HARRISON NDEKE / TECHNICAL JOURNAL
Practical engineering notes on AI systems, automation, and the boring-but-important work that keeps a workflow dependable after launch.
FEATURED NOTE / 00148-Hour Assessment / Buyer Guide / Workflow Audit / Security Review / Kenya Business
A fixed-fee, 48-hour engagement that maps your business workflows, ranks the automation opportunities by value, reviews where personal data and money flow, and hands you a fixed-quote proposal — with no obligation to build.
Read the field noteTHE ARCHIVE
Each note starts with a real problem, separates confirmed practice from opinion, and includes the trade-offs a production team needs to consider.
0148-Hour Assessment / Buyer Guide / Workflow Audit / Security Review / Kenya Business
How Harrison Ndeke's 48-hour workflow assessment works — what gets mapped, what you receive, and what it costs. KSh 35,000 fixed fee.
Read note
02KDPA / Data Protection / Compliance / AI Automation / Kenya Business
Kenya's Data Protection Act meets AI automation — consent, data minimization and the compliance checkpoints every automation project needs.
Read note
03WhatsApp / Messaging / Comparison / Buyer Guide / Kenya Business
API or app? A practical comparison of WhatsApp automation options for Kenyan businesses — cost, limits and when to upgrade.
Read note04Budget Scoping / SME / Kenya Business / Automation Scoping / Buyer Guide
Real automation scopes at real Kenyan budgets — what KSh 100,000 buys, what it doesn't, and where the value actually lands.
Read note
05Production Reliability / Failure Modes / Guardrails / Monitoring / AI Automation
The failure modes that kill automation projects after launch — silent errors, runaway agents, data drift — and the guardrails that stop them.
Read note
06RAG / enterprise search / access control / vector retrieval / n8n / provenance
A practical architecture for automated knowledge base synchronization, RBAC vector filtering, citation schemas, and retrieval evaluation in enterprise workflows.
Read note
07automation reliability / graceful degradation / n8n / fallback design
A practical guide to preserving safe core work, deferring optional enrichment, bounding retries, and routing unresolved automation work to human review.
Read note
08API design / batch processing / idempotency / workflow reliability
A practical pattern for item-level outcome ledgers, selective retries, reconciliation, and idempotency when a batch API completes only part of a request.
Read note
09AI agents / agent evaluation / tool calls / external effects
A practical guide to testing an AI agent's full trajectory, approvals, tool calls, external effects, and final system state before release.
Read note
010AI agents / tool calls / JSON Schema / authorization
A practical guide to validating AI tool arguments, enforcing business rules and authorization, and testing consequential agent actions before execution.
Read note
011backpressure / n8n / queue mode / workflow scaling
A practical guide to concurrency limits, bounded queues, retry budgets, load shedding, and overload monitoring for n8n workflows and AI agents.
Read note
012prompt injection / RAG security / AI agents / tool security
A practical architecture for treating retrieved content as untrusted data, constraining agent tools, validating actions, and testing poisoned sources.
Read note
013observability / n8n / workflow monitoring / automation reliability
A practical observability guide for connecting automation logs, metrics, traces, alerts, and synthetic checks to verifiable business outcomes.
Read note014idempotency / n8n / workflow reliability / duplicate prevention
A practical idempotency pattern for preventing duplicate messages, records, charges, and other side effects when automation events are retried.
Read note015AI agents / human approval / tool security / risk controls
A practical risk model for deciding which AI agent tool actions can run automatically, which need human approval, and what an approval gate must verify.
Read note016n8n / webhooks / security / API security
A practical security checklist for n8n webhooks covering authentication, request validation, replay protection, resource limits, safe responses, and production testing.
Read note
017n8n / reliability / automation / error handling
Learn when to retry an n8n failure, when to use a fallback, how to configure an error workflow, and when to send work to human review.
Read note018n8n / Make / automation platforms / workflow tooling / self-hosting
An evidence-based comparison of n8n and Make covering self-hosting economics, code-node flexibility, node ecosystem, and when each platform is the better production choice.
Read note
019RAG / vector search / embeddings / AI fundamentals / Supabase pgvector
A foundational explainer on Retrieval-Augmented Generation: chunking, embeddings, vector search, grounding, and why naive RAG hallucinates.
Read note
020Supabase / pgvector / RAG / row-level security / Postgres
A practical guide to using Supabase as the data layer for production RAG systems: pgvector setup, row-level security for multi-tenant retrieval, realtime, and edge functions.
Read note021AI Agents / Finite State Machines / Orchestration / Pydantic / LangGraph
Why deterministic finite state machines and Pydantic schema contracts beat unconstrained multi-agent swarms in production.
Read note022AEO / GEO / SEO / Information Gain / Google AI Overviews
Algorithmic strategies, entity-first semantic triples, and Information Gain extraction to rank in Google AI Overviews and Perplexity.
Read note023AI Memory / Vector Search / pgvector / Cognitive Architecture / RAG
Why raw vector search fails for long-term agent memory, and how a 3-tier cognitive architecture solves state, time, and causality.
Read note024AI Pricing / Automation Costs / Kenya Business / ROI / Consulting
A transparent, practical breakdown of AI and workflow automation costs in Kenya — from fixed assessment fees to custom agent platforms and infrastructure.
Read note025AI Agents / Plain English / Business Guide / Workflow / Kenya
A clear, jargon-free explanation of AI agents for business owners: how they differ from chatbots, concrete jobs they automate in Kenyan SMEs, and essential guardrails.
Read note026n8n / Zapier / Platform Comparison / KDPA / M-Pesa / Kenya
An objective comparison between n8n and Zapier for Kenyan enterprises: monthly task pricing, data residency under KDPA, and M-Pesa Daraja API integration.
Read note027Hiring Guide / AI Consultant / Kenya Business / Production Systems / Vetting
6 vetting questions to evaluate AI automation consultants in Kenya, spotting the red flags of automation theatre, and hiring for verified production reliability.
Read noteEDITORIAL STANDARD
There are no generic AI roundups here. Every Field Note is written to help a builder make a clearer technical decision: how to model failure, when to introduce review, what should be measured, and where a clean demo is not enough.
Discuss a system