A.I. Power Index Party: What A.I.’s Top Leaders Say Comes Next

On Wednesday night, 145 A.I. founders, CEOs, investors and executives gathered at Dante Elizabeth Street for Observer’s inaugural A.I. Power Index Party. The evening celebrated the 2026 A.I. Power Index honorees, and the mood was bullish. But the conversation kept returning to the same question. Now that the technology works, who is it for, and who answers for it?

Jeff McMillan built and led artificial intelligence at Morgan Stanley for more than a decade before launching his own A.I. consulting firm. He told Observer he felt humbled to be in the room. “A.I. can serve good, or it can serve evil,” he said. “And that’s a choice.” Paraphrasing Pope Leo XIV’s recent remarks on A.I., he added, “A.I. should serve humanity. Humanity should not serve A.I.”

McMillan places the responsibility on the companies deploying the technology. “We’re spending a lot of time worried about what Sam Altman and Anthropic do. But they’re just the mechanism,” he said. “It’s the companies that will deploy this that have the ultimate responsibility.” He is optimistic that A.I. will create new kinds of work, including jobs managing and supervising A.I. agents, but the pace worries him. “That’s the good story. The bad story is that’s going to happen very quickly,” he said. “What is the role of governments, companies and educational institutions to make sure those people don’t get left behind?”

Anusha Dandapani, United Nations International Computing Center chief data and A.I. officer, shared the concern. “We need to avoid a world in which adoption moves faster than accountability,” she told Observer. “A.I. should scale human capability, but it should not dilute human accountability.”

Shortly after 7 p.m., Brennan Pothetes, founder of Infinity Constellation, took the podium and the room hushed. The “turmoil around A.I. and A.I. adoption” was why he wanted to be the evening’s presenting sponsor, Pothetes said; he was drawn to awards that offered “shining examples” of A.I. creating real leverage in business, and that championed what he called “the good capitalist.” Pothetes pointed to the $6 trillion global services market as ripe for rewriting, and to the gap between capability and adoption.

“A model is not a business. Technology is not a workflow. You can’t bolt A.I. onto an old process and call it a day—not if the customer doesn’t trust it, the team doesn’t use it, and it doesn’t change the P&L,” he said. The unglamorous work, in his view, still comes down to human judgment and operating expertise. “The people being recognized tonight are making that shift. They’re turning A.I. from a shiny demo into how work gets done. Tonight, we’re not celebrating promise, we’re celebrating proof.”

The Enterprise Award

The first proof belonged to Shawn Edwards. Bloomberg’s CTO received the 2026 A.I. Enterprise Award for his work turning the Bloomberg Terminal into an A.I.-native system without losing the trust of the financial services industry that depends on it. That work included a series of A.I. releases that culminated in ASKB, a conversational A.I. interface.

In accepting, Edwards spoke about building A.I. capable of making “mission-critical decisions.” He said ASKB had received a “phenomenal” response. “In all my years of working, I’ve never seen a technology have such a profound impact so fast,” he said. “It’s a great time to be in finance, and it’s a great time to be in tech.”

The information Bloomberg’s customers needed was often already there, Edwards said, but there was too much of it for any one person to process. “The time it takes to explore one idea in the past, now you can explore ten,” he said. The “spark of an idea” still has to come from a person. He rejects the notion that A.I. breeds cognitive laziness. “We’re seeing a lot of proof that A.I. is an amplifier,” he said. “The people with the most creativity, the most intelligence, the ones that have the experience, the ones that can draw the connections, they use A.I. at a rate that makes them even better than everybody else.”

He does agree that jobs are at stake. Work built on data collection or basic spreadsheets was already under threat, he said, and advances in robotics could eventually reach manual labor. That raises a training problem: if A.I. takes over the simpler tasks, how do you develop the next generation of portfolio managers, analysts and programmers? Edwards told Observer he is working through that question with the dean of Columbia Engineering. Bloomberg is still hiring young people, he said, but “we have to introduce these tools right away and skip the part where they’re gathering and doing the simple things. We thought that was a way to train them. Now we have to rethink that.”

He also raised the environmental cost. “I’m very concerned with the climate impact of all of this,” he said. He hopes models will become smaller and more power-efficient. “You don’t always need the biggest model.”

For Joe Zhao, founder and managing partner of Millennia Capital, the choice to recognize Edwards was easy. An advisor to Zhao’s firm presented the award and made the case in geopolitical terms. “Shawn has set a new standard that the rest of the financial data industry must now meet,” the advisor said. The argument was that the Terminal, as American-built, American-controlled infrastructure running on U.S. communication rails, helps keep global capital flowing through U.S. markets and the dollar system. “What I mean is that the people at Bloomberg, even if they’re not aware of it, do much of the causal work of sustaining U.S. influence and dollar dominance in the broader world, just like Fedwire or the U.S. legal system does.”

