longer-term trajectory of the space as statistical agencies themselves explore alternative and administrative data sources to modernise their own releases.

Who is really sustaining consumer demand, and what could make it turn? Consumer data offer more than an early read on the next release. Drawing on original research, Filippo Pallotti shows how transaction data challenge the "K-shaped" narrative, explain how inflation shifted wealth from savers to borrowers to keep spending alive, and reveal when rate changes and fiscal transfers start to hit demand. The result: sharper answers on growth, inflation and the timing of policy.

How can developments in physical commodity markets inform systematic research across global equities and futures? This session explores how Mysteel’s commodity prices, inventories, production and industrial fundamentals can be mapped to equity exposures and futures markets. Drawing on historical backtesting, we will discuss applications in equity portfolio research, futures signal development and fundamental monitoring. We will also examine point-in-time alignment and out-of-sample evaluation, offering a practical framework for turning industrial data into model inputs and assessing the strength and stability of historical relationships.

MCP has gone from emerging standard to a major building block for agentic AI in little more than a year. But with rapid adoption has come a proliferation of tools, servers, standards and security questions. This panel looks back at the year of MCP - what has worked, what hasn't, where the ecosystem is heading, and what it could mean for how financial firms connect AI agents to data, research and investment workflows.

Funds expect free trials. Vendors foot the bill. Is that fair? Trials can run for months, eat engineering time and end in silence. Many providers say the sales cycle is stacked against them. Buyers say they cannot commit budget to data they have not tested. This debate puts both sides on stage. Should trials be paid, part-paid or credited against a contract? What makes a trial worth running for either side? And as data moves into AI workflows, should vendors charge for MCP access from day one, or is it just another trial by another name?

AI agents are taking on more of the investment process, from sourcing and testing datasets to generating signals and executing trades. As machines start to buy, evaluate and act on data with little human involvement, the data itself becomes the foundation everything rests on. This panel asks what autonomy means for the future of investment data: whether agents can be trusted to judge data quality and provenance, how licensing and compliance work when the end user is a machine, and where human judgement still matters most. What would it take for firms to trust AI with the data decisions that drive their biggest calls?

Our AltDating sessions facilitate interactions between data buyers and sellers via a program of personalized matchmaking based on your needs.

Slot 1 - 11:05 - 11:20
Slot 2 - 11:20 - 11:35
Slot 3 - 11:35 - 11:50
Slot 4 - 11:50 - 12:05
Slot 5 - 12:05 - 12:20
Slot 6 - 12:20 - 12:35
Slot 7 - 12:35 - 12:50
Slot 8 - 12:50 - 1:05

In this new vendor showcase, hear from a curated selection of emerging data providers as they introduce their latest datasets and product launches shaping the future of alternative data.

‍

‍

Buying the same datasets as everyone else no longer guarantees an edge. So where should investors put their money and effort? Buy means licensing external data; build means investing in the teams and infrastructure to source and extract more value from it; create means generating proprietary data and signals. This panel explores the trade-offs, sourcing challenges and alpha potential of each approach and how GenAI is shifting the balance.

Every fund wants more alternative data. Every data team has a backlog. Todd Fearn demos what happens when an AI agent does the onboarding: acquiring a new alternative dataset, profiling it, flagging sensitive fields, applying quality rules, landing it in the warehouse and cataloging it with full lineage, without ever holding a credential. If a bad file arrives and the agent explains and repairs the break. Datris is an open-source MCP data control plane. This is what governed agents look like.

From research to execution, AI agents are taking on more of the investment process, and agent-to-agent investing raises the prospect of markets where machines increasingly deal with other machines. How far that goes is as much a question about people and institutions as about technology. This panel steps back to ask what autonomy means for the future of investing: the role of human judgement, where responsibility sits, and what it would take to trust machines with the decisions that matter most.

‍This session highlights how generative AI reshaped the workforce in 2026, from hiring activity and layoffs to shifting skill requirements. It then looks ahead to 2027, setting out our predictions for the year: what proportion of AI-related layoffs will be reversed, and how central AI literacy will become across open vacancies.

Private markets are opaque by design. Disclosures are sparse and deals are infrequent, so identifying genuinely comparable companies is one of the hardest problems in valuation. Traditional industry or geography filters often miss what drives value. Dhagash Mehta, Head of Applied AI Research at BlackRock, will share research on a machine learning approach that defines peers by shared valuation drivers. He explains how to handle incomplete private-market data, compare methods, and ensure explainability in due diligence and portfolio construction.

As data volumes surge and AI accelerates how information is consumed, data quality has never been more critical. From quietly revised histories and incorrect entity mappings to shrinking coverage, subtle errors can undermine investment signals. As funds increasingly rely on automated workflows and AI agents, these issues can spread faster and become harder to detect. This panel explores how investors validate datasets, manage emerging AI-related risks, and hold providers accountable for transparency, methodology and timely change notifications.

Our AltDating sessions facilitate interactions between data buyers and sellers via a programme of personalized matchmaking based on your needs.

Slot 9 - 2:20 - 2:35
Slot 10 - 2:35 - 2:50
Slot 11 - 2:50 - 3:05
Slot 12 - 3:05 - 3:20

‍