Inside the Next Phase of AI: Agents and On-Device Intelligence
Ask about Knowledge at Wharton
Episode Description
ABOUT THE EPISODE Stefano Puntoni said AI in 2026 will be defined by specialization, not general intelligence, as models are engineered for speed, reasoning, or cost efficiency. He noted that agent systems will replace single-model workflows, allowing AI to perform tasks while other AI evaluates quality and accuracy. Puntoni said everyday AI will expand through on-device deployment, embedding intelligence into phones and consumer electronics without constant cloud dependence. He highlighted that regulation and geopolitics will increasingly shape how AI is developed and deployed across markets.
ABOUT THE PODCAST This Week in Business features interviews with Wharton faculty about the latest news, fascinating trends, and issues impacting both consumers and the business world. Episodes are recorded at the Wharton School and published twice per week.
#AI2026 #AIAgents #AIRegulation #OnDeviceAI #AISystems #EnterpriseAI ----- The Wharton School of the University of Pennsylvania is committed to sharing its intellectual capital through Knowledge at Wharton, the school’s online business analysis journal.
Launched in 1999, Knowledge at Wharton offers free access to:
- Articles, podcasts, and videos highlighting Wharton faculty research and analysis of current business trends - Interviews with book authors from Wharton School Press - In-depth series of curated content like Wharton Executive Education’s Nano Tools for Leaders - A searchable database of more than 10,000 articles covering all aspects of business - A weekly newsletter that delivers Knowledge at Wharton insights directly to your inbox
Visit the Knowledge at Wharton homepage: https://whr.tn/3qUpSBE Subscribe to the Knowledge at Wharton YouTube channel: https://whr.tn/3DEAgAU
Read the full transcriptHide the full transcript
- Hi everybody, and welcome to our first conversation of the new year about artificial intelligence. I'm Dan Loney.
- Great to be joined once again by Stefano Puntoni, Marketing Professor here at the Wharton School, and also co-director of
- the Wharton AI Human Research Program. Stefano, great to talk to you again. How are you, sir?
- I'm great and happy New Year to everybody. All right, so as the calendar turns to 2026— and certainly 2025 had a lot of
- conversation around around artificial intelligence. What is it that really has your interest piqued as maybe some things to
- look out for as we go into the new year? I was thinking about what will be exciting to think about for 2026.
- These are— you know, there's lots of stuff going on. Many, many trends, many things to watch out for, but maybe I'll
- share quickly a list of six things that I think is exciting to look forward to in 2026, and that are going to shape a lot of
- processes and conversations around AI. So let me go through, quickly, that list. I think the first one will be really on the
- engineering side. We're going to see an increasing specialization of models. We're going to have faster models, more specialized
- models. If you take, for example, ChatGPT already today is not one model, but it's more like a switchboard that directs
- you to the model that is best suited to answering a particular query that a user has. And we're going to see that kind of
- specialization and flexibility trade off between, you know, thinking, reasoning models or test time compute kind of— you
- know, emphasis— versus speed and quick and cheaper models. And so for companies that are going to deploy language models across lots of
- different functions to do lots of different jobs, the procurement process is going to shift a little bit away from
- which vendor is the best, towards more this idea of optimizing a portfolio of models. I think that's one trend. The
- second trend will be towards agentic systems. And so we've been talking about AI agents for a year now already. So it's not
- a new thing. But here, I emphasize the word system. So what we might see more and more is where AI agents are not going
- to be just one agent performing a multi-step operation. What you're going to have, you're going to have a set of models
- which are going to be working in concert to produce this work output across a set of steps. And you're going to have models
- doing the work, and then you're going to have other models checking the quality of the work done by the first model. So
- you'll have more of a system- like agentic AI. That's the second. The third one is something more for consumer
- markets, where we're going to see this everyday AI, trend continuing. We're going to see more on-device AI. So devices
- like consumer electronics, such as, you know, say smartphones, that are going to be optimized for AI models. So, for example,
- just announced by Samsung, the news is that they are planning to ship about 800 million devices in 2026 that are AI-
- first. So this on-device trend and everyday AI is one. Then, to me, it's interesting from the AI labs and the platforms' point of
- view, how are we going to monetize the free products? And we talked shortly about this in a previous segment. But
- basically, these AI companies say OpenAI are offering this free product, which is serving hundreds of millions of
- customers, and they are not, you know, paying anything for it. And so right now it's a massive cash burn that will end. And so
- somehow these companies will need to find a way to make money out of the free product, and that basically means
- advertising. So we're going to see increasing efforts to monetize the free offerings.
