At AlphaSummit 2026, much of the conversation focused on results: where AI is already delivering value for investors, strategists, and finance leaders. Alongside those examples, speakers offered a picture of what comes next. Together, their comments point toward AI that works continuously, connects expertise across teams, and changes how organizations assign and manage work.
Every Covered Company Will Have an AI Agent
At last year’s AlphaSummit, Matt Reustle predicted that agents would become more proactive and more personal. This year, Ian Lieberman, director of research at Talos Asset Management and an early user of AlphaSense’s always-on AI agent SuperAnalyst, described what that could mean for investment research.
I really think that we’re going to get to a place where the team walks in every morning and there’s just a readout. There’s an agent for every name in coverage.
Each company in the team’s coverage universe would get a dedicated agent that understands how the team thinks about it, which KPIs matter, and what to pull from primary research.
That continuous coverage would free analysts to spend more time talking to management teams and experts, reading technical journals, and doing bespoke quantitative analysis. For Lieberman, having an agent for every company is a way to expand coverage while giving people more time for deeper research.
AI Will Move Beyond the Screen
George Lee, co-head of Goldman Sachs Global Institute, expects this year to bring more developments in AI devices.
New intelligence has emerged into the world that allows us to ask more questions, think more broadly, be more creative, and yet we're still bound by our phones or our PCs.
As that intelligence becomes ambient through glasses, pendants, or earbuds, he sees a new horizon for how people use it.
He also pointed to robotics as another frontier for applying AI to the physical world. Lee highlighted Chinese labs’ practical approach to deployment and said he believes they are well ahead of much of the world in robotics.
AI Will Connect Expertise Across Teams
Jazmin Medina, investment partner at D.E. Shaw Ventures, described how she sees AI use evolving inside firms.
Phase one of AI diffusion has really been individual productivity and driving outputs for a particular project. I’m really excited for the next phase, which is essentially collaborative AI decision-making.
As an example of that first phase, she pointed to a signal tracker her team built across 5,000 private companies, enabling research at scale that previously required 40 to 50 hours for a deep dive into a single vertical.
The next phase will address a limit she sees in how expertise spreads. Domain experts have finite time and cannot share their knowledge with everyone in the organization whenever it’s needed. She sees agents helping close that gap.
Medina envisions AI helping teams share expertise across the firm. The private-markets team could explain how it expects a scenario to unfold, while colleagues in public equities, credit, and semiconductor research contribute their perspectives.
Companies Will Need to Manage AI Like a Workforce
Lexi Reese, CEO of Lanai, said that at the current course and speed, work that took 30 years to fully transition to machines in the industrial age will be delegated to AI over the next five to eight years, before companies have the management systems to understand the full consequences.
Her answer is to manage AI the way companies manage people. She described AI as a colleague with varying degrees of autonomy that “can and should be managed the same way,” with goals and performance reviews. Tokens turned AI from a fixed software expense into a variable one.
Tokenomics are looking a lot like labor economics, but what am I getting for this value?
The unit worth measuring is the task. AI, she said, “is not coming for jobs; it comes for verbs.”
She also argued for naming one accountable owner, with budget authority and a goal tied to revenue and profit per head. As she put it, “if everyone owns it, no one owns it.”
Humans Will Stay in the Lead, With Broader Skills
Manish Sharma, chief strategy and services officer at Accenture, rejected the phrase “human in the loop” in favor of “human in the lead.” He said Accenture’s headcount is growing, entry-level hiring continues, and the firm is investing heavily in reskilling.
It is human in the lead every single time.
The challenge is that every person needs more than one skill set. “Typically, you’re a data guy. You’re a technology guy… But you need at least one more element as you think about the future,” Sharma said.
Katie Fogertey, chief management officer at Gluck & Co., offered a complementary perspective: AI can help professionals develop that broader knowledge. “I can’t think of a time where it’s been easier for somebody who sits in marketing to actually become well-versed in operations, and somebody who’s sitting in operations to become well-versed in supply chain.”
George Lee argued that the field needs a “grand unified theory of bottlenecks.” In banking and investing, he said, AI widens the breadth and pace of research, and the bottleneck then moves downstream to human judgment.
Together, these perspectives suggest that AI could expand what individuals can do, while raising the bar for the breadth of knowledge and judgment they bring to their roles. As AI takes on more of the research process, people still need the expertise to question its findings and take responsibility for the decisions those findings inform.
The Road Ahead
Judging by these conversations, the next phase of AI looks woven into everyday work: agents monitoring continuously, expertise reaching across teams, and professionals working across disciplines. For research, that could mean new kinds of competition. Michael Eastwood, director of Americas research at Jefferies, said his closest competitor may well be an AI lab instead of another bank’s analyst.
Even so, the full shift will take time. Sharma estimated it will take four to five years to build an AI-enabled enterprise, with early results expected in six months to two years. Lee put the timing in context, noting that Time named the personal computer its “Machine of the Year” in January 1983, yet it took 10 to 15 years before PCs changed how offices worked.
That gives business leaders room to make deliberate choices about which tasks to delegate, how to measure the results, and where human expertise and judgment will remain essential. As Lee put it, AI can “amplify your creativity, your reach, your curiosity, and your decision making.”
There may soon be an agent for every company or market a team follows. People will still have to decide which findings matter and what to do about them.
Missed the event? Catch all the AlphaSummit 2026 on-demand sessions and register to join us for next year’s conference.