The Transformer Award

The second honor, the A.I. Transformer Award, went to Candidly founder and CEO Laurel Taylor. In presenting the award, Candidly CMO Meera Oliva took a moment to define the mission-driven company’s work as moving A.I. “out of the chatbot and into the ledger,” using real financial data to guide people through student debt, retirement savings and benefits decisions that once required a human advisor many people could not afford.

Visibly moved, Taylor embraced Oliva, then spoke to Candidly’s mission “to crush debt and build wealth,” which grew out of her own experience. “This is my path and purpose,” she said. “I have personally lived this problem.” Her mother, a highly educated social worker who earned $24,000 a year, taught her that education was “the way up and out.” Taylor financed her own education with private and federal loans, scholarships and grants. What she hadn’t understood was what paying down that debt would cost her.

“I went to MIT. I’m good at math. And I still missed that memo—because moving through your finances sequentially, debt first, wealth later, isn’t just emotionally satisfying. It’s financially devastating, and it’s what over 80 percent of users do that are stuck in the education financing labor complex,” Taylor said. “Thus, Candidly was born.”

A decade on, Taylor pitches Candidly as the A.I. layer that banks and financial institutions build on. “We help the largest financial institutions in the world leverage the data and distribution they already have and turn it into candid, authentic, one-to-one conversations with every person they serve,” she said. Candidly, she said, is “driving a 121 percent lift in offer uptake, a 40 percent reduction in cost to serve,” and has “generated over $2 billion of impact for the people using our platform.” The room erupted in applause.

“Yes, the 1 percent, we’re all excited about the wealth transfer. That’s $30 trillion,” Taylor said. “But let’s not forget to serve the 99 percent and to bring them up and along with us and along the way.”

Earlier that day, Taylor told Observer that the question of who the technology serves is a leadership decision. “We have no excuse not to offer the same level of expert care and compassion that we give to the ultra high net worth,” she said. “It’s a choice by leaders in the industry as to who they’re going to serve and what the mission of their company is.”

She described Candidly’s A.I. tool, CAIT, pronounced “Kate,” as “empathy-rich, judgment-free.” Users often arrive with roughly a seventh-grade level of financial literacy. “The meaning that we’re assigning to that,” Taylor explained, “is not that the user is at a seventh-grade literacy level, but that they want to have a conversation using simple language.” A typical question: “Tell me what’s up with the SAVE program I was put into under the Biden administration. Is it still available?”

Candidly currently reaches users through employers and financial services partners. That will change soon. “Later this year, we’re going to open up that front door direct to consumer,” Taylor said. She called the honor “a symptom of values.” Telling her own story, she said, had taken her “to kind of a visceral place. It’s very raw for me.”

Who It’s For

The idea of A.I. as judgment-free came up again with Chris Silver, chief revenue officer of Parloa. The Berlin-based startup tripled its valuation from $1 billion to $3 billion between May 2025 and January 2026, and its co-founder and CEO, Malte Kosub, is a 2026 A.I. Power Index honoree. Parloa builds A.I. agents that handle customer service calls, chats and messages for enterprises. Silver said the product is voice-first because contact center volume still runs predominantly on the phone.

Silver cited a global financial services company that used Parloa for debt collection and saw its recovery rate rise by more than 50 percent. “You remove the element of speaking to the human,” he said. “So you’re removing human judgment.” The reason, he said, is that “humans are embarrassed to talk to another human and tell them about their negative financial situations.” The company could then offer payment plans suited to a customer’s circumstances “without them having to overexplain the excuse and the root cause and the reality of why they can’t pay what they committed to paying.”

Wendy Diamond, founder of the Women’s Entrepreneurship Day Organization, spoke about bringing A.I. training to women in countries including Afghanistan, Haiti, Pakistan and Honduras. In Afghanistan, she said, women gather privately in groups of about 20 because they can’t afford individual internet access. “What’s sweet about it is it opens their minds,” she said. Some now find work through Upwork and do it from home. Then she leaned forward. “You know what I learned? The Taliban doesn’t understand. They don’t know WhatsApp. Think about that.” A moment later, she pulled out her phone. “Look what we’re doing in Zimbabwe,” she said. She showed Observer a video of women her organization is teaching to turn tomatoes into ketchup and tomato sauce, and to market what they make with A.I.

