01 / Where AI pays
Most of the return sits in unglamorous work.
Three places to look first.
Moving information around
Data entry, reports, inbox triage, approvals. If a task is "read this, decide that, write this", AI handles it well and the hours are real.
Where speed is the product
Quotes, proposals, answers to customers. If you're slower than a competitor, you lose deals before you know they existed.
Where errors cost money
Invoice mismatches, compliance checks, repetitive quality control. The return is what you lose today, not what a vendor promises.
02 / Where it's theatre
Three ways to spend on AI and get nothing back.
- A strategy deck written before anything works.
- Tool licences bought for everyone, with no process that changes.
- A chatbot nobody asked for, so the board can say "we do AI".
None of these are bad on their own. They just don't pay unless they start from a real cost in the business.
03 / How I work
One process at a time, measured.
Find the cost
Where does the business lose hours or margin right now? Not where AI could theoretically help.
Decide: build, buy or skip
Often an existing tool does the job. Sometimes the honest answer is "not yet".
Make it work in daily use
One process, measured against how it runs today. A demo that wins applause is not the goal.
Get the team to own it
People before tools. A team that understands what AI can and can't do gets more from a cheap tool than an untrained team gets from an expensive one.
04 / Why a CTO, not an agency
Built by someone who has had to answer for it.
At Aximpro I worked with enterprise clients including Dell, HPE and Vertiv, and built PartnerValue, an enterprise SaaS AI platform. At NYALA I was VP Engineering and then CTO of a regulated fintech platform. That's the lens I bring: what it does for revenue, margin, speed and risk. The technology comes second.
What sets Sanjeev apart is his exceptional ability to diagnose and resolve software challenges.
05 / Questions
Questions people ask
Where should a mid-sized company start with AI?
With one process that costs real hours or margin every week and has clear inputs and outputs, such as quotes, document handling or customer requests. Not with a strategy deck or a tool licence. Prove it there, then decide what comes next.
Do we need an AI strategy before we start?
Not a long one. You need to know which business problem you are solving and how you will measure it. A strategy written before the first working use case is usually guesswork.
Do we have to train our staff on AI under the EU AI Act?
Article 4 of the EU AI Act has applied since 2 February 2025 to companies that use AI systems at work. Since the 2026 Digital Omnibus amendment, the duty is to take measures that support the AI literacy of staff who use AI on the company's behalf, without having to guarantee a specific level. It is not a certificate or an exam. Check the details against the official text or with a lawyer. More in EU AI Act Article 4, explained.
Do you work outside Munich?
Yes. I am based in the Munich region and work on-site in Munich and remotely across Germany and Europe. I work in English.