At AlphaSummit 2026, AlphaSense’s second annual global user conference, financial industry leaders and professionals came together to share insights and perspectives from their enterprise AI journeys.
As workflows have markedly shifted to autonomous, agentic AI, the consensus brings the challenge to light: Speed and depth are equal table stakes. Decision-grade AI depends on trustable, traceable integrations that accelerate human execution and equally drive conviction.
Among many diverse perspectives, key AI themes emerged: the amplification of human capacity, the new norm of agentic workflows, the importance of judgment and governance, and the diligent orchestration of speed to conviction.
View all sessions from the 2026 AlphaSummit here.
AI Amplifies, Rather Than Diminishes, Human Capacity
A common theme across several sessions at AlphaSummit found that AI adoption amplifies human capacity, rather than replaces it. In the “From Theme to Thesis- AI-Powered Idea Generation for Investors” session, Hala Kosyura, managing director of portfolio strategy at the Canada Pension Plan Investment Board, described how integrated adoption has streamlined and added differentiated value to their investment process.
Kosyura explained: “The new [AlphaSense] screener tools are excellent for our PM or analyst who has an investible universe of 100 names and says ‘give me which companies are following the pattern, for example of sustainable margin expansion, and this and this.’ And suddenly, the kind of initial work that used to be brute force over the span of days and weeks can be compressed to hours.”
While the fundamental workflow remains the same, AI compresses ramp-up, screening, and comparative research from days to hours. This allows idea velocity to flourish, as analysts and portfolio managers are freed up to focus on deeper, more complex ideas that drive alpha. With SuperAnalyst, AlphaSense’s always-on AI agent, workflows run autonomously on behalf of investment teams, so that complex research tasks can be delegated with confidence.
Products like SuperAnalyst are in the mode of aiding humans instead of replacing them. Think of how this can amplify your creativity, productivity, and decision making.
Agentic Workflows Are the New Norm
As AI shifts to more autonomous, agentic functionality, the sentiment across many sessions captured heightened engagement and on-the-ground use cases of agents running daily workflows.
In the “Conviction in Days, Not Weeks” session, Charles DeNoto, Director of Corporate and Investment Banking at MUFG, shared the agentic workflow he set up for morning, evening, and weekly briefings. He leverages SuperAnalyst to develop foundational content such as a 10-page thematic deck, half of which will eventually make its way into client-facing deliverables.
“I don't even have to tell it what to understand. It comes back to me and it has pricing trends, information about the different financings in the sector across debt and equity, information about prepayments, stuff that I really care about, competitive positioning, software acquisitions that they’re making. I can take all that information and tailor it into different work products.. This gives me a base where 40% of it I can repurpose and actually use in a work product, and then the other 60% is just helpful to know,” according to DeNoto.
In “The Future of Work: Your Always-On Analyst” session, Ian Lieberman, partner and director of research at Talos Asset Management, shared how as an early AI adopter, the firm sought to build scalable workflows that kept the organization lean while expanding fundamental coverage. Specifically, the firm values intelligent inference and a trusted, low-friction data foundation.
While large models like Claude can generate insightful analytical inferences, a challenge remains with complex integrations, trust, and verifications, according to Lieberman. Talos therefore implemented AlphaSense as one single source of truth and a repository of research, so that tools like SuperAnalyst can operate natively on that dataset without friction.
AI must haves are a blend of intelligent inference with the right data set. We value lack of friction and having one source of truth that we trust: that’s AlphaSense for us.
Judgment and Governance are Key
While AI has created clear gains in research acceleration, the need for an overarching layer of human judgment is greater than ever. Michael Eastwood, Director of Americas Equity Research at Jefferies, emphasized that equity researchʼs value lies beyond speed in the analystʼs verification, context, and judgment.
Eastwood outlined a real-life scenario in which human analysts and AI exist harmoniously and play equally important roles. In a recent project, an analyst at Jefferies used AlphaSense and other large language models to cover 26 small but complex companies in addition to three sector primers in a month, with AI handling the first 48 hours and a human lens refining output thereafter.
Importantly, Eastwood shared his belief that AI cannot yet “see around corners” or innovate like humans. As such, sell-side intelligence thrives on relevant, contextual insights, networks, and corporate access that span beyond earnings notes.
Every single person on our team is more valuable.. It’s very difficult to be an expert on everything. One day you’re working on a bespoke fintech merger, the next you’re working on a software deal, then you’re working on a digital infrastructure deal. So when you’re able to take AI and synthesize information and understand what’s going well for these businesses, what’s not, I can then converse with our team and actually have them brief me with much better efficiency and better engagement internally. It’s been a tremendously positive catalyst.
Speed with Conviction, Diligently Orchestrated
Throughout sessions at AlphaSummit, speakers indicated that AI research workflows can achieve speed with conviction, with the right pairing of verifiable, traceable data and robust human judgment and governance pillars.
In the “Standing Behind It - Building AI Workflows That Survive Scrutiny” session, Jazmin Medina, investment partner at D.E. Shaw Ventures, described an internally built system covering nearly 5,000 private companies that uses AI to collect sentiment, usage, durability, and commoditization signals across sectors. What previously took 40–50 hours per instance can now be run at scale, allowing for faster debate and conviction speed across teams.
Alanna Crank, COO of global equity research at Jefferies, explained how the firm is using prompt libraries and use cases to institutionalize AI usage. For example, an analyst compressed a coverage workflow from six months to four weeks by leveraging AlphaSense, detailed prompts, a second model to challenge outputs, and intensive human review line by line.
The prep work for traveling to Asia for a semiconductor tour takes 12-18 hours of work. I had a detailed conversation with SuperAnalyst, building it just like that. I consumed two movies on the flight instead.
Conviction as the Competitive Edge
For high-stakes investment research and diligence, speed and depth no longer need to be mutually exclusive. Attendees at this year’s AlphaSummit conference shared the consensus that integrative AI with robust content, data, and governance layers allows this to be a reality, and sets the stage for agentic function and the next chapter of AI evolution.
AI adoption is not what matters, it’s the output. You don’t measure business success based on how many people scanned their badges each day.
Missed this year’s AlphaSummit? Registration is open for AlphaSummit 2027 here.
