We build it, improve it, and run it. We don't just advise.
Technology is rarely the business. It becomes the business's problem the day it stops working. The applications people work in, the cloud and networks under them, the security around them: we build those, improve what is already there, and stay to run it.
Vendor-agnostic across the cloud OEMs. We resell none of them. And if the answer is "leave it alone", that's what we'll tell you.
Why Godwit
A bird almost nobody has heard of, doing something no other animal can.
The bar-tailed godwit crosses the Pacific from Alaska to New Zealand without landing, not because it is impressive, but because there is nowhere to land. No island, no ship, no halfway. It cannot fly seventy per cent of the way and call that a crossing.
- 11,000 kmAlaska to New Zealand
- 9–11 dayscontinuously airborne
- No landingno food, no water, no rest
We hold ourselves to the same test. A migration is finished when the business is running on it, not when the last workload copies across. A dashboard is finished when somebody opens it on a Monday. A platform is finished when more than one person can keep it alive. We stay until then.
The part that actually matters
We engage people with deep industry experience: more than two decades of hands-on work across cloud infrastructure, security, delivery operations, web and mobile engineering and IT strategy.
They are the people on the task, and they own the outcome to the end of it.
What we do
Nine practices we run today, in the three stages we sell them in. Follow any one for how the work runs and what it involves.
Talk to us about this
Most AI projects we are asked about do not need AI. A three-week review tells you which of yours does, what data you would need first and what it would cost. Then we build the ones that survive it, with guardrails before anything ships. It sits here rather than above because we would rather be asked for it than lead with it.
A full assessment of your IT landscape and a short, medium and long-term digital transformation roadmap, in the form a CIO or CFO can take to a board. It sits here for the same reason.
These are the practices we run today, not the limit of what we take on. We also stand up infrastructure from nothing, and do the engineering that follows from an architecture or a build-versus-buy decision. If you need something adjacent to this list, ask.
If it is not our work we will say so plainly, instead of taking it on and learning at your expense.
Where we're useful
Six situations that bring people to us. Recognise one and we already know most of the conversation.
Talk to us about thisNobody knows the passwords, the topology or which renewal is next. We document it, stabilise it, then decide together what actually needs rebuilding.
It doubled, nobody can explain which line item did it, and finance wants an answer this quarter. We find the number, then fix the causes.
You need a network, identity, email and backup standing up to a date, and you'd rather not co-ordinate five vendors to get there.
After an acquisition, a founder exit or an MSP you've outgrown. We take it over without the six-week discovery theatre.
Or the deal now depends on ISO 27001, or someone nearly paid a fraudulent invoice. We work out what genuinely has to be fixed, and what the questionnaire will accept.
An approval that waits three days in an inbox, a report typed in twice, a month-end reconciled by hand. We build the small application that takes it off people, then host and run it.
What we commit to in writing
Most of what goes wrong between a client and an IT partner is structural, not technical. So these go in the contract.
Talk to us about thisOur fee is attached to a working migration, a network that is live, an application in production or a service being run, not to advisory time. If an engagement produces a recommendation and nothing else, that is a failure on our side of the table, not a deliverable.
Where there is an internal team, they stay the people who know the estate, and we are scoped around what they do not have the time or the specialism for. We will not quietly accumulate the knowledge that makes them hard to replace, and we will not route their work through us. Where there is no team yet, we are it until there is, and we hand over rather than hold on.
We are vendor-agnostic across the cloud OEMs and carry no quota, rebate or referral fee with any of them. There is no version of a recommendation here that is worth more to us than another, so when we say a workload should stay where it is, that is the whole reason.
Where we supply hardware and licences, the markup is minimal, realistic and shown on the quote as its own line. If you would rather contract with the seller and pay them directly, that works too and nothing is routed through us. Either way the engineering is a separate line you can question.
- Hardware & licences
- minimal markup · shown on the quote
- Engineering
- the only line that is ours
You deal with the people who built it, and a phone number for when it matters. If we ever can't staff that properly, we'll tell you before we take the work.
As-builts, runbooks, Terraform, credentials and monitoring live in your accounts and your repo. Leaving us should be boring. That's deliberate.
If the fix is a config change or a licence review rather than a project, we'll say so. We'd rather have a small piece of work and a client who calls us again.
A short engagement, a fixed fee
Short enough that you're not betting anything on us, and useful even if you never call again.
Each one is a fixed scope with a fixed fee, agreed before it starts. We quote against what the work actually is rather than against a rate card, which is how the number stays honest, and how the smallest of these stays small enough to be worth trying us on.
Talk to us about thisA snapshot of one estate
The smallest step worth taking. Billing export access plus a read-only role is enough for us to come back with what your cloud or network is actually costing you, and where it is exposed.
Cloud assessment & cost optimisation
A line-by-line read of your cloud or hybrid estate. Over-provisioned hardware, governance gaps and idle spend, each costed and ranked by effort and risk.
Network review
Your as-is topology documented properly, often for the first time. Risks, single points of failure, support-contract gaps, and what to do in what order.
Security posture review
Identity, endpoints, cloud configuration and whether you would notice an incident, scored by the risk you actually carry, not by how alarming each item sounds. Includes a DPDP data map where it applies.
An application for one process
The approval that waits in an inbox, the report typed twice, the month-end reconciled by hand. Built, hosted and run, for a nominal cost upfront and a price per transaction after that.
A migration or a build
A defined set of workloads or a defined site, an agreed date, an agreed SLA and a rollback path. We don't start until the scope is small enough to be certain about.
Your money, not oursAn estate spending ₹10 lakh a month, at the 20% we aim for, keeps about ₹24 lakh a year, which is the number the assessment is really about. Ours is a fraction of it, and we would rather talk about the saving than about our invoice.
Ongoing, not a one-offManaged services and network AMC run as a monthly support retainer or a variable model, sized on the estate rather than on a headcount we would like to bill.
Who you'd be dealing with
Godwit AI Labs Pvt. Ltd.
Gurugram, Haryana, India
Where we work
Wherever the work is
Nothing about the offer is regional. Design, configuration, monitoring and out-of-hours support run remotely, so distance changes nothing about them; where a job needs hands on site we travel, and for a rollout we staff the site for its duration.
What does depend on location is the on-site response time, so we agree that per site and write down the number we will actually hit. Ask us about anywhere. You will get an answer, not a postcode check.
How new we are
New company. Not new to this.
We would rather say that plainly than imply a track record we cannot show you yet. Every engagement above is fixed-fee and short for exactly that reason: the risk of trying us should sit with us, not with you.
Tell us what's actually broken
A bill that keeps climbing, a network held together by one person's memory, a migration that stalled.
- A reply within one working day, from someone technical.
- A 30-minute call. No deck, no discovery workshop.
- Then a scoped proposal, or an honest "you don't need us for this".