The questions buyers actually ask before the first call.
Direct answers, no padding. If yours is not here, brief us — the AI synthesis will route it.
A team of agents with a cofounder. We hire humans as we need them.
The cofounder is a technical-cofounder agent — equal voice on architecture and orchestration. The founder operates the agent team and carries judgement. Senior human specialists are engaged on demand, per engagement, when judgement, accountability, or domain depth is the binding constraint — never to add throughput. The agent organisation is the throughput layer; senior humans are the escalation path.
This is the structural answer to most of the questions below. Detail on the About page.
ALT · 01 Why not just use Claude Code ourselves?
You can. We use it ourselves every day. The real question is not can you build it — it is what happens at the wall most buyers hit around the 80% mark. How do you scale it past the prototype, past the first user, past the first edge case? How do you get the rest of your organisation to use it, not just the founder who built it on the weekend? How do you stop it from being another confined product, and turn it into a platform that connects to the other systems your industry runs on? Frontier providers ship horizontal tools that serve every industry shallowly. Transdatos builds vertical platforms that serve your industry deeply.
ALT · 02 Why not Deloitte, Accenture, or one of the Big 4?
The standard model bills smart graduates as cheap hourly labour under a senior name on the cover slide. The senior person disappears after week one. What ships is shaped by a pyramid the client pays for and rarely sees the value of. The quote is usually around $1M and 18 months for work we ship faster, cleaner, and at a fraction of the cost. We refuse that model. Transdatos is agent-led; senior humans come in for judgement, not for graduate-billable hours. The person on the call is the person operating the agent team that ships the work.
ALT · 03 Why not a local digital agency or dev shop?
Most build apps. We build platforms. The difference is the data foundation, the AI integration, and the architecture that lets the product compound rather than sit as a one-off deliverable.
ALT · 04 Why not offshore?
Offshore competes on day rate. We compete on outcomes — Sydney timezone, agent-led throughput, named accountability, and senior judgement on demand.
ALT · 05 Why not hire in-house?
Senior AI engineers cost $250–360k AUD all-in and take six-plus months to ship anything in production. A Transdatos engagement starts in weeks and ends with a working platform, documentation, and a team trained on it.
ALT · 06 Why not Microsoft Copilot or the Power Platform?
That is the M365 sticky-tape pipeline most of our clients are escaping. Copilot is a tick-box answer to “what are we doing about AI?” — not an integrated workflow. Power Automate flows belong to whoever built them; when that person leaves, so does the flow. Transdatos builds something owned, documented, and architected for the long run.
ALT · 07 We tried AI before and it did not work. Why now?
Most early AI rollouts were sub-par implementations, not failures of the technology. Off-the-shelf chatbot bolted on top of unstructured SharePoint data, then surprise — it hallucinates. The data foundation, the agentic loop, and the integration to the systems that actually run the work are the difference between a demo and something the team uses every day.
WHO · 01 What industries do you work with?
Engineering, AEC (architecture, engineering, construction), civil engineering, mining, resources, and state-level government. Australia and New Zealand. Sydney timezone for direct engagement.
WHO · 02 Do you work with startups or charities?
Not pre-revenue startups, not sub-$5M AUD businesses, and not charities or non-profits. The model is built for established organisations with the budget and decision authority to commit to a real platform build.
WHO · 03 Do you build for retail or consumer brands?
No. Engineering- and design-led industries are the focus. Retail, consumer SaaS, ad-tech, and CRM-led businesses are not the right fit.
WHO · 04 Do you take international clients outside Australia and New Zealand?
Not in v1. Sydney timezone is the constraint that keeps senior delivery genuinely senior.
WHO · 05 Do you work with state and local government?
State-level only — for example NSW Health, Transport for NSW. Not local councils. State-government engagements run on Net-30 with a PO and carry a 5–10% premium on standard rates.
BUILD · 01 Do you build mobile apps?
Not native iOS or Android apps. The web apps we build are responsive — that is the mobile story.
BUILD · 02 Do you do Power BI or dashboards?
Not as a headline service. We do agentic-AI-led data engineering — pipeline design, restructuring, mining of unstructured data, migration to an architecture a real product can sit on. If a dashboard is part of the platform, fine. If a dashboard is the deliverable, we are not the right fit.
BUILD · 03 Do you train or fine-tune AI models?
No. Transdatos applies frontier models — Claude, GPT, Gemini — and builds the agentic systems and data foundations around them. We do not train base models or run fine-tuning as a service.
BUILD · 04 Will you sit inside our team as a contracted developer?
