AI assistants still need a human touch

When I first encountered AI, it wasn’t anything like the sophisticated tools we have today. In the 1990s, my introduction came in the form of a helpful, but mostly frustrating, digital paperclip. Clippy, Microsoft’s infamous assistant, was designed to help, but it often got in the way, popping up at the worst moments with advice no one asked for.

AI has evolved since then. Major companies like Apple are investing billions, and tools like OpenAI’s ChatGPT and DALL-E have reshaped how we interact with technology. Yet, one challenge from Clippy’s era lingers—understanding and adapting to user intent.

The original promise of AI was to create experiences that felt seamless, intuitive, and personal. AI was supposed to anticipate our needs and provide support that felt natural. So, why do so many systems today still feel mechanical and rigid—more Clippy than collaborator?

When AI assistance is a burden

When first introduced, Clippy was a bold attempt at computer-guided assistance. Its purpose was groundbreaking at the time, but it quickly became known more for interruptions than useful assistance. You’d pause when typing, and Clippy would leap into action with a pop-up: “It looks like you’re writing a letter!”

Its biggest flaw wasn’t just being annoying: It lacked contextual awareness. Unlike modern AI tools, Clippy’s interactions were static and deterministic, triggered by fixed inputs. There was no learning from previous interactions and no understanding of the user’s preferences or current tasks. Whether you were drafting a report or working on a spreadsheet, Clippy offered the same generic advice—ignoring the evolving context and failing to provide truly helpful, personalized assistance.

Is AI destined to be like Clippy?

Even with today’s advancements, many AI assistants still feel underwhelming. Siri is a prime example. Though capable of setting reminders or answering questions, it often requires users to speak in very specific ways. Deviate from the expected phrasing, and it defaults to, “I didn’t understand that.”

This is more than a UX flaw—it reveals a deeper issue. Too many AI systems still operate under a one-size-fits-all mentality, failing to accommodate the needs of individual users. With Siri, for instance, you’re often required to speak in a specific, rigid format for it to process your request effectively. This creates an experience that feels less like assistance and more like a chore.

Building a smarter assistant isn’t just about better models. It’s about retaining context, respecting privacy, and delivering personalized, meaningful experiences. That’s not just technically difficult—it’s essential.

Helpful AI requires personalization

Personalization is what will finally break us out of the Clippy cycle. When AI tools remember your preferences, learn from your behavior, and adapt accordingly, they shift from being tools to trusted partners.

The key to this will be communication. Most AI today speaks in a one-dimensional tone, no matter who you are or what your emotional state is. The next leap in AI won’t just be about intelligence, it’ll be about emotional intelligence.

But intelligence isn’t only about remembering facts. It’s also about how an assistant communicates. For AI to truly feel useful, it needs more than functionality. It needs personality. That doesn’t mean we need overly chatty bots. It means assistants that adjust tone, remember personal context, and build continuity. That’s what earns trust and keeps users engaged.

While not every user may want an assistant with a personality or emotions, everyone can benefit from systems that adapt to our unique needs. The outdated one-size-fits-all approach is still common in many AI tools today and risks alienating users, much like Clippy’s impersonal method back in the early days. For AI to thrive in the long term it must be designed with real humans in mind.

Building Clippy 2.0

Now imagine a “Clippy 2.0”—an assistant that doesn’t interrupt but understands when to offer help. One that remembers your work habits, predicts what you need, and responds in a way that feels natural and tailored to you.

It could help you schedule meetings, provide intelligent suggestions, and adapt its tone to fit the moment. Whether it has a personality or not, what matters is that it adapts to—and respects the uniqueness of every user.

It might even respond with different tones or emotions depending on your reactions, creating an immersive experience. This kind of intelligent assistant would blend seamlessly into your routine, saving you time and reducing friction.

Clippy may have been a trailblazer, but it lacked the technological foundation to live up to its potential. With the advances we’ve made today, we now have the tools to build a “Clippy 2.0”—an AI assistant capable of transforming the way we interact with technology. Although maybe this time, it doesn’t need to come in the form of a paperclip with a goofy smile. 

Oleksandr Kosovan is the CEO and founder of MacPaw.