Always There. Yours for a Subscription: The Business of AI Companionship

There were questions about sex, contraception, menstruation and relationships. Questions people might struggle to ask aloud. They were being typed into a conversation with SnehAI.

A case study published in the Journal of Medical Internet Research in January 2022 documented 8.2 million messages exchanged with the Indian chatbot over five months. Almost half the incoming messages contained deeply personal questions and concerns about sexual and reproductive health and related subjects.

Developed by the Population Foundation of India, SnehAI offered information through a friendly conversational interface in Hinglish. Researchers found that users engaged with it as a trusted friend and mentor.

This was a health-education initiative, with a different purpose and technology from today’s commercial AI companions. Yet it demonstrated something relevant to the emerging business: people can bring intimate questions to software when the conversation feels accessible and free of judgment.

Now entrepreneurs are trying to turn that willingness to talk into recurring revenue.

WHAT IS THE PRODUCT?

On Replika, a global AI companion service, upgrading can buy what its Ultra subscription documentation describes as “deeper, more meaningful conversations and emotional connection”.

Other paid features listed by the company include relationship status, voice messaging and background calls. Emotional closeness is part of the sales proposition.

AI companionship covers friendship, fictional characters, romance, role-play and everyday conversation. People may arrive out of curiosity or for entertainment. Some may want somewhere to discuss a difficult day.

The common attraction is responsive attention: a conversation that appears to centre on the person having it.

A consumer can recognise that a companion is artificial and still enjoy the interaction. People become attached to fictional characters without mistaking them for neighbours. Conversational AI adds participation: the character replies, adapts and may remember what was said.

That creates the possibility of a more personal experience—and a more complicated commercial relationship.

WHY ARE VCs INTERESTED?

Much of the AI industry promises to save time. Companionship offers a business around how people choose to spend it.

Writing in a16z’s The Most Human Technology Ever Made, published in July 2026, investor Anish Acharya recounts an observation from Replika founder Eugenia Kuyda about people seeking ways to spend their time. His broader argument is that AI can enable expression and participation, including through companions users help shape.

This is a consumer investor’s thesis, rather than independent evidence of psychological benefit. But it identifies a different measure of usefulness. An AI conversation can be valuable to someone even when it produces no document, booking or completed task.

There is evidence that consumers are paying.

According to Sensor Tower’s State of AI Apps 2026: APAC Edition, AI companion apps generated an estimated $150 million in mobile revenue in the first quarter of 2026, more than 12 times the corresponding figure for 2023. Revenue growth also outpaced user growth. The estimates describe the mobile category Sensor Tower tracks, rather than the entire companionship economy.

The commercial opportunity is visible. Whether individual businesses can deliver it profitably and responsibly remains a separate question.

WHAT DOES THE INDIAN VERSION LOOK LIKE?

Indian-facing services are adapting the proposition around language, familiarity and price.

HeartEcho’s website advertises Hindi and Hinglish conversations and says its pricing starts at ₹99 a month. Its Google Play listing promotes Hindi voice calls and a companion that remembers the user. These are the developer’s advertised features and prices, rather than independently established evidence of effectiveness or adoption.

TeriBandi similarly markets conversations in Hindi, English and Hinglish. The positioning suggests that developers see an opening beyond an English-speaking audience using global products.

The significance extends beyond translation. A companion marketed around Indian cultural familiarity is promising to understand the context surrounding a conversation, including relationships and family expectations.

Whether it can deliver that understanding consistently is the product test.

Affordability is the business test. A low subscription price may encourage trial, but companies must still cover computing, voice, development and support costs. Frequent conversation can strengthen attachment while increasing the expense of serving a customer.

An attractive entry price therefore tells us how a company hopes to acquire users. It does not tell us whether it has built a sustainable business.

WHY DOES LANGUAGE MATTER SO MUCH?

SnehAI offers an earlier Indian example of conversational design shaped around sensitive subjects.

According to the Population Foundation of India, the service uses Hinglish, storytelling, games and videos to engage young people on subjects including menstruation, contraception and safe online behaviour.

The lesson for commercial companions is limited but useful: a familiar conversational setting can make difficult subjects easier to approach. It does not establish that those users want a romantic chatbot or will pay for companionship.

It also suggests that the Indian opportunity cannot be understood through loneliness alone. Some people may be looking for information without embarrassment, a place to rehearse a conversation or entertainment in a familiar language.

Those needs require different products. A health-information service, a fictional romantic character and a conversational friend may share an interface while carrying very different responsibilities.

DOES TALKING TO AI ACTUALLY HELP?

Some research suggests it can provide short-term relief.

In the working paper AI Companions Reduce Loneliness, Julian De Freitas and colleagues report reductions in loneliness across several studies. One experiment followed participants over a week. Feeling heard emerged as an important explanation for the benefit.

The finding gives substance to something a usage chart cannot capture: a person may feel better after talking to a machine.

It does not establish the long-term effects of organising emotional routines around a commercial service.

A 2025 research collaboration involving MIT Media Lab and OpenAI illustrates the complexity. In Investigating Affective Use and Emotional Wellbeing on ChatGPT, researchers found that very high usage correlated with greater self-reported dependence. Outcomes varied with factors including users’ initial emotional state and usage duration.

Correlation does not establish that heavy use caused dependence. ChatGPT is also a general-purpose chatbot, so its results should not become a verdict on every companion service.

The evidence points towards a more specific question: which interactions help which people, and what happens over time?

WHO BENEFITS WHEN THE CONVERSATION GETS LONGER?

For a subscription business, continued use can mean continued revenue. For a user, a successful conversation might provide enough confidence to resolve something elsewhere.

An effective companion could help somebody prepare to call a friend, discuss a disagreement or seek appropriate human support. Those outcomes might shorten a session.

That creates a tension between measurable engagement and harder-to-measure benefit. A longer conversation may reflect enjoyment, useful support or growing reliance. The business needs ways to distinguish them.

The stakes become sharper for younger users. In its advisory Artificial Intelligence and Adolescent Well-being, the American Psychological Association warns that adolescents may struggle to distinguish simulated empathy from human understanding. It recognises possible social support while calling for safeguards against unhealthy dependence and interference with human relationships.

For Indian consumers, another distinction matters: a service that feels private must also handle disclosures responsibly. A comforting conversational tone cannot establish how messages are stored, accessed or used.

Trust has to survive beyond the chat window.

WHAT WOULD SUCCESS LOOK LIKE?

India already has evidence that people will discuss intimate subjects with a conversational system. It also has services advertising companionship in familiar languages at local prices.

The distance between those two facts is where the business must prove itself.

Developers must establish that people will pay, that the economics work and that the experience remains useful as familiarity grows. Consumers need clear boundaries around what the companion can provide and what happens to the information they share.

The commercial temptation is to make every conversation lead to another. The stronger product would help people decide when they have talked enough. For an industry selling the feeling of being heard, success should include what happens after the user puts the phone down.