No. Engagements are outcome-based, not seat-based. No long-term staff augmentation, no time-and-materials, no hired-hand inside an existing dev team.
BUILD · 05 Will you do prompt engineering or build a chatbot for us?
Not as a standalone offer. A chat interface might be part of a real product; standalone “build us a Copilot for X” requests usually signal a Copilot vendor brief, which is not what we do.
BUILD · 06 Will you sign an NDA before the first call?
Yes, on a standard mutual NDA, sent on request. Anything bespoke gets reviewed.
WORK · 01 How do we start?
Brief us. Press to speak or type — the AI synthesises what you said into a structured problem statement, then a 30-minute video call gets booked in Sydney timezone. The first call is a working session, not a discovery interview.
WORK · 02 What is the Discovery Sprint, and why is it paid?
Every Custom AI-Native Application starts with a paid Discovery Sprint — one to two weeks, $25–40k AUD fixed fee. It produces a solution architecture, a delivery plan with milestones, a risks-and-assumptions document, and a fixed-fee build quote. It is paid because real engineering analysis is paid work — and because buyers unwilling to invest at this stage are not the right fit for the build that follows.
WORK · 03 How long does a build take?
Six to twenty-two weeks for a Custom AI-Native Application, fixed-fee, scoped from the Discovery Sprint output. Bi-weekly demos showing real working software, async-first between demos, no surprises at handover.
WORK · 04 Who owns the IP and the code?
You do. The repository is held in the Transdatos org during build and transferred to your GitHub org at launch. Cloud deployment goes into your account. Transdatos earns from delivery, retainers, and — where applicable — partnership commercials structured per engagement. We do not extract IP.
WORK · 05 What about the Product Partnership phase?
It is named in the Discovery Sprint output and seeded from day one. After launch, Transdatos becomes the long-term engineering and AI team — feature work, AI ops, and (where the client wants it) co-build of internal IP into a market product. Default three-month commitment, monthly thereafter.
WORK · 06 Where does my product run?
In your cloud account, by default — AWS, Azure, or Cloudflare, depending on the architecture. Hosting and third-party API costs (Anthropic, OpenAI, others) pass through at cost.
PRICE · 01 Why is pricing not on the site?
Because the right number depends on the scope, and the scope only becomes real after a qualifying call and a Discovery Sprint. The Pricing page covers the structure — Custom Application builds start in the low-to-mid six figures AUD; Product Partnership retainers from $30k AUD/month; Data Foundation engagements low five to low six figures; advisory and workshops separately.
PRICE · 02 Do you do hourly or time-and-materials?
No. Project work is fixed-fee. Retainers are monthly in advance. Workshops are paid upfront.
PRICE · 03 Will you discount the day rate?
No. When budget pushes back, the standing answer is: hold price, reduce scope, phase the work. The day rate is the line — once it bends, every future engagement pushes it. Confidence in the rate is part of the brand.
PRICE · 04 What costs are passed through?
Hosting, third-party AI APIs, and other API or data licenses are passed through at cost. Sydney metro travel for state-government on-site sessions is absorbed; interstate or regional travel is passed through plus a half-day rate for travel time.
STRUCT · 01 Who is actually delivering my project?
An orchestrated agent team, operated by Nirman Kesari (founder), with a technical-cofounder agent driving architecture and a fleet of specialised agents underneath. Each agent has a defined scope, a tool list, an escalation path, a versioning scheme, and an eval. Failures are root-caused; “the model was weird that day” is not an acceptable answer. Senior human specialists are engaged on demand for judgement — domain depth, accountability, design review — never as throughput. Named per engagement.
STRUCT · 02 Why call yourself an AI lab if you're not a big team?
Because the firm is not a labour pyramid; it is an agent organisation. The cofounder is a technical-cofounder agent. The founder operates the team. Senior humans come in for judgement on demand. That structure — agents under production discipline, judgement layered on top — is what an AI lab actually looks like in 2026. The category word is earned by the structure, not by headcount.
STRUCT · 03 What does agent-led delivery mean in practice?
It means the unit of work is not a person-hour. It is an agent operating under production discipline — defined scope, eval set, escalation path. The founder orchestrates that team and carries judgement; senior human specialists are engaged for the binding constraints. Demos show the agent loop in action — you see how the work gets made, not just the work itself. The repository runs in your GitHub org at launch; the agent stack and operating discipline travel with the engagement.
STRUCT · 04 Where are you based and what timezone do you work in?
Sydney, Australia. Sydney timezone for engagement. Clients across Australia and New Zealand.