The Consumer AI Category Nobody Expected
When people talk about artificial intelligence today, the conversation almost always gravitates toward productivity. We discuss copilots, automation, coding assistants, customer service agents, and the possibility of doing more with fewer resources. Every new model release is evaluated through the lens of benchmarks, enterprise adoption, and economic impact. The dominant narrative is clear: AI exists to help us work better.
Yet some of the most revealing experiments in artificial intelligence are taking place far away from enterprise software and productivity workflows. They are happening in spaces traditionally associated with entertainment, creativity, emotion, and imagination.
This became strikingly clear during yesterday’s AI After Work (AIAW) Podcast season 12 finale conversation with Beatrice Bushati, COO and Co-Founder of Pirr, and Rebecca Oskarsson, Co-Founder and CTO. Their company has spent years building an AI-native storytelling platform where users co-create romantic narratives with artificial intelligence. At first glance, the idea sounds unusual. In a market obsessed with enterprise efficiency, why would anyone focus on AI-generated romance stories?
The answer turns out to be surprisingly important.
What Pirr has discovered is not simply that AI can help people write stories. The company has uncovered something far more fundamental about how humans interact with generative AI. While much of the technology industry assumes that better AI means less human involvement, Pirr has found exactly the opposite. Their users do not want AI to replace their creativity. They want AI to amplify it. They do not want a machine that produces a finished story at the press of a button. They want a creative partner that helps them build worlds, develop characters, explore relationships, and shape narratives.
This insight may ultimately prove relevant far beyond storytelling. It challenges some of the assumptions that have guided the AI industry since the arrival of large language models and suggests that the future of AI may not be defined solely by automation, but also by participation.
Building Before the World Understood Generative AI
One of the most fascinating aspects of Pirr’s story is how early the founders arrived at this insight.
Today, almost every startup claims to be AI-native. Venture capital funding flows freely toward companies built on top of foundation models, and every entrepreneur appears to have a generative AI strategy. Pirr, however, began long before the world understood what generative AI would become.
The company traces its origins back to 2020. At that point, language models were still largely confined to research communities. Most people had never heard of GPT models. There were no viral demonstrations, no mainstream adoption, and certainly no ChatGPT moment.
What existed instead were early models that occasionally produced flashes of brilliance surrounded by large amounts of nonsense.
Rebecca described those early days as moments where the models would generate a few surprisingly coherent sentences before quickly descending into gibberish. Yet even within those limitations, the team could see something extraordinary. The models were not merely predicting text. They were participating in storytelling.
For the founders, that was enough.
The original vision emerged from a simple observation. If AI could generate even a few compelling sentences, then perhaps one day it could help people create entire stories. Rather than viewing generative AI as a technology for productivity, they viewed it as a technology for creativity.
That perspective was far from obvious at the time.
Investors struggled to understand what they were building. Users had little understanding of what artificial intelligence could do. The technology itself was still immature, but the founders believed that if they could create an engaging experience, the models would eventually catch up.
Looking back, that conviction appears remarkably prescient.
The Discovery That Changed Everything
As generative AI improved and Pirr’s user base began to grow, the company discovered something unexpected.
Most observers assume that users want AI to do more of the work. This assumption drives much of today’s AI development. Every generation of models aims to automate larger portions of human effort. The underlying belief is that the ideal user experience is one in which the machine handles everything.
Pirr’s users behaved differently.
Instead of asking the AI to write entire stories, users actively participated in the creative process. They wanted to control character development. They wanted to guide emotional arcs. They wanted to determine what happened next. They wanted to shape relationships and influence outcomes.
The AI was not replacing the author. The AI was collaborating with the author.
This distinction became one of the defining principles of the platform. While many storytelling applications focused on generating large blocks of text from simple prompts, Pirr continued to emphasize human involvement. Users could steer the narrative, make decisions, select directions, and gradually construct stories alongside the AI.
What emerged was something that resembled creative partnership more than content generation.
This insight deserves far more attention than it currently receives because it challenges a widespread assumption about artificial intelligence. We often talk about AI as if its ultimate purpose is to eliminate human effort. Yet creativity does not always work that way. In many creative domains, the process itself is part of the value.
People do not write stories solely because they want a finished story. They write because they enjoy creating.
The same principle applies to music, art, design, and many other forms of creative expression. Removing the creator from the process may increase efficiency, but it can also remove much of the enjoyment.
Pirr’s success suggests that the future of generative AI may involve collaboration as much as automation.
Why Romance Became the Perfect Testing Ground
Another fascinating lesson from the discussion is why romance proved to be such a powerful category for AI storytelling.
Many people outside the publishing industry underestimate the size and sophistication of the romance market. Romance is one of the largest genres in publishing, attracting highly engaged readers who often consume multiple books each month. More importantly, romance readers tend to know exactly what they want.
This became a significant advantage for Pirr.
The two of the founders explained how romance readers organize stories around specific tropes. These tropes function almost like narrative building blocks. Readers search for enemies-to-lovers stories, friends-to-lovers stories, grumpy-versus-sunshine dynamics, forced proximity scenarios, and countless other combinations.
To outsiders, these preferences may appear repetitive. To readers, they provide emotional frameworks that define the experience they are seeking.
From an AI perspective, this is incredibly valuable.
Users are effectively providing structured descriptions of emotional outcomes. They are communicating not only what they want to read, but how they want to feel while reading it. This creates a rich feedback loop that allows AI systems to learn far more about user preferences than traditional content platforms.
Pirr observed that their users provide dramatically more input than someone creating a story through a general-purpose chatbot. Every decision, character adjustment, plot direction, and narrative preference becomes a signal. Over time, these signals create a detailed picture of how people engage with stories.
