Skip to content
AppDues
Log in
Contact salesSign up

How Tolan turned a lovable alien into a $1M ARR AI app

Ajay Mehta, founder and president of Portola, built Tolan around a simple bet: an AI product wins when people feel the other side actually knows them. Since a soft launch in late 2024, the app has passed $1M ARR, passed half a million downloads, and held a 4.7 App Store rating. This page breaks down the product, cost, and marketing choices behind that result — and how to judge whether the same playbook fits your own app.

The AI revolution: personalisation people can feel

For years, voice assistants and chatbots promised understanding but often delivered frustration. Mehta's argument is that newer models finally make a different kind of consumer experience practical: one where the product remembers you, reads your tone, and responds as if a real counterpart is present.

Tolan applies that idea to a single character — a colourful alien companion. Instead of a general-purpose assistant, every conversation is personal and emotionally aware. That focus matters for product-market fit: users are not testing a tool, they are returning to a relationship. The founders — Mehta, Quinten Farmer, and Evan Goldschmidt — treated authenticity as the product, not a feature layered on top.

Competition and costs: why an early paywall helped

Natural AI chat is not cheap. Each exchange may need to retrieve earlier conversations, choose an appropriate tone, and generate a reply with low latency. Those steps consume significant computing resources, so token spending rises with every active user.

For Tolan, that pressure arrived quickly. Two weeks into soft launch, operating costs forced a paywall. Mehta credits that early decision with improving both growth and product development. A free user who disappears after 12 days tells you little; a subscriber who renews — or fails to renew — in month two, four, or six tells you exactly where value breaks down.

The cost also shaped competition. Because infrastructure is expensive, fewer teams can sustain a polished, always-on AI companion. That scarcity gave Tolan room to establish itself before direct rivals appeared.

Saving big with TikTok marketing

Tolan's characters do marketing work before any ad runs. Bright colours, expressive animation, and a distinctive alien design make a two-second clip recognisable in a crowded feed. That visual identity kept paid install costs low from the beginning.

The team then leaned into organic TikTok with creators. When a video landed, installs could spike to five figures in a day. Virality is never guaranteed, but the underlying system is repeatable: memorable characters, frequent short posts, varied formats, and creator partnerships that let the personality travel beyond the app itself.

Use this checklist before copying the playbook:

  • Can your product be recognised in a silent, two-second video?
  • Does a character, mascot, or interface element carry the brand without a logo?
  • Can creators show the core experience without explaining it?
  • Do you have a paywall or trial that reveals real retention, not just curiosity?
  • Have you priced in the per-user cost of every AI interaction?
  • Are you posting often enough, in enough formats, to give one post a chance to break out?

If most answers are no, fix the product and cost model first. TikTok amplifies what already works; it rarely rescues a forgettable or unprofitable app.

Frequently asked questions

What is Tolan?

Tolan is an AI companion app from Portola, founded by Ajay Mehta with Quinten Farmer and Evan Goldschmidt. Users talk in real time with a personalized alien character that carries emotional context across conversations.

Why did Tolan add a paywall so early?

Yes. About two weeks after soft launch, high token costs pushed the team to introduce a paywall. Paying users then gave clearer retention signals — for example, whether someone renewed in month two, four, or six — than free users who simply stopped after a few days.

How did TikTok help Tolan grow?

The bright, visual characters made short-form video easy to produce and recognise. Paid installs were relatively inexpensive from the start, and organic TikTok posts sometimes drove 10,000 to 15,000 installs in a single day.

Why are AI apps expensive to run?

Fast, natural AI conversation requires recalling past chats, matching tone, and responding quickly — all of which consume substantial computing resources. That cost also limits how many teams can operate a polished AI app well, creating a natural barrier for competitors.