By Jorge Mercado Wednesday, October 28th, 2026

Santa Barbara’s Cabrillo Pavilion was packed out on Oct. 21 as nearly 200 people gathered to hear Ventech’s panel titled “The Future of Work: Humans + AI Agents,” and the beginning of the program already provided an interesting insight into how many people are using AI.

Moderator and Ventech board member Mike Wald asked a simple question during his introductory speech: “How many people are, at least once a week, actively engaged with a Chatbot?” 

Hundreds of people in the room — about 90% — shot their hands straight up. He noted that, for most people, that is what they think AI is. But Wald noted that AI has taken different forms over the last four to five years.

Before, most of what AI was being used for is what he calls “party trips,” playing with photos, fake videos, “AI slop,” he said.

He talked about AI agents — AI that outperforms something like a chatbot by actually having the ability to receive a goal, break it down into actionable steps and accomplish it. He said that these agents are “the real AI we think about when we see AI in movies.”

“Now, we’re seeing these agents and tools and different systems pop up every day and they’re handling rather complex tasks,” Wald said. “And that’s going to mean different things for different groups of people.”

According to McKinsey Global Surveys, 62% of companies are testing AI agents at work, with most of those being larger firms. He noted that predictions suggest that by the end of 2028, 113 billion agents will be working together across different apps.

“We’re going to continue seeing this very quickly,” he said. 

As a result, more energy will be required.

“By 2030, data centers are going to be using more power than Japan,” Wald said. 

Panelists for the event included Josh Narva, a partner at Riviera Partners, Diana Pereira, the CEO and founder of Balance Financial Management and Jimi Smoot, CEO of Assist Technologies. All three provided insight into how small businesses are incorporating AI and what that could mean for the general workforce.

Smoot, who is working to add AI into his workflows, noted that while AI is more prevalent, people should still be skeptical to fully rely on it. 

“It can be really good at getting you something that looks plausible, but could be entirely wrong,” he said. 

Narva agreed, saying that some bots can be “wildly overconfident.” 

“In my business, you might cut to page 7 of an email listing seven different companies and you can ask Claude, ‘Give me a summary of this exchange and four highlights,’ and it will hallucinate; it will confuse the company names,” Narva said. “You have to use and apply your own common sense.”

Pereira agreed with the other two panelists, but also noted why it is important to have ways to verify the information it is giving you and also to build the AI model over time and not just trust it right away.

“You build it with ways to validate information, the same way you hire a new employee and you train them,” she said. “You have to put guardrails in place and then you can rely on it.”

“People treat these agents as if it’s this special thing. People give AI this task that they would never give a person, but they expect it to come correct,” Smoot said. “When you hire someone, you don’t expect some really complex plans on your desk, you have to do the same thing with AI.”

Asked about how companies are handling having AI teams, Narva noted he doesn’t believe a company needs a chief AI officer.

Narva noted that his firm launched an AI-related survey last year and one of the key findings was that in 61% of cases, the company’s CTO was the primary AI owner, a chief product officer in 17% of cases and a chief AI officer in 20% of cases, down from the year prior. 

“I think if you’re a small company, this is the CEO’s job. Technical or not, you need to understand the buck stops with you; you have to make these decisions, this is what leaders do,” he said.

Pereira said she believes the biggest opportunities for AI are in intensive, error-prone research work. She noted that work she and her team did in the past that has taken hours could be cut down by 3/4ths the time it took by using AI.

“Any type of slow, painstaking process within the organization, AI can cut that down,” she said.

email: [email protected]