The result is not simply personalization. It is collaborative storytelling driven by an unusually deep understanding of user intent.
From Erotica to Emotional Intelligence
One of the most interesting strategic decisions in Pirr’s journey was its evolution from erotica toward romance.
At first glance, the distinction may appear minor. In practice, it changed almost everything.
The founders explained that erotica focuses primarily on specific scenes and moments. Romance, by contrast, focuses on the emotional journey surrounding those moments. The relationship itself becomes the story. Character development, emotional tension, vulnerability, conflict, growth, and resolution all become central components of the narrative.
This shift forced Pirr to solve a much more difficult problem.
Generating isolated scenes is relatively straightforward for modern language models. Maintaining emotional consistency across an extended narrative is significantly harder. Characters must remain believable. Motivations must evolve logically. Relationships must develop naturally. The system must remember what happened previously and build upon it in ways that feel authentic.
In many ways, Pirr’s evolution mirrors the broader evolution of generative AI itself.
The first generation of AI systems focused on producing outputs. The next generation increasingly focuses on maintaining context, memory, and coherence over time. The challenge is no longer generating text. The challenge is sustaining meaningful experiences.
What Pirr discovered is that users care deeply about emotional continuity. Readers become invested in characters. They remember previous interactions. They expect consistency. The storytelling experience succeeds or fails based on whether the emotional journey feels believable.
This pushes AI beyond language generation and toward something closer to emotional modelling.
Storytelling Versus Artificial Companionship
As the conversation turned toward emotional AI, another important distinction emerged.
The rapid growth of AI companions, AI girlfriends, and AI boyfriends has sparked widespread debate about the role of artificial relationships. Some see these products as harmless entertainment. Others worry about their impact on human connection.
Pirr occupies an interesting position adjacent to this category while deliberately avoiding it.
The founders made a compelling argument that storytelling serves a fundamentally different purpose than artificial companionship. While both involve emotional engagement, storytelling encourages creativity, imagination, and participation. Readers interact with fictional characters, but they do so within a narrative framework that remains clearly distinct from reality.
The goal is not to replace human relationships. The goal is to explore stories.
This distinction may become increasingly important as emotional AI continues to evolve. Society will likely spend the coming decade debating where to draw boundaries between healthy engagement and unhealthy dependence. Companies operating in this space will need to think carefully about how their products influence user behaviour.
Pirr’s approach suggests one possible path forward. Rather than positioning AI as a substitute for human connection, the platform uses AI to facilitate creative expression. The technology becomes a tool for storytelling rather than a replacement for relationships.
That difference may prove more significant than it initially appears.
What Consumer AI Can Teach Enterprise AI
Although Pirr operates within entertainment and storytelling, many of its lessons apply directly to enterprise AI.
One of the most striking parallels involves user agency.
Enterprise AI discussions frequently revolve around automation. Organizations want systems that reduce workloads, eliminate repetitive tasks, and increase efficiency. Those objectives remain important. Yet consumer applications often reveal deeper truths about human behaviour.
Pirr discovered that users value control. They value participation. They value the ability to shape outcomes.
These same principles increasingly appear in enterprise environments. Employees often resist systems that remove autonomy while embracing systems that enhance their capabilities. The most successful AI products frequently feel like collaborators rather than replacements.
This suggests that future AI design may require a more nuanced understanding of human motivation. Efficiency matters, but so does engagement. Productivity matters, but so does ownership.
The organizations that understand both dimensions will likely build more effective products.
The Future of Emotional AI
Perhaps the most thought-provoking part of the discussion centered on a simple question: does AI actually need to understand emotions?
Rebecca argued that today’s systems do not truly understand in the human sense of the word. They identify patterns, recognize signals, and generate responses based on statistical relationships. Yet from the user’s perspective, the distinction may not always matter.
Beatrice referenced the famous ELIZA experiments, where users formed surprisingly strong emotional connections with a system that simply reflected their own statements back to them. Decades later, modern language models are vastly more sophisticated, but the underlying lesson remains relevant.
Humans are remarkably willing to attribute emotional understanding to machines.
This reality creates both opportunities and responsibilities.
On one hand, emotionally aware AI systems may help people feel heard, understood, and supported. On the other hand, developers must remain mindful of how these systems influence human behaviour.
The future of emotional AI will likely depend less on whether machines genuinely experience emotions and more on how humans experience interactions with those machines.
That may be one of the defining technology questions of the next decade.
Beyond Productivity
The AI industry often frames progress in terms of efficiency. We measure time saved, costs reduced, and tasks automated. These metrics are important, particularly for enterprises seeking measurable returns on investment.
Yet conversations like this one remind us that technology has always shaped more than work.
The internet transformed communication.
Social media transformed identity.
Streaming transformed entertainment.
Artificial intelligence may ultimately transform creativity.
Pirr’s journey offers an early glimpse into what that future might look like. The company started with a simple idea that many people initially dismissed. Today, it stands at the intersection of storytelling, personalization, emotional engagement, and human creativity. Along the way, the founders have discovered that people do not necessarily want AI to create for them. They want AI to create with them.
That distinction may prove to be one of the most important lessons emerging from the current AI wave.
As models become increasingly powerful, the question will not simply be what AI can do on its own. The more interesting question may be what humans and AI can build together.
The answer may not emerge first from boardrooms, enterprise software, or productivity tools. It may emerge from something far more human: our desire to tell stories, explore emotions, imagine possibilities, and connect with one another through narrative.
If that turns out to be true, then AI love stories may tell us far more about the future of artificial intelligence than many of today’s AI agents ever will.
*This article was enhanced with the help of AI tools, drawing on the podcast transcript and complementary online research. To go deeper into the source material, I encourage you to listen to the full episode and make your own learnings. www.aiawpodcast.com
