# Perplexity Interface > A calmer way to move from a question to finished work, all in one place. By Antonio Lopez · UF · April 2026 · https://antonio.ac/thesis/perplexity-interface ## The idea in five minutes [Video: Perplexity Interface presentation](https://antonio.ac/work/perplexity-interface/presentation.mp4) Chapters: - The proposal - Where work breaks - Why one workspace - How Interface works - What to measure ## Full deck [Download the deck (PPTX)](https://antonio.ac/work/perplexity-interface/perplexity-interface-deck.pptx) 1. Perplexity Interface 2. Categories open when one product makes a new capability usable 3. Perplexity Pro — before 4. Perplexity Interface — after 5. Presentation agenda 6. No environment runs all four in sequence 7. The fragmentation tax 8. Productivity follows workflow design 9. Perplexity already has the pieces for Interface 10. The integrated workspace 11. The category window 12. The three-month sprint 13. Interface — making work easier ## Keep every step close Interface helps people turn an answer into something ready to share without rebuilding the work somewhere else. ## Full transcript [00:01] Now, my work as a product manager and product designer at IBM, deploying artificial intelligence tools to some of the largest companies, is focused on this exact problem. What you see on the screen right now is a leading AI search platform called Complexity. [00:13] Now, the typical way of doing things for any knowledge worker at a company like JP Morgan, but they would send in their query to the chatbot, and they would get a wall of text, and then maybe, if they're lucky, an artifact or two. [00:25] What I'm here today to talk about is a proposal called Perplexity Interface, a new product that unifies the entire chain of work into one surface, into a single screen, where a full week's worth of work is consolidated into one surface. [00:38] I'll be talking about today why we made some of the design decisions that we made, and some of the problems that surfaced our investigation, what users are looking for, and why this addressed it, correct? [00:49] I'm gonna talk about 3 things in particular. First, I'm gonna talk about the fragmentation tax. I'm gonna talk to you guys about, like, context switching and switching tabs. [00:56] I'm going to talk about why I selected complexity as an existing stack to be the suitor for this product, and then I'm gonna talk about why timing is incredibly important. Now, the choice of how we design a product would define a category and hopefully take the market. [01:10] Now, I said earlier that artificial intelligence is capable of running four of the main primary sources of what we do at work, which is reasoning, we write, we code, and also we draft artifacts and we write emails. [01:22] The issue is that the typical way of doing this is opening about 12 different tabs and switching tabs from one place to another. Here at Newswire estimated this costs almost half a trillion dollars a year in lost productivity. [01:34] So simply by just switching from one task, from one tab to another, completely eliminating all the synergies that we could create from all the different tools we use in a given workday. [01:44] McKinsey also did a really interesting study, where they studied, all the Fortune 500 companies, or a good select amount of them, and they saw that almost all of them were using AI tools regularly. [01:54] Only 40% of them, 4 in 10, were seeing any meaningful gains from it. There's a gigantic gap that makes us question if people want to use these technologies and aren't seeing the returns, there must be some kind of issue in the way we design it and deliver it to face users. [02:08] The good news is that Goldman Sachs did some research, on 4-4 earnings reports and analyzing all their transcripts, and they saw that, [02:17] That 30% of productivity raised when you unify the workspace into one environment. Now, these are typically, like, siloed environments where you can't necessarily see what the work is doing behind the screen, which is why I want to amend the design of how we do things. [02:31] Now, I talked about earlier why Perplexity is the existing suitor, the best prime suitor for this type of product, and that's because they consolidate all of their different product sources into one. But currently, that's not being delivered. That is what I want to talk about today, and that's what I want to present to you all. [02:47] So what you see right now is a download of the product, and usually what you would do is you would send in a query into the chatbot, and then it'll root its source in the resources, like your Bloomberg terminal, if you're working for JP Morgan. [02:59] And then, typically, this is where the story ends for most AI chatbots. You'll get your answer, you'll maybe get, some kind of artifact, and then it ends there. Now, when you're facing, like, a portfolio manager, or a manager that is relying on your analysis and your decision, this is just totally insufficient. [03:15] So what we have devised is a system where it routes all of your data, all the company's specific tools and data sets, into one specific environment. And so you can see that it can create Excel spreadsheets and really, like, your capital IQ, which is a particular database that a lot of portfolio managers and analytical sources use. [03:34] And then you can create a source trace, which is basically an appendix for all of your research, so you can go back and look at it in the future to see if the work you're doing is actually rooted in the grounded truths of the company's intellectual property and datasets. And then below that, you have a Jupyter notebook that's able to run simulations, do work in real time. [03:51] Finally, you have the portfolio manager handoff, or the actual exportable artifact, where you're gonna face whoever, whatever manager, whatever individual gave you that task. [04:00] Now, what's exceptional about this product is that an entire week's worth of work is consolidated onto a single page. Typically, these would be different tabs. [04:09] You can see that there's 12 different apps and tools. This would completely clog your day, if not your week. We've successfully been able to cut out all the noise and centralize a system that is intelligent and actionable on your proprietary information. [04:23] Now, the reason why this is so important to change the design is because these categories typically get defined once, and then once they take the market, they never change. We saw this with the iPhone, and right now, Gemini and Microsoft, or Google and Microsoft, I'm sorry. [04:36] are actually unifying, starting the process of unifying all their distinct capabilities into one environment. So this gives us a very narrow opportunity, about 9 to 12 months, to actually take the market and to outcompete a lot of the leading AI labs that are trying to figure out what users are looking for. [04:52] Which is why I asked for a 3-month sprint. It's about 90 days to get our product to use… to get our product to about 200 to 500 power users. [05:00] We're gonna ask two very specific questions. Are people coming back for a second and third task, and are they expanding the amount of tasks that they're asking a particular, a particular tool to use on a given workday? [05:14] Then we'll be able to quantify exactly how much it's worth, and we'll be able to take the market and hopefully reach millions of people to make work a lot easier. I appreciate you guys taking the time to listen to my presentation, and open to any questions. Thank you.