Why we switched to packs, not single assistants.

We launched just99 with a clean framing: one specialist AI assistant, one job, ₹99 a month. A Custom GPT tuned tightly for a specific operator task. The logic seemed obvious. The results were not.

The original pitch was wrong.

Version one of just99 sold individual assistants. You subscribed to the Cold Email Operator. You got access to a Custom GPT configured for writing cold emails. That was the whole product. One thing, one price, done.

The framing was clean. The value was not visible. You cannot show the depth of a single Custom GPT in a screenshot. A tuned system prompt and a well-configured assistant look exactly like a blank chat window to someone evaluating whether ₹99 is worth it. That is a hard sell.

The math operators kept asking.

The question we heard most in the first few weeks was some version of: "₹99 for one prompt that I could write myself?" Not always that blunt. Sometimes it came as a support message. Sometimes it was a refund request two days after subscribing.

The operators asking this were not wrong. A single Custom GPT, however well-tuned, is a one-trick tool. Use it once and you have seen what it does. Use it the same way a week later and it still does the same thing. There is no library. There is no variety. The perceived value shrinks fast once the novelty is gone.

We kept hearing this across enough different operator types that it stopped being feedback and started being a design flaw.

What we got wrong about positioning.

We were selling craftsmanship when operators wanted inventory. We were proud of having one carefully tuned assistant per job. Operators wanted a drawer full of starting points they could pull from depending on what that day's work looked like.

The truth about how operators actually use AI tools: they do not open a chat window wanting "the best prompt". They open it wanting something that beats a blank page. The first five words of a cold email are the hardest part. The correct framing of a Facebook ad brief is the hardest part. Once they have a starting point that is 70% right, they are fast. The starting point is the product, not the perfect output.

A single tuned assistant was solving a different problem than the one operators actually had.

The pack model.

₹99 now buys you 20 specialist templates for one job. They sit inside an assistant pre-loaded with all of them. You open the assistant, pick the template that fits today's work, fill in a few placeholders, and send. Tomorrow a different template fits. Next week a different one. The same ₹99 subscription stays useful across the full range of situations that job throws at you.

The Cold Email Operator is not one prompt anymore. It is 20: a founder-to-founder opener, a cold follow-up, a re-engagement after three months of silence, a referral ask, a short punchy version for people who are clearly getting too many emails, a longer one for categories where operators want to show depth. You pick based on what you know about today's recipient. The assistant holds the library.

This is what ₹99 should feel like. Not a one-shot tool you exhaust in a week. A reference you keep open.

Why this is honest.

We do not pretend each template is a hand-tuned masterpiece. They are well-written starting points built by people who have done that specific job. The value is the curation: you did not have to write 20 of these from scratch, stress-test which angle works for which situation, or maintain them as models update.

We update the pack when the underlying models shift in a way that changes what a good starting point looks like. Your subscription link always points at the current version. You do not manage that. We do.

That is the deal: ₹99, 20 curated starting points, kept current, one job. If the pack does not fit your version of that job, cancel. No lock-in, no annual prepay penalty.

What stays the same.

The price is still flat ₹99. Still one pack per subscription. Still cancel anytime from the account page in two clicks. Still India-first, which means rupee pricing and GST invoicing, not dollar prices with a disclaimer.

The switch from single assistants to packs was not a pivot. It was a correction. The job operators needed doing was always "give me a library of starting points for this work". We just took a version to find that out.