Field notes
Notes
Short, practical answers to problems teams hit before they hire anyone.
Payments and invoicing
- eTIMS integration: OSCU or VSCU, and what changes in your system
KRA eTIMS system-to-system integration explained: OSCU vs VSCU, sandbox testing and certification, and what your invoicing system has to change.
6 min
- How to reconcile M-Pesa payments with your invoices
How to reconcile M-Pesa and bank payments with invoices: what to match on, wrong account numbers, partial payments, duplicates, and what to automate.
6 min
- M-Pesa STK push callback not received? Check these in order
Your customer paid but the callback never came. Why M-Pesa STK push callbacks go missing, what to check in order, and the fix that stops relying on them.
5 min
Business systems
- What a useful WhatsApp CRM should automate
A practical guide to WhatsApp Business automation and CRM solutions for SMEs: shared inbox, lead capture, ownership, follow-up, customer history, integrations, and where AI helps.
6 min
- Your cloud VPS blocks SMTP. Send email over an API instead.
A friend's freshly deployed SaaS app could not send a single email from a DigitalOcean droplet. The cause was not the code or the credentials. Cloud providers block SMTP ports by default, the failure is silent, and the fix is to send over an HTTPS email API.
6 min
- Small business tech fixes that should not become rebuilds
A practical list of website, email, DNS, SSL, booking, form, M-Pesa, and AI-built app problems that can often start as a small fix before anyone quotes a rebuild.
4 min
- What to automate first in a law or accounting firm
Professional firms do not need AI everywhere on day one. A practical order for automating client intake, booking, document prep, and follow-up without breaking trust or compliance.
6 min
- What to do when your business website goes down
Your site, domain, email, or M-Pesa link is down and customers cannot reach you. A practical triage order for small businesses: what to check first, what not to touch, and when to call for help.
5 min
AI in production
- The same calculator, different grades: a lesson for using AI
Everyone in a maths exam can have the same calculator and still get different grades. AI makes the same distinction visible: a powerful tool helps, but understanding, judgment and checking the work still matter.
5 min
- AI cost per user: how to model it before you ship
Founders ask 'how much will the AI cost?' and quote API prices. The API price is the least useful number in that conversation. Here's a practical model for cost per active user per month: the six terms that matter, and the five levers that actually move them.
4 min
- Why AI keeps fixing your app into new bugs
When your AI coding tool gets stuck, another prompt often makes the app worse. Break the debugging loop with reproduction, evidence, small diffs, and tests.
4 min
- Your AI-built app is a prototype until production proves it
AI coding tools can get a SaaS demo online fast. Use this AI-built app production-readiness checklist before real users, payments, private data, and integrations.
5 min
- Four ways to ship AI in your fintech (and how to pick)
Build, buy, hire outside, or wait. Every fintech founder facing an AI decision has these four options, and the right one is rarely the obvious one. The costs and tradeoffs nobody puts on a slide.
4 min
- How to scope an AI engagement when you have never done one
A practical framing for founders about to commission their first AI project: the right starting question, the four things to define up front, and the red flags that usually mean trouble.
4 min
- From AI demo to product: where the real work is
Many AI projects never ship. In my experience, the reason is rarely the model. A short note on the engineering work that lives between a working demo and a feature in real customers' hands.
3 min