Skip to content

Current AI Services Engagements

I come into a business, assess what’s actually there, and build what it needs.

See completed engagements →

This is the work I do : hands-on engagements with small and medium businesses that have a system, a process, or a function that isn’t working the way it should — and need someone to come in, find out why, and fix it.

I work independently. I use AI tools to move faster than a traditional engagement would allow — every fix, every report, every recommendation is something I personally investigate and verify. What a client gets at the end is not just a working system. It’s something their own team owns and can build on even after my engagement ends.

How I Work

1. Assess and Audit. I start by looking at what’s actually in the system — the data, the configuration, the workflows, the reports already in use — rather than relying on what I’m told is wrong. Most organisations know something is off but not precisely what, or why. The first deliverable is always a structured findings report: what’s broken, what’s missing, and what’s possible, in priority order.
2. Fix the foundation first. Nothing built on top of broken data can be trusted. Active defects — data getting silently overwritten, records saved incorrectly, validation gaps — get fixed before any new capability is layered on top. This sequencing isn’t a preference; it’s the order that produces something reliable.
3. Deliver each fix as a complete unit, not just a patch. Every fix I make ships with the code change, the root cause analysis, before-and-after evidence from the actual data, reproduction and verification scripts, and an annotated walkthrough. The client isn’t asked to trust that something was fixed — they can see it.
4. Build a methodology record, not just a fix. Every fix also produces a record of how I found it: the investigation sequence, the underlying pattern, the evidence standard, the verification steps. The next instance of the same class of problem gets resolved without rebuilding the mental model from scratch. For clients running on Claude, these records become executable skills — structured, human-gated AI workflows the client’s own team can reuse long after the engagement ends.
5. Sequence everything so value lands early, not just at the end. Work is structured in phases. Each phase is validated by the client before the next begins, and each one delivers something usable on its own — not held back until a single big reveal at the end.

Current Engagement

● Active

Client
A regional distributor of precision scientific instruments, operating and serving customers across the Middle East. The business runs both equipment and consumables sales alongside ongoing field service operations for installed instruments.

Scope
A full remediation and intelligence programme for the company’s CRM system, covering:

  • Data integrity fixes — active defects where the system was silently corrupting or mishandling records, including a bug that could permanently overwrite equipment configuration data on a routine form submission.
  • Capability gaps — configuration and structural gaps limiting what management could see and decide on, each with a defined resolution path.
  • Operational and commercial reporting — new reports built entirely from data the system already holds, covering delivery tracking, quote ageing, contract renewals, warranty-driven sales opportunities, and supplier and currency reporting.
  • Process improvements — targeted changes to how sales, service, and back-office teams use the system day to day.
  • Service visit intelligence – An independent audit of all completed field service visits found that engineers were documenting their work, but almost nothing else: customer reactions, pending problems, parts used, and upgrade opportunities were going unrecorded in the vast majority of visits. The work in progress restructures how the service form captures this — turning what engineers already observe in the field into structured data the business can act on, without adding a new tool or any new login.

Approach
The engagement is sequenced in five phases, each validated before the next begins — starting with fixing what’s broken, then quick operational wins, then commercial intelligence, then structural and service-level capability. The client sees real output at every phase.

Engaged since May 2026

Interested in working together?

If you’re running a small or medium business and something in your operations — a system, a process, a function — isn’t working the way it should, I’d be glad to talk.

Get in touch: ranga.sampath@gmail.com