The Competitive Stakes

For the evening’s co-host, Neal Batra, the stakes are national. “Pandora’s box has been opened,” he told Observer. “As a country, America must remain technologically competitive, both economically and as a matter of national security. Some of the greatest advancements are on the immediate horizon.” Batra is an attorney, early-stage investor and film executive who had long considered A.I. inevitable. When the technology crossed from theory into reality in 2022, he reassessed nearly everything and went looking for companies at the front of A.I., robotics, defense and biotech.

“As Americans, we will have to continuously reposition ourselves over the next several years toward where human contribution remains most valuable in an economy that is becoming exponentially automated. For now, judgment, taste, trust, interpersonal ability, distribution and IP remain among the strongest moats,” Batra said. “Innovation must also fall within the confines of the Constitution as we grapple with what a reasonable expectation of privacy means in the A.I. era. Capitalism will also be tested as automation reshapes both labor and the conditions under which people and businesses can realistically compete.”

Zhao sees A.I. as the latest step in a line running from computers to the internet, mobile and the cloud. “What is A.I.? A.I. is just the children and the grandchildren of computers, mobile and cloud,” he said. “Hopefully, the private sector and the public sector can work together to come up with the right adequate standards to allow the technology to flourish, but in a way that’s responsible.”

Not everyone claimed to know where it all leads. “This is the most sure I’ve been in my career of 20 years in tech that this is one of the most important times in the history of innovation,” said Noah Lichtenstein of Crossover VC, “but I’ve never been less sure about where the winners are going to come from.”

The Art of It

“A.I. will bring a fundamental shift to the art world, but not in the production sense,” art market economist Magnus Resch told Observer. Resch has built a career examining what the art world would rather not discuss: which artists break through and why, how galleries stay afloat, and what drives collectors to buy. His newest project revives Magnus, an app he first released in 2013 as a guide to exhibitions and art-world events. It went dark during the pandemic. Now back in beta, it has been rebuilt around A.I. and pairs image recognition with a database of art prices. Point a phone at a work and, within seconds, the app returns the artist, the title and, crucially, the price. Resch calls it the “Shazam for art.”

“Artists will not be replaced by A.I.,” he said. “But A.I., using the right data, will help to convert visitors into buyers by helping them to identify what they like and to give them confidence in the purchase decision.”

Jeffrey Yin, CEO of the recently merged Artsy and Artnet, described an industry that embraces A.I. and is wary of it at the same time. Inside his company, A.I. has sped up coding, product development and internal workflows. Some work, he said, hasn’t moved at all, and journalism is his example. “You have to actually write a story, right?” he said. “You’ve got to go find the news, collect news and write. That hasn’t changed.”

Inside Dante

The evening’s setting had a story of its own. For more than a century, the original Caffè Dante was a fixture of Greenwich Village, a place where the front window was reserved every afternoon at four for neighborhood widows, who ate for free. In 2015, Australian couple Linden Pride and Natalie Hudson bought the café and were warned they would ruin it with flat whites and avocado toast. Pride has said he never expected to become “a custodian of such a public institution,” where “the people have so many memories.” The couple won over the locals, and in 2019, Dante was named the best bar in the world. Dante Elizabeth Street, the newest of the group’s six restaurants, opens next week, on Tuesday, Sept. 29.

Pride said Hudson designed the space, with its “classic elevated European sensibility” and soft pistachio-green palette. A wall of arches separated the front of the room from the back, with blooming bouquets set into the cutouts. The feeling was both grand and cozy, or what executive chef Daniel Garwood called hygge, a Danish word. “You kind of get that vibe,” he said, “even though it’s an uplifting space. It’s also very comfortable.”

Guests gathered around the horseshoe bar as servers wound through the room with caviar tarts topped with lemon emulsion, champagne pearls and Polish caviar; Wagyu short-rib skewers; and cacio e pepe lasagna served on spoons.

After thanking the evening’s partners and sponsors, Observer partnerships director Justin Arnold announced that “FLYBLACK Jets is giving everyone in this room a travel credit tonight towards their first charter.” Then he quipped, “With the way that commercial flights have been going lately, FLYBLACK may have picked the easiest room in New York to sell on a private jet.”

FLYBLACK’s own pitch echoed the one the honorees had been making all night: technology works best with a person behind it. “We’re marrying that cutting-edge marketplace technology with human expertise,” said partner Nick Rouner. “Clients of ours are taking advantage of everything we have on the tech stack, but they have a real person that they can call, and they can email. We know that your kid has a peanut allergy. We know that your wife likes Gavi, but not another Piedmont wine, whatever it is. It gets super, super granular. But at that level, what you’re paying for is the comfort, it’s the convenience. And it’s working with people who know what you like.”