The a16z Show - Big Ideas 2026: The Agentic Interface

Starting point is 00:00:00 I chatted with the head of IT recently, who told me for the first time in his two-decade-long career, he believed that IT support was fundamentally going to change. If all of us want this software to be doing work for us, ideally it's doing work with at least, if not more competency than a human could. We're no longer designing for humans, but for agents. The new optimization isn't visual hierarchy, but machine legibility. And that will change the way we create and the tools that we used to do it. Every year, we step back and ask a simple question.

Starting point is 00:00:35 What will builders focus on next? Our 26 big ideas bring together the themes our investing teams believe will shape the coming year in tech. This episode is built around three big ideas that, together, explain where AI products are actually heading next. The shift is not just that models are getting smarter. It's that software is changing shape. AI is moving from a tool you consult to a system

Starting point is 00:00:56 that can understand intent and take action. You're going to hear three different perspectives on that transition. What it means for the interface, what it means for how we design software and information, and what it means for how work gets executed inside organizations.

Starting point is 00:01:12 The first big idea is that the prompt box is not the final interface for AI. Mark Andrusco argues that the winning products will feel less like chat and more like proactive teammates. They'll notice what you're doing,

Starting point is 00:01:22 anticipate what you need, and propose actions you can approve. Here's Mark. I'm Mark Andrewsco, a partner on our AI apps investing team. My big idea for 2026 is the death of the prompt box as the primary user interface for AI applications. The next wave of apps will require way less prompted.

Starting point is 00:01:42 They'll observe what you're doing and intervene proactively with actions for you to review. The opportunity we're attacking used to be the $300 to $400 billion of software spend annually in the world. Now what we're excited about is the $13 trillion of labor spend that exists in the U.S. alone. It's made the market opportunity or the TAM for software about, 30 times bigger. If you start from there and then you think about, okay, if all of us want this software to be doing work for us, ideally it's doing work with at least if not more competency than a human could, right? And so I like to think about like, well, what are the best employees do? What are the best human employees do? And I've recently been talking about this graphic that

Starting point is 00:02:23 was floating around on Twitter. It's a pyramid of like the five types of employees and the ones with the most agency and why they're the best. So if you start at the bottom wrong of the pyramid, it's like people who identify a problem and then come to you and ask for help and ask what to do. And that's like the lowest agency employee. But if you go to the S tier, like the most high agency employee you could possibly have, they identify a problem, they do research necessary to diagnose where the problem came from. They look into a number of possible solutions. They implement one of those solutions and then they keep you in the loop or they come to you at the very last minute and say, like, do you approve this solution I found? And that's what I think the future of AI apps will be. And I think that's what everyone wants. That's what we're all working toward it. So I feel pretty confident that. we're almost there. I think LLMs have continued to get better and faster and cheaper, and I think there's a world in which the user behavior will still necessitate a human in the loop at the very

Starting point is 00:03:13 end to sort of approve things, certainly in high-stakes contexts. But I think the models are more than capable of getting to a point where it's suggesting something really smart on your behalf and you basically just have to click accept. As you guys know, I'm pretty obsessed with the notion of an AI-native CRM. And I think this is like a perfect example of what these proactive of applications could look like. So in today's universe, a salesperson might go open their CRM, explore all the open opportunities they have, look at their calendar for that day and try to think about, okay, what are the actions I can take right now to have the greatest impact on my funnel and my ability to close deals? With the CRM of tomorrow, your AI agent or your AI CRM

Starting point is 00:03:50 should be doing all these things on your behalf, in perpetuity, identifying not only like the most obvious opportunities that are in your pipeline, but going through your emails from the last two years and harvesting, you know, this was once a warm lead, and you kind of let it die. Like, maybe we should send them this email to drum them back up into your process, right? So I think there are so many ways in which drafting an email, harvesting your calendar, going through your old call notes, like, the opportunities are just endless. The ordinary user will still want that last mile approval almost 100% of the time. They will want the human part of the human in the loop to be the final decision maker.

