AI's impact on product design
By Jason Martin, SoCal Tech Forum Board Member
In May, Andy Oliva returned to the TechConnect Podcast to continue a conversation we started six months ago around digital employees, AI agents, and the changing role of software. Andy is the founder of Lucky Day Labs, a creative AI agency he describes as an AI transformation partner for companies trying to understand strategy, build AI-enabled products, and adapt their processes as the technology moves quickly. He is also a forum leader in OpenAI’s official forums and an ambassador for HeyGen. In this episode, our discussion focused on how AI is reshaping product design, why software companies need to rethink the way users interact with products, and what developers, founders, and teams should understand about building for an AI-native future.
Key Learnings
AI is lowering the barrier to building software, but raising the expectation for quality, usability, and speed.
The future of SaaS may look less like software as a service and more like service as software, delivered through agents, plugins, skills, voice, and messaging.
Prompt boxes, voice interfaces, and text-based interactions are becoming primary product surfaces, not just add-ons.
Agent-first design requires context, access, backend hooks, APIs, MCPs, and thoughtful orchestration.
Companies that want real AI adoption need to tie experimentation to incentives, processes, and visible leadership behavior.
The Changing Purpose of Software
When I opened the conversation with Andy, I wanted to center it on a question that many software teams are now asking directly: what is the purpose of software when users and companies can increasingly build their own tools with AI?
Andy framed this around what he called the “SaaSpocalypse,” a moment of anxiety for established SaaS companies as AI-assisted builders make it easier to create software in days or weeks. His view was not that software is going away. Instead, he argued that the shape of software is changing.
The phrase that stood out was “service as software.” Rather than thinking only in terms of traditional SaaS dashboards, Andy encouraged companies to ask what service they actually provide and how that service can become modular, accessible, and useful wherever the user already works.
For developers, that shift matters. The product is no longer just the web app. It may be a plugin inside Codex, a skill used by Claude Code, a voice agent, a text interaction, or a backend capability exposed to another agent. In Andy’s view, product design now includes deciding how an AI agent discovers, understands, and uses your company’s expertise.
He gave the example of a database company offering a plugin that carries its schema knowledge, best practices, troubleshooting guidance, and architectural expertise directly into the coding environment. That plugin becomes more than a feature. It becomes the way users interact with the product.
Meeting Users Where They Already Are
A major theme in the conversation was the need to stop forcing users into yet another app, dashboard, or workflow.
Andy pointed out that people already have too many platforms to manage. As AI makes it easier for everyone to create new products, that overload could increase. His recommendation was direct: companies should think voice-first, iMessage-first, or RCS-first.
The insurance example made this concrete. Instead of logging into a dashboard, finding the right form, uploading a file, and navigating a workflow, a customer should be able to speak naturally with a voice AI. If proof is needed, the agent could ask for a photo by text, confirm receipt, and handle the backend process.
For developers, this pushes product architecture into a new place. It is not enough to build a clean UI and expose a few APIs. The experience needs to support natural interaction, context switching, file exchange, follow-up, and task completion across channels.
Andy also made an important distinction about dashboards. He does not believe every dashboard disappears. Instead, he sees them becoming more like storefronts. They communicate brand, provide orientation, and support exploration, while much of the actual work happens through voice, text, or agentic workflows.
That framing is useful for engineering teams. The dashboard may still matter, but it may no longer be the primary work surface.
Lower Barrier, Higher Expectations
One of Andy’s strongest observations was that AI lowers the barrier to entry while raising the ceiling of expectation.
He used website builders as an analogy. Tools like Wix and Squarespace made websites far easier to create, but that did not automatically produce great websites everywhere. The technical barrier dropped, but cultural priorities, business incentives, and taste still mattered.
The same pattern is emerging with AI software development. Teams can build faster, but users will expect more. If software is slow, confusing, unattractive, or unintuitive, the old excuse of engineering complexity is less convincing.
Andy was careful not to dismiss the real constraints teams face. But his point was that the margin for poor user experience is shrinking. If AI makes it easier to build and iterate, then products need to feel better, work better, and respond faster.
For professional developers, this is both pressure and opportunity. The value of engineering does not disappear when code generation improves. It shifts toward judgment, orchestration, architecture, product sense, and the ability to decide what should be built in the first place.
As Andy said, design is still about deciding what to show, where to show it, and how to guide attention. AI can help generate interfaces, but it does not automatically know which two or three actions matter most to the user in a specific product context.
Prompt Boxes as Product Interfaces
I shared with Andy that I am seeing more software products move toward a prompt box as the primary mode of interaction. In marketing automation platforms, for example, users can ask the system to audit content, find pricing references, or build customer segments without manually navigating every step.
Andy agreed that this is becoming a core software pattern. He compared it to the iPhone’s early use of skeuomorphic design, where apps resembled familiar real-world objects until users became comfortable with mobile interaction. Once people understood the medium, interface design evolved.
He sees a similar transition happening with AI. Users are becoming more comfortable with prompt boxes much faster than they adapted to earlier interface shifts.
But Andy’s view goes beyond adding a chat widget. He argued that the prompt box should be able to control the interface. If a user asks for last month’s revenue report, the agent should navigate to the right dashboard, apply filters, adjust columns, and respond to follow-up instructions.
For developers, the implementation detail is important. Andy described a model where the agent does not necessarily need full computer-use control with a mouse and screen. Instead, the application can expose backend hooks into different pages, filters, and actions so the agent can operate the product in a controlled, reliable way.
That is a meaningful design shift. The UI is no longer only for humans clicking buttons. It becomes a structured environment that agents can inspect, manipulate, and explain.
