About iVaak AI: An IICL Company, Hyderabad & Seattle | Company
Region & billing currency
Book a demo
About

We built the layer we could not buy

Businesses lose revenue at the point of first contact, in India, in the Gulf, in Southeast Asia, everywhere customers reach for whichever channel is nearest, and every available fix fails in a specific way. The cheap tools turn a human question into a numbered menu. The capable ones demand a platform migration a forty-person business will never sign off on. iVaak exists precisely in that gap: a conversational layer that sits on top of what you already run and refuses to become your system of record.

The five failures we set out to fix

01
Missed and abandoned first contact
A brokerage takes forty enquiries a day and answers none after 7 PM. A hospital drops calls through the OPD peak. A D2C inbox has hundreds of unread “is this in stock” messages. The business pays for demand and then fails to catch it.
02
Rigid automation makes it worse
A flow builder converts a human question into a numbered menu. Customers do not ask questions in the shape of a menu, so the business ends up deflecting satisfaction rather than deflecting work.
03
Every channel is a separate island
The customer calls, gets nowhere, opens WhatsApp, repeats everything, messages the website and repeats it a third time. The business sees three unrelated tickets instead of one person.
04
Capable tools demand a migration
The business already runs on Tally, on a Google Sheet the owner’s brother-in-law maintains, on an HIS bought in 2019. Asking it to migrate is asking it to bet the operation.
05
Handoff destroys context
The human agent receives a ticket ID and a truncated transcript. The customer, having explained the problem to a machine, explains it again to a person, and that is the moment automation makes an enemy.

How we work

The conversation state layer is the product
Channels are surfaces attached to it. Any feature that would compromise the channel-agnostic state model gets rejected regardless of the revenue it appears to unlock.
Ship what removes a constraint
Not what demos well. A connector library is less exciting than a customisable dashboard and roughly ten times more valuable, because it decides how fast a customer can actually go live.
Code-complete is not customer-live
In Indian messaging, regulatory onboarding often takes longer than the engineering. We publish two dates and quote the honest one.
Voice quality outranks feature count
A new channel shipped while Hinglish latency is bad adds surface area to a product people are already unhappy with. Quality work is roadmap work.
Say no to segments we cannot serve well
A failed pilot costs more than a delayed entry: for the customer first, and for us second. Declining well is part of the service.
IICL
Intelligence
India.Com Limited

An IICL company

iVaak AI is built and operated by Intelligence India.Com Limited, IICL. That parentage is not decoration: it is why data protection, consent and Indian regulatory posture are treated as design constraints at the start of a feature rather than a compliance review at the end of one.

It is also why we are willing to say no to a segment. When we tell a hospital group that our DPDP consent suite has not shipped and we would rather they waited, that is the house view, not a sales tactic.

Where we are, honestly
Channels live
5 of 8
Channels built, not yet sendable
Email · SMS · RCS
Data residency
India, US, EU, Gulf
Audit logging & retention
Live
DPDP consent-management suite
In build
SOC 2 Type I → Type II
Not yet
Healthcare general availability
Gated
We publish this because a procurement team should be able to establish fit from a public page. If any line affects your criteria, raise it in the first conversation and we will confirm exactly where we stand.

What we will not build

Naming this is part of the product. Each of these has been considered and rejected for a stated reason, and every one of them would make the promise of a neutral layer untrue.

A CRM, ERP or ticketing system
The whole positioning is that we do not replace your systems. Building one destroys the pitch and the trust that comes with it.
A general-purpose chatbot builder
Different product, different buyer, commoditised, and a race to the bottom on price.
Our own foundation models
Capital-intensive and not the moat. The commodity layer improves faster than any startup could match, so we stay multi-provider.
Marketing automation
Adjacent but a different buyer in a crowded market. We integrate rather than build.
Live video and co-browsing
Low demand in our segments, high complexity, and a distraction from the layer.
Per-minute price matching
A race we would win by destroying the value of what we sell. We compete on continuity, resolution rate and integration speed, and accept the deals we lose on rate alone.

The team behind it

Portrait
placeholder
Founder & product
Owns the conversation state model and the decision about what we refuse to build.
Portrait
placeholder
Voice & ML lead
Hinglish quality, latency on real Indian networks, interruption handling, grounding.
Portrait
placeholder
Platform engineering
Metering, cost attribution, connectors, the 21-pair continuity suite.
Portrait
placeholder
Solutions & customer success
Onboarding, usage modelling, spend caps, and the pilots we scope before anyone signs.
Names, photographs and biographies to be supplied. Roles shown reflect the Year 1 team structure.

If any of this sounds like the way you would want a vendor to behave, talk to us.

Fifteen minutes, one live demo, and a straight answer about whether we are the right fit for your volume and your systems.