Starting point is 00:04:27 And that's great. I think that's like the natural way in which this will evolve. I could imagine a world in which the power user is basically taking a lot of extra effort to train whichever AI app it's using to have as much context about their behavior at how they perform their work as humanly possible. These will utilize larger context windows. These will utilize memory that's been baked into a lot of these LLMs and make it such that the power user can really trust the application to do 99.9% of the work or maybe even 100. And they'll pride themselves on the number of tasks that get done without a human needing to approve them. Mark gives the interface shift, from prompting to execution. The second big idea follows naturally.

Starting point is 00:05:08 If agents are the ones navigating software on our behalf, then we have to start building software to be understood by them. Stephanie Zang calls this machine-ledgible software. In an agent-first world, visual hierarchy matters less and structure matters more. The advantage shifts to products, content, and systems that machines can reliably interpret and operate inside. Here's Stephanie. Hi, my name is Stefan Deng, and I'm an investing partner.

Starting point is 00:05:31 on the A16 Z growth team. My big idea for 2026 is creating for agents, not for humans. Something I'm super excited about for 2026 is that people have to start changing the way they create. And this ranges from creating content to designing applications. People are starting to interface with systems like the web or their applications with agents as an intermediary. And what mattered for human consumption won't matter the same way for agent consumption. When I was in high school, I took journalism. And in journalism, we learned the importance of starting with the five Ws and H in the lead paragraph for news articles, and to start with a hook for features.

Starting point is 00:06:13 Why? For human attention. Maybe a human would miss the deeply relevant, insightful statement buried on page 5, but an agent won't. For years, we've optimized for predictable human behavior. You want to be one of the first search results back from Google. You want to be one of the first items listed on Amazon. And this optimization is not just for the web, but as we designed software too. Apps were designed for human eyes and clicks.

Starting point is 00:06:38 Designers optimized for good UI and intuitive flows. But as agent usage grows, visual design becomes less central to overall comprehension. Before, during incidents, engineers would go into their grafana dashboards and try to piece together what was going on. Now AISREs take in telemetry data. They'll analyze that data, and they'll report back with hypotheses and insights directly into Slack for humans to read. Before, sales teams would have to click through and navigate Salesforce or other CRMs to gather information.

Starting point is 00:07:10 Now, agents will take that data and summarize insights for them. We're no longer designing for humans, but for agents. The new optimization isn't visual hierarchy, but machine legibility. And that will change the way we create and the tools that we used to do it. It is a question we don't know the answer to what agents are looking for. But all we know is that agents do a much better job at, you know, reading all of the texting an article versus maybe a human would just read, you know, the first couple paragraphs. There are a bunch of tools out there that different organizations use to just make sure that they show up when consumers are prompting chat GPT asking for the best corporate card or the best shoes to buy.

Starting point is 00:07:52 And so there's like a bunch of what we call GEO tools out there in the market that people are using. but everybody is asking the question what AI agents want to see. I love this question. When humans may choose to exit the loop entirely, we're already seeing that happen in some cases. Our portfolio company Decagon is answering questions for a lot of their customers already autonomously. But for other cases, security operations or incident resolution,

Starting point is 00:08:22 we typically see a little bit more human in the loop where the AI agent takes first stab at trying to be. to figure out what the issue is, running the analysis and serving to the humans different potential situations. Those tend to be cases of higher liability, more complex analyses that we see humans staying in the loop and will probably stay in the loop for much longer until the models and the technology get to incredibly high accuracy. I don't know if agents will be watching Instagram Reels. It's really interesting. At least on the tech side, It is really important to optimize for that machine legibility piece,

Starting point is 00:09:03 optimize for insight, optimize for relevance, especially, versus in the past it was more about hooking people in, capturing attention, in flashy ways. What we're seeing already is a case of high volume hyper-personalized content, and maybe you don't create one extremely relevant, article, extremely relevant and insightful article, but maybe you're creating extremely high volumes of low quality content, but addressing different things that you may think an agent wants to see, almost like the equivalent of keywords. In the era of agents, where cost of creation of content