Designing for Agents, Not Just Humans
When we shifted into agent-first design, Andy described three requirements: the agent needs eyes, context, and access.
That means the system must provide the agent with the right information, the right permissions, and the right pathways into the product. This could happen through computer use, backend navigation hooks, APIs, or MCPs. The exact design depends on the product, but the principle is the same. Agents cannot be treated as an afterthought.
Andy also talked about continuity across tools. He described wanting an agent embedded in a software dashboard to carry the same kinds of skills, plugins, and memories available in his coding environment. The user experience becomes more seamless when context moves across interfaces.
At the same time, he noted that friction points still create moats. If one assistant knows one part of your history and another assistant knows something else, that fragmentation keeps users inside certain ecosystems. Even when memory export is technically possible, users still need to know it exists, take the time to do it, and trust that it works.
For software companies, this creates a strategic question. Is your product creating useful continuity, or is it creating avoidable friction? The difference matters.
Plugins and Skills as New Moats
One of the more practical takeaways from Andy’s perspective is that plugins and skills may become a new competitive moat.
He used Granola as an example of a product that moved quickly by offering a plugin so users could access their notes where they were already working. In Andy’s view, that speed mattered. Without that integration, a user might decide to rebuild the functionality elsewhere.
This is an important point for product teams. In the past, companies often measured value through logins, dashboards, clicks, and in-app engagement. Andy challenged that thinking. If users are getting value through a plugin, an agent, or a prompt interface, the product may be succeeding even if traditional engagement metrics look different.
He compared this to the Apple Store. Not every person who walks in buys something immediately, but the store communicates the brand, the experience, and the promise of the product. Some of that value is hard to track in a spreadsheet.
For developers and product leaders, this means instrumentation still matters, but it cannot become the only lens. A product’s value may increasingly show up in how well it participates in a broader AI workflow.
Experimentation as an Operating Model
Andy described Lucky Day Labs as a product innovation lab built around AI-native experimentation. The model is simple: run experiments, learn from them, and create products based on what those experiments reveal.
His argument was that this approach may become necessary because the pace of change is too fast for traditional planning cycles. Waiting too long for requirements, documents, or approvals can mean missing the window where a product decision matters.
For companies, that may require rethinking team structures and development processes. Andy was not prescriptive about exactly how each business should operate, but he was clear that teams need room for rapid prototyping, fast iteration, and hands-on experimentation with the latest tools.
This is especially relevant for software developers. Staying current is no longer just about reading announcements. Andy’s view is that real signal comes from building, testing, and seeing how tools behave in practice.
AI Adoption Requires More Than Access
We also discussed why AI adoption fails inside companies. Andy’s view was that buying tools or telling employees to use AI is not enough.
Employees already have constrained bandwidth, existing KPIs, and incentives tied to their current responsibilities. If AI adoption is not connected to those incentives, it often becomes lip service. People may use AI for small workflow optimizations, but they are less likely to rethink the work in a way that meaningfully changes the organization.
Andy also raised the issue of trust. Employees may worry that AI adoption is really about replacing them. Leaders cannot fix that with words alone. They need to show what they value through decisions, promotions, and visible examples.
That point applies directly to engineering organizations. If a company wants AI-native development, it needs to reward the people who embrace it responsibly, improve productivity, and help the team learn.
The Role of Developers in an AI-Native Stack
Toward the end of the episode, Andy pushed back on the idea that AI simply eliminates software jobs. His view was more nuanced.
He distinguished between the skills used in a role and the role itself. Engineering has already moved up the stack before. Developers once worked closer to firmware. Then operating systems, frameworks, and higher-level abstractions changed the nature of the work. AI is another move up the stack.
In Andy’s framing, developers still matter because expectations are rising. If one engineer becomes more productive with AI, a team of five AI-native engineers can become dramatically more productive. The opportunity is not only cost cutting. It is building more, moving faster, and competing at a higher level.
Takeaway and Key Learnings
My biggest takeaway from the conversation with Andy is that AI-native product design is not just about adding a chatbot or using AI to write code faster. It’s a broader shift in how software is packaged, accessed, priced, experienced, and maintained. Developers need to think about products as services that can be used through prompts, voice, messaging, plugins, skills, APIs, MCPs, and agentic workflows. At the same time, companies need to build cultures that reward experimentation and reduce hesitation without ignoring ethics or implementation quality. The teams that adapt best will not be the ones that chase every new tool blindly. They will be the ones that understand where their users already are, expose their product’s value in those environments, and keep raising the quality of what they build.
About the SoCal Tech Forum
The SoCal Tech Forum: Building a Community of Innovators
The SoCal Tech Forum was founded with a vision to create a dynamic and inclusive space for technology enthusiasts, innovators, entrepreneurs, and students. Based in the Inland Empire, a region brimming with untapped potential, the forum serves as a centralized platform where individuals can come together, gain valuable experience presenting their ideas, and learn from seasoned speakers. Our mission is to amplify technology growth and foster entrepreneurship within the Inland Empire and beyond.
With over 1,000 members and growing, the SoCal Tech Forum is the premier hub for tech-minded individuals from across Southern California. Our community thrives on the hunger for knowledge, collaboration, and innovation. Each month, we host engaging meetups that feature diverse topics ranging from blockchain and artificial intelligence to business strategies for startups and beyond. By adapting our format regularly, we ensure our events remain fresh, relevant, and tailored to the interests of our members.
At the SoCal Tech Forum, we value feedback from our community to guide the topics we explore. Whether you're delving into highly technical subjects or engaging with non-technical content, our events promise something valuable for everyone. Together, we aim to elevate education, spark meaningful discussions, and tackle challenges that drive the tech ecosystem forward.
