Focus Built for Kenyan & African SMEs

Technology that fits
how your business works.

Custom software, smart automation, and AI-powered tools — designed around the real constraints of small and medium businesses in Kenya and across Africa.

Focused on clarity, simplicity, and long-term maintainability.

SME-first Local market realities
Automation Less manual work
AI workflows Practical, not experimental
End-to-end Idea to running system

Context

African SMEs are not lite versions of Western enterprises.

Core focus

Local context first

M-Pesa reality

Most SaaS treats mobile money as an edge case, so reconciliation becomes somebody's Sunday. We design around how money actually moves here — the statement formats, the partial payments, the reference that doesn't match anything.

Lean team scaling

Automation that helps small teams operate like larger ones — fewer manual handoffs, clear workflows, and systems your staff can run day to day.

Resilience

Connectivity is not guaranteed. We design for offline-tolerant flows, sync, and operations that keep working when the network does not.

Core areas of work

Three capabilities, delivered independently or together — depending on where you are and what you need.

01

Custom Software & Systems

Backends, APIs, and SaaS products built for operational complexity. Event-driven architecture, USSD/mobile-first systems, data pipelines.

  • Backend & API design
  • SaaS product development
  • Data modeling & pipelines
  • USSD & mobile-first systems
Active engagements
02

Automation

Replace manual processes with systems that run themselves. Particularly relevant for businesses dealing with M-Pesa transaction volumes and reconciliation.

  • Reconciliation automation
  • Workflow automation
  • Document processing
  • Reporting pipelines
Active engagements
03

AI & Intelligent Systems

AI agent workflows, LLM integration, and model fine-tuning for businesses ready to move beyond rule-based automation.

  • AI agent workflows
  • LLM integration
  • AI categorisation & classification
  • Model fine-tuning (PyTorch)
Emerging practice

Principles, not just deliverables

Every engagement is shaped by these — not as ideals, but as constraints we build within.

01

Document decisions, not just code

Every project produces ADRs and runbooks alongside the code. When something breaks at 11pm, your team should know why it was built this way — not just how.

02

Simple architectures that your team can maintain

The right system is one the people responsible for it can understand, debug, and extend. Clever solutions that only one person can maintain are liabilities.

03

Long-term systems, not quick fixes

We design for 3–5 year horizons. That means making trade-offs explicit, choosing boring technology where possible, and building for the team that will inherit this — not just the deadline.

Who this is for

Audriam works best with teams that want to understand the systems they're building — not just have them built.

Choose the profile that fits you.

Strong fit

  • SMEs in East Africa building financial or operational infrastructure
  • Teams investing in systems they'll maintain for 3–5 years
  • Founders who want to understand the system they're paying for
  • Businesses dealing with M-Pesa complexity, reconciliation, or manual workflows
  • Teams that want documentation and trade-off clarity, not just shipped tickets

Not a fit

  • Feature factories — tickets shipped fast with no ownership
  • Spec-driven outsourcing with no client involvement in decisions
  • Building without wanting to understand trade-offs
  • One-week MVPs with no plan for what comes after

How Audriam works

Every full engagement follows the same four stages — so you always know where you are and what comes next.

01

Discovery

We audit the current system (or lack of one), talk to the people who use it, and produce a discovery document that maps what exists, what's broken, and what needs to be built.

Deliverable: Discovery doc
02

Architecture & Design

Before any code, we design the system. Architecture Decision Records document every significant choice and the trade-offs considered — so the decisions are legible long after the build is done.

Deliverable: ADRs + system diagrams
03

Build

We build to the design. Code is documented, tested, and structured so your team can read and maintain it. No black boxes.

Deliverable: Documented code + runbooks
04

Handoff & Iteration

We don't disappear after launch. We stay close through the first weeks, document what changes and why, and make sure your team can own what was built.

Deliverable: Knowledge transfer + iteration log

Engagement types

No pitch, just a real technical discussion about your system.

Get in touch

We build for the market we serve

Alongside client work, Audriam builds its own products — proof that the approach works, and tools designed specifically for East African businesses.

Live

Areacode — estate management

A property management platform for multi-unit residential estates in Kenya. Today it runs the operational side of a gated community: who is on site, who lives in which unit, which guard is on shift, and which provider is at the gate. In live testing with real estates — the gate is where it earns its keep.

  • Gate logging — entries and exits, walk-ins and expected visitors
  • Visitor scheduling by residents, one-time and recurring
  • Occupants, households, and move-in / move-out
  • Employees, shifts, and guard check-in / check-out
  • Service provider directory with vehicle and fleet registration
  • Kenyan DPA 2019 consent gate with a consent audit trail

In development

  • Rent configuration and invoicing
  • Matching tenant payments to invoices
  • Arrears, late fees, and aging
  • Multi-estate accounts for management companies
Open Areacode
Ideation

M-Pesa & Bank Reconciliation

A tool for SMEs drowning in M-Pesa statements and manual reconciliation. It parses bank and M-Pesa statements, matches transactions against invoices, and uses AI categorisation to surface what needs attention.

  • M-Pesa & bank statement parsing
  • Invoice matching & reconciliation
  • AI-powered transaction categorisation
  • Discrepancy flagging & reporting
Discuss the problem space

Case studies

Deep dives on architecture decisions, trade-offs, and lessons learned — not vanity metrics.

Rental estate finance — in design

Rent configuration, invoicing, payment matching and arrears for the same platform. Schema and module specs are written; nothing is built. The write-up covers why money is stored as integer minor units, and why multi-tenancy had to land first.

Each case study focuses on architecture decisions, trade-offs, and lessons learned rather than UI polish or growth metrics.

Decision log

Building in public.
Thinking out loud.

Technical decisions, architecture trade-offs, and lessons learnt — written as I build. No newsletter cadence, no marketing. Just a notification when a new post goes up.

Open the log
Architecture decisions Automation AI & agents Building in public SME tech Lessons learnt
What's queued

Notes in progress: why estate multi-tenancy had to come before the finance module, storing money as integer minor units, and running a Kenyan DPA 2019 consent gate in production. Written when they're ready, not on a schedule.

New posts appear here and on /blog as they ship.

Next step

Working on something
that could run better?

If you're dealing with manual processes, reconciliation headaches, or a system that's grown past what it was designed for — let's talk. No pitch, just a real technical conversation about your situation.