- The fifth and sixth trends are more macro. The fifth is regulation. On January 1, in
- California and Texas, new laws on AI came into force. And we can expect more state level and maybe government level
- across the world actions around AI. Some of those conversations are going to be shaped also by geopolitical forces and
- competition. And finally, of course, is the one everybody talks about. It's about investments and valuations and share prices.
- Are we in a bubble? Are we not in a bubble? And what kind of bubble is it? We recently wrote a short article in Knowledge at
- Wharton discussing this issue. And everybody is excited to see what is going to happen in 2026. This is a very short overview of
- a few things I think are interesting to watch out for. So let me ask you this then. From your perspective as a
- professor, and even to a degree, the Wharton School as well, which I guess for the purposes of this discussion, we could
- call the Wharton School a business. How are— how is the Wharton School and your work being impacted by AI? And how do
- you expect it to continue to develop in 2026 because of AI? Oh, this academic year, I think, is a very important academic year for
- the Wharton School when it comes to AI. Because in September, we launched our AI major and AI concentration for the MBA and
- undergraduate programs, respectively, being the first year school to try that. So we are in the middle of that
- experiment, and we are seeing how this is panning out. I am actually very excited about my course. I'm starting to teach in
- about two weeks. My course is called AI In Our Life, and it's a quarter-length course for MBAs and undergraduates. And I can
- tell you, this year, I'm adding three new sections, three new topics to the course. The first one is a session on creativity
- and ideation using language models. The second one is a session on skills, and also deskilling. What kind of skills do
- we need to develop to be effective working with AI, but also what kind of skills we should be careful about not
- losing when we work with AI? And the third one is about AI agents and basically delegation of work and integration of AI in
- workflows. These are basically the three topics I'm adding. And I'm excited to meet lots of bright students this quarter.
- Yeah. What is the students' reaction when you're doing classes like this? And they obviously understand the
- importance of AI and what it means for not only their education, but for the businesses that they work with,
- especially the MBA students. There's a huge thirst for insight into AI. All the students recognize that AI is going to be something that's
- going to shape a workplace to a very significant extent. And also they realize that if they can be competent and, you know,
- insightful in the way that they understand how AI can impact business practices, they will have an edge in the job market.
- And so— because all companies are excited to to hire talent like that. So I think that's where we see— we are curious to see what
- kind of interest the major is going to get. We don't know, basically, how many students are going to take the major. We'll
- see when they accumulate credits and they try to credential in that. But judging from enrollment in my course, there's
- certainly quite a lot of interest in the topic. Stefano, always great to talk with you, and again, we look
- forward to chatting with you every month, talking about AI this year. Thanks. All the best.
- Thanks, Dan, great to be here. You got it. Stefano Puntoni, Marketing Professor
- here at the Wharton School, and co-director of Wharton AI Human Research Program.
Brief Summary
Stefano Puntoni discusses AI trends for 2026, including model specialization, everyday AI, monetization, and the Wharton School's new AI major.
Key points
- Trends to Watch in AI for 2026. Stefano outlines six key trends shaping AI, from specialization of models to regulation.
- Wharton School Launches AI Major. The Wharton School introduces an AI major for MBA and undergraduate programs, marking a significant educational shift.
Episode highlights
- Trends to Watch in AI for 2026Stefano outlines six key trends shaping AI, from specialization of models to regulation.“These are exciting things to look forward to in 2026.”0:56Jan 7, 2026
- Wharton School Launches AI MajorThe Wharton School introduces an AI major for MBA and undergraduate programs, marking a significant educational shift.“This academic year is very important for the Wharton School when it comes to AI.”5:25Jan 7, 2026