Starting point is 00:09:50 kind of goes to zero, and it's really easy to create high volumes of content, that's a potential risk around just high volumes of things to be able to try to capture agent attention. If software becomes machine legible and agents can execute tasks across tools, that the biggest challenge is not cosmetic. It's organizational. That leads to the third big idea. Sarah Wing describes the rise of an agent layer that sits above the traditional system of record and becomes the place where work actually happens. It collapses the distance between intent and execution in changes which software systems control the flow.

Starting point is 00:10:27 Here's Sarah. I'm Sarah Wang, general partner on A16 Z growth, and my big idea for 2026 is that systems of record start to lose their edge. A passive system of record layer stops making sense when agents can independently execute on a signed intent. I expect to see a new dynamic agent layer that actually make sense for employees to replace legacy systems of record. This is a very exciting development on the long road of inserting intelligence into companies. I don't say that systems of record are losing primacy lightly at all. I used to work at a firm that almost exclusively invested in ERPs and other systems of record because of the stickiness of the data gravity.

Starting point is 00:11:07 There was a wave of SAS 2.0 that was well funded and tried and failed to take on the system of record, mostly through a better UI. This is the first time that we've seen a genuine threat to that. And that's because the distance between intent and execution is collapsing. And that's creating not a 20 to 50% better experience for the user. but how you get to that magical TEDx. Let's take the concrete example of ITSM, IT service management. This has traditionally been the domain of powerhouse company service now.

Starting point is 00:11:40 I chatted with the head of IT recently, who told me for the first time in his two-decade-long career, he believed that IT support was fundamentally going to change. It will look completely different in five years. So why is that? If you think about the way that the old systems work, how long it takes do something like request access, to new software in the firm, and you contrast that with the ITSM agents that are arriving.

Starting point is 00:12:04 They plug into your stack, and this type of request becomes nearly instantaneous. Through advancements in LLMs, you can now extract intent, you can classify the request type, you can map it to a known workflow, identify user entities, and the request from the user becomes fulfilled in a way that's efficient and accurate. So we think there's a couple of valuable layers in this new paradigm. Of course, there's the foundation model layer. We believe that stays valuable. But it's really the emerging agent layer that sits as close as possible to the user

Starting point is 00:12:35 and is collecting data on that user, understanding user preferences that we think accrues value in the future. Based on everything that we're seeing in the wild, we believe this is a huge opportunity for new players to come in and win. Why is that? We're in a phase right now where the product is getting better on a weekly, if not daily basis. And you need teams that move fast. If you're going to collapse intent and execution, what bridges that is actually having an accurate, reliable solution for your customer. Otherwise, they're not going to use it. They're not going to trust the agent that you're building.

Starting point is 00:13:07 That's why we're starting to see even agents built on top of classic iconic platforms like Datadog lose to some of the new AISRE companies like a resolve or a traversal. We're extremely excited about this opportunity and 2026 is going to be the year that the dynamic agent layer overtakes the system of record. Taking together, these three big ideas form a single story. First, the interface shifts from chat to action. Second, the design shifts from human first to agent readable. Third, the workflow shifts from systems of record to agent layers that turn intent into outcomes. This is what agenic really means here. AI stops being something you ask and become something that does.

Starting point is 00:13:51 Thanks for listening to this episode of the A16Z podcast. If you like this episode, be sure to like, comment, subscribe, leave us a rating a review and share it with your friends and family. For more episodes, go to YouTube, Apple Podcasts, and Spotify. Follow us on X at A16Z and subscribe to our Substack at A16Z.com. Thanks again for listening, and I'll see you in the next episode. As a reminder, the content here is for informational purposes only. Should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund.

Starting point is 00:14:30 Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see A16Z.com forward slash disclosures.

The a16z Show - Big Ideas 2026: The Agentic Interface

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.