Multi-language IVR workflows, web apps, and AI automation — built to be handed over.

There is no price list, because the same sentence — “we need an IVR” — covers a two-week call flow and a nine-month platform. Cost tracks four things: how many screens or call flows exist, how many systems have to be integrated, how much of the current data is trustworthy, and whether anything has to be live before a fixed date.

You get a written quote after a scoping call, with the scope, the exclusions, and the acceptance criteria attached to it. If the budget you have in mind will not buy what you have described, we say so on the call rather than in a proposal three weeks later.

Any of them. An IVR is a number, a call flow, a routing decision, and a lookup against your data — the parts that change between industries are the words and the system it reads from.

We have shipped this for healthcare and home care, logistics, financial services, retail, utilities, and public helplines. The healthcare work happens to be the most regulated version, which is why the failure paths get designed first regardless of who is calling.

Yes, and that is usually the right sequence. On a phone system, a speech or LLM layer is invoked at the point where your menu sits today, and the existing keypad tree stays as the fallback path. On a web product, an assistant sits beside the current interface rather than in front of it.

A first phase covers the handful of intents or tasks that account for most of the volume — not the whole surface. Everything else keeps working exactly as it does now, and there is a measured before-and-after rather than a claim.

With. Projects are written to end with your engineers owning the system: infrastructure as code, documentation, runbooks, and review sessions while it is being built rather than a handover document at the end.

If your team will not have the capacity to own it afterwards, that is worth knowing in week one. We raise it during scoping, not after launch.

You do, on both counts. Code lives in your repository from the first commit, and we build in your AWS or Azure account rather than ours, so nothing has to be migrated at the end and nothing is hostage to a renewal.

Where we hold credentials during a project, they are scoped to what the work needs and revoked at handover.

Data stays in your account, in the region you choose, encrypted at rest and in transit, with least-privilege access per component and an audit trail switched on. Nothing is copied to a workstation and nothing reaches a third-party service that has not been agreed in writing.

For regulated work we have built to HIPAA expectations on healthcare telephony — a signed BAA, PHI kept out of logs, redaction before anything reaches a model provider, and a stated retention policy on recordings. The same shape applies to PCI or GDPR obligations: the controls are specific, they are written down, and they are testable.

A 45-minute call — bring an engineer, not only a buyer. We want to see the system, not a summary of it.

Larger builds start with a short paid discovery: current-state review, target architecture, and a sequenced plan with an estimate you can take to finance. You own that document whether or not the build follows. Smaller, well-defined pieces skip straight to a fixed scope and a start date.

Yes, on a monthly retainer covering monitoring, incident response, dependency and security updates, and a block of hours for changes. It is optional and priced separately — some teams take the handover and run it themselves, which is a good outcome.

Production incidents are covered 24/7 for production incidents.

We are based in India and contract with clients in the US, UK, and Europe directly, in your currency and under your law.

Working hours overlap 9:00am–1:00pm ET, every weekday, messages are monitored to 6:00pm ET, and production incidents are covered 24/7 for production incidents.

In practice the distance costs one thing: we do not take 8am ET stand-ups. Design reviews, demos, cutover calls, and incident bridges all happen inside the overlap or on escalation.