Inflection AI launches Pi, a high-EQ personal chatbot
Inflection AI just launched Pi, an AI chatbot built to feel more like a supportive, witty conversational partner than an automated work tool. Founded by DeepMind and LinkedIn veterans, the startup explicitly designed Pi to avoid coding or essay generation, focusing purely on day-to-day dialogue and high emotional intelligence. Why it matters: as the race heats up around workplace productivity bots, Pi is betting that people really want an empathetic sounding board that constantly reminds them it's actually an AI.
While every tech giant is rushing to build AI that writes code and aces exams, two of tech's biggest names are taking the exact opposite route.
Inflection AI, the startup founded by LinkedIn co-founder Reid Hoffman and Google DeepMind co-founder Mustafa Suleyman, just launched its first AI chatbot named Pi.
A chatbot with EQ, not a coding assistant
Just like ChatGPT, Pi uses generative AI to chat with people and answer questions. But that is about where the similarities end.
Built entirely in-house, Pi is designed for personal, day-to-day interactions with a heavy focus on being emotionally intelligent and human-like.
The company is very upfront about what Pi cannot do: it is not designed to generate code, and it will not write your school essays.
According to Suleyman, Pi has been specifically tuned to balance humour and creativity while staying politically correct. The training relied on conversational content and user interactions to sharpen its dialogue flow.
Right now, Pi is free to use across the web, mobile apps, and even social platforms like Instagram, though Inflection AI hinted that paid subscriptions could roll out later.
Behind the scenes, the startup has serious backing. Incubated by venture firm Greylock, Inflection raised 225 million dollars. Hoffman, a partner at Greylock, even stepped down from OpenAI's board back in March to clear any conflict of interest.
Hard boundaries on romance and human imitation
Inflection AI is placing strict guardrails on how Pi behaves emotionally.
For starters, Pi has strict policies against engaging in romantic conversations. The team wants zero emotional entanglement with its users.
More importantly, Pi is explicitly banned from imitating a human. The system will frequently remind users that it is, in fact, an artificial intelligence.
Why it matters
We think Inflection AI is making a fascinating counter-bet against the rest of the industry.
Instead of chasing the enterprise productivity race, they are going after daily personal companion use. For everyday users, Pi is worth testing today if you want an articulate, friendly sounding board that meets you right where you already spend time, like Instagram, without demanding prompts or complex setups.
Their guardrail strategy looks clever. By banning romantic roleplay and constantly breaking the fourth wall to remind users it is machine-made, Inflection is dodging the massive ethical and regulatory traps that usually plague companion chatbots.
Still, there is a real trade-off here. Stripping out coding, long-form writing, and romantic intimacy leaves Pi in a very narrow conversational niche. It will be interesting to see whether users find enough long-term utility in a supportive conversation partner to eventually pay for a premium subscription.
Sources (1)
- businesstoday.in(opens in a new tab)Originalbusinesstoday.in