Why we built 20X: the problems we are trying to solve

Why we built 20X

We built 20X because repetitive work caps growth more than demand does. Missed calls, slow first response, enquiries sitting in a queue, no-shows, and a pile of half-connected tools do not get solved by hiring one more person. 20X is a platform where those workflows actually finish, across phone, web chat, WhatsApp, SMS, and email, with your team handling the exceptions.

That is the short version. Here is how we got to it.

What people actually asked us for

Nobody wakes up wanting an AI operating system. What they want is coverage: something that answers when the team cannot, follows up before the enquiry goes cold, and finishes the dull steps that currently live in five tabs.

The way it was described to us was always concrete. Calls go unanswered after six. First response takes hours instead of minutes. Enquiries sit in the CRM until someone feels like working them. There are three internal tools and a web of automations holding them together. Hiring another salesperson just to keep up is not affordable.

That is why the 20X solutions are named after the work rather than the technology: customer support, lead qualification, appointment booking.

What we watched fail

A chatbot is fine when the answer is already in a help article. It is no use when a refund needs checking, a slot needs booking, or an enquiry needs qualifying and passing on. Teams that trained a widget on their FAQs were not wrong to try. The ticket still needed someone with access to a system.

The patched-together stack was worse. A helpdesk bot, a phone menu, a dialler, a booking tool, and automations in between. Each piece worked on its own. Together they produced fragile links, no single record of what happened, engineers babysitting connectors, and an operations team that could not change a single line without raising a ticket.

We spent years wiring exactly that for clients at Setoo, which is why adding one more tool never felt like the answer. What was missing was one place to run the agents, the knowledge, the limits, and the handovers.

Hiring works too. It also adds cost, training, and inconsistency. For high volume work with stable rules, like order status, appointment reminders, and first-pass qualification, another full-time seat is an expensive way to buy eight more hours of cover. Response speed is one of the clearest examples: research on online enquiries has shown for years that waiting minutes rather than hours changes the odds of ever reaching the buyer (Harvard Business Review on the short life of online sales leads).

What would have guaranteed failure

It helps to ask the opposite question. An AI operations product fails if it invents answers when the knowledge is missing, calls itself autonomous with no way to reach a human, ships a chatbot and calls it an agent, prices idle seats, or makes your operations lead wait on an engineer for every change.

We built to avoid that list. Approved knowledge, a defined handover, human approval where it matters, usage-based pricing, and a no-code path are not slide decoration. They are the failure modes we refused to ship. It is also why we will not publish customer numbers we cannot show the source for.

What we built instead

20X is a no-code AI agent platform for business operations, built and run by Setoo. You describe a workflow, connect the systems it needs to touch, add knowledge you have approved, set the limits and the handover, test it, and launch it on the channels your customers actually use. The same agent and the same knowledge carry across those channels, rather than a different bot per channel.

For the category definition in one read, see what is 20X. To put a number on one workflow before you speak to anyone, the free AI business tools cover ROI, voice cost, response speed, no-shows, a short discovery quiz, and a wider diagnostic. No account needed.

We also run real operations rather than only landing pages. The Thrifti case study is the public example: a resale marketplace running listings, fulfilment questions, and customer conversations on 20X.

Who this is not for

If you only need an FAQ box on a website, this is more than you need. If your procurement requires a completed SOC 2 Type II certificate today, we are ready for that audit rather than certified. If the decision must be made by a clinician or a lawyer, it should stay with them. Low volume work where the numbers never pay back should stay manual.

Saying no is part of building the right product.

How to start without a six-month programme

Pick the workflow that already hurts: after-hours calls, slow follow-up, no-shows, or the ticket type your team copies and pastes. Put a number on it with one calculator or the automation opportunity quiz. Then either start free or book a 30-minute strategy call.

You do not need to become technical. Describe the job, approve what the agent may use, and decide when a person should step in.

What we are still working on

We are not done. Phone-line edge cases, moving off an existing setup, and the fear of an agent embarrassing you in front of a customer are all real. Limits, audit logs, access control, and handovers exist because those fears are reasonable.

The underlying problem does not change with the next model release. Too much work still waits for a specific person to be free at the exact moment it arrives. 20X exists so that moment is covered, and so a person is still there when the agent is not enough.

What changed our minds along the way

Three things surprised us once real traffic started arriving.

The knowledge matters more than the model. Teams expect the hard part to be the intelligence. In practice, the hard part is deciding which documents are true, current, and allowed to be quoted to a customer. A tidy, small, approved set beats everything you own.

Handovers decide whether people trust it. The agents customers remember fondly are not the ones that handled everything. They are the ones that recognised the limit quickly and passed the conversation on with the context intact.

The first workflow is never the impressive one. It is after-hours calls, or the first reply to a form, or confirming appointments. Nobody demos that. It is where the payback is.

Why we price by usage

Seat pricing made sense when software was something a person opened. An agent is not a seat. It works when the work arrives, which on most phone queues and enquiry forms is spiky and seasonal.

So the platform is billed on usage, in tokens, where 1 token is ₹3 and a standard voice minute inside a package costs about one token. One balance covers voice, chat, WhatsApp, and email, and top-ups do not expire. A quiet month should cost almost nothing, because that is the honest shape of the work.

Where teams want help rather than software alone, the fees are published rather than quoted case by case: Explore at ₹4,999 for a discovery session, Start at ₹25,000 plus ₹9,999 a month to get a first workflow live with our team, Transform at ₹75,000 plus ₹24,999 a month for a wider build, and Scale from ₹1,50,000 plus ₹49,999 a month for organisation-wide work. Consulting fees pay for people. Tokens pay for the platform. We keep them separate so you always know which is which.

What we mean by an AI workforce

We use the phrase carefully. An AI workforce is not a replacement for your team, and we have never seen it work as one. What it replaces is the queue: the pile of routine work that waits for a person to be free, then arrives all at once and pushes the important work to the evening.

The version that works looks dull from the outside. An agent answers after hours and books the appointment. Another replies to every enquiry in under a minute and asks three questions. A third confirms tomorrow's bookings and reschedules the ones that need moving. Your team arrives to fewer, better conversations.

That is the product we were trying to build, and it is the reason the solution pages are organised around those jobs rather than around features.

Where to start reading

If you want the definition, read what is 20X. If you want a number instead of a definition, the free tools will give you one for your own business in a few minutes. If you would rather talk it through, book a strategy call.

Either way, start with one workflow. We would rather you launch one small agent that works than approve a programme that never ships.