Writing by Arup Maity
- Where clinic revenue actually leaks
Not to insurers: to the rejection nobody worked before the window shut. What an AI worker in a clinic finance office should do, and how to test the payback yourself.
- Compliance rarely fails on judgement. It fails on the checks nobody had time for.
An AI worker can run every written check, every time, record that it did, and hand a person a short list to sign. Four UAE examples.
- Same technology, different economics: personal AI assistants and business AI workers
Why a cloud personal assistant is hard to sell for a flat fee, why the device may be its better home, and why a business AI worker is judged against a different budget.
- Jawda counts the patients you didn’t measure
Abu Dhabi’s twelve outpatient rehab Jawda indicators are computed from the clinic’s own record, and a missed reassessment counts as “not improved”.
- Three questions an “AI-native” development partner should be able to answer
The label rarely survives three follow-up questions: how much of the code is AI-generated, what gate it passes, and what happens when the AI is wrong.
- Managing a software team when agents write most of the code
When code arrives in hours, the bottleneck moves to the specification, independent testing and deciding what not to build. Accountability doesn’t move.
- You hired a coder. You needed an owner.
Hiring developers who are good with coding agents makes the typing cheap. It leaves the requirement and adoption as hard as they were, and moves the delivery risk onto you.
- UAE e-invoicing: eight checks to run before you choose a provider
Eight checks on last quarter’s invoices that show where the e-invoicing work really is. The provider fixes none of them; your order-to-invoice process does.
- AI pilots rarely fail because the model was bad
They fail because the data and process underneath were never digitised. Ask which core operations still run on spreadsheets and tribal knowledge, and fix that first.
- “AI makes this cheaper now, right?” Mostly, no.
What AI has changed in a software budget is where the cost sits, not how big it is. Worth knowing before you put a number in front of finance.
- Where AI belongs in real estate, and where it should stay out
The AI conversation in property is about the chatbot on the listing page. The decisions that cause real problems are rent caps, commission disputes and title checks.
- Why Your Strategy Dies (And How AI Can Changes That)
Thirty years in this industry teaches you something humbling: the best strategies rarely ship. I've watched it happen dozens of times.
- Working with AI Agents: Build or Compose. 3 Questions to Guide.
Two scenes from the last three months. Scene one.
- The Downturn Paradox: Why AI Transformation Compounds Fastest When Capital Is Tightest
The instinct in a downturn is to defer transformation: tighten belts, run leaner, wait for clearer skies. That instinct is wrong, and increasingly the data shows why.
- The Code Didn't Change. The World Did.
An AI model judged too dangerous to release has exposed an uncomfortable truth: the legacy systems quietly running our banks, hospitals, and power grids are far more vulnerable than we let ourselves believe. The threat isn't old code — it's abandoned code.
- The Transformation Investment Gap: Why Ambition and Spend Must Match
Most leaders of a growing company can tell you, to the decimal, what they spend on rent, on payroll, on customer acquisition. Ask them what they should spend on technology to actually reach their goals, and the room goes quiet.
- The Always-On Boardroom: When Strategy Stops Being an Event
- I Came Back to Coding After 20 Years — and the Fault Line on My Team Was Nothing Like I Expected