Investing in AI Companion Apps

Is there more than just flash and novelty, and can they unlock long-term potential?

My partner at F4 Fund, David Kaye, wrote about the future of AI earlier this year in his piece, "How will we touch the robots?"

I want to share my thoughts on investing in AI companion, tutor, mentor, etc. companies. Here’s a list of challenges and opportunities to go beyond the initial “wow” factor, which current products excel at but often fail to sustain.

Challenge: ChatGPT already a mentor and companion

Here's my prediction: The LLMs keep maturing and advancing, probably to the point where an OpenAI product can do many of the things that a consumer AI app would do right now. Eventually, the companionship that LLMs provide will become commoditized to the point where your virtual buddy on your mobile screen is just a flashy interface.

Another challenge: Flashy interface and long term user retention

You could argue that a flashy interface has its purpose. We've had Talking Tom, a hugely successful, entertaining character on mobile devices, and Clumsy Ninja, a unique physics-based simulation where players trained a clumsy ninja to develop skills. Both products captured attention in their time, but since the pandemic, no character-driven interface has achieved lasting appeal. Even Replika, the AI companion app, seems to have peaked in popularity based on recent numbers (source: AppMagic).

So, how can companion apps move beyond the initial “wow” factor that typically fades after a few days? How do you create a product that users continue to enjoy for years?

Opportunity: Where is the multiplayer ingenuity

Many companion apps emphasize personalization, recalling past conversations to create a lifelike and emotionally engaging experience. However, these single-player interactions can only go so far. For true user retention—keeping people engaged for weeks, months, or even years—a new approach is needed.

I believe the key to sustained engagement is multiplayer. AI can serve as a facilitator, enhancing multiplayer experiences by bridging interactions between users. While we may one day see AI replacing real humans in virtual worlds and metaverses, moving beyond scripted NPC roles, the immediate opportunity lies in using AI to connect people and foster shared experiences.

In the metaverse, AI could facilitate complex social structures by organizing ad-hoc live events, building simulated economies, and managing virtual spaces where users can gather, interact, and collaborate. Essentially, AI could serve as the ultimate dungeon master. An AI companion could even take on character roles, making virtual worlds feel more vibrant, dynamic, and populated.

But to fully engage players in competitive or cooperative tasks—whether player versus player or player versus environment—the presence of real human participants remains essential.

Another opportunity: Utility combined with entertainment

Companion apps often rely on entertainment, but adding utility can help them become more valuable in users’ daily lives. We've seen several AI companions that actually do more than just entertain you: Buddy AI is a recent company that has raised millions to develop a product that has a 3D character teaching their users languages.

But it's hard to escape the fact that all of these startups are focused on this promise of glorified chatbots in your mobile phone where the best LLM will easily win.

Who knows—within a few years, many companies in this space might shift their focus from mobile apps to developing applications for advanced platforms like Tesla robots.

Final words

The future of AI companions depends on their ability to remain relevant in their users’ lives. For founders, this is a chance to push boundaries and redefine what it means to create enduring relationships AI beings. For investors, it’s about identifying the pioneers who can turn fleeting novelty into long-term engagement and retention.