← Services06 / Data analytics & automation

Smarter business operations.

One trusted data model, three dashboards, decisions made faster.

Why this matters

Most businesses are drowning in data and starving for decisions.

Every business we work with has more data than they realise, in their POS, their CRM, their accounting package, their delivery app, their WhatsApp inbox. The problem is rarely a shortage of numbers. It is that the numbers live in seven systems that disagree with each other, and nobody trusts the report enough to act on it.

Our analytics work starts with picking three decisions your leadership needs to make better, and tracing back to the data those decisions need. We then build one model, usually in dbt or Python, that everyone in the business can agree is the source of truth. Dashboards come last, not first. Three of them, each for a role and a decision.

Once the foundation is in place, automation gets cheap. Stock reordering, invoice chasing, KYC checks, daily ops reports, the obvious wins start paying back within weeks.

Our capabilities

What we actually deliver in this practice.

Each engagement is scoped against one or more of the capabilities below. Pick what you need, leave what you do not. No surprises, no hidden line items.

Capability 01

Data warehouse & modelling

One trusted data model everyone in the organisation agrees with, instead of seven dashboards quietly disagreeing.

Source auditEvery system that produces data catalogued and scored for freshness and trust before anything gets modelled.
Extract & loadConnectors from POS, CRM, school MIS, hospital HMIS, accounting, M-PESA and your existing Sheets workbooks.
Modellingdbt or Python project producing the metrics finance, operations and leadership all agree to use as gospel.
DocumentationWritten data dictionary so a new hire can use the model productively on day one rather than week six.

Capability 02

Executive dashboards

Three views, not thirty. Each one tied to a real decision a real leader makes. Built for CEOs, principals, medical directors and treasurers.

Revenue & growthPipeline, conversion, revenue and forecast for businesses. Admissions and giving trends for institutions.
Operations dashboardThroughput, fulfilment, attendance, SLA and exception alerts for whoever runs the floor day to day.
Finance dashboardCash, AR, AP and margin trends tied to your accounting truth rather than to estimates and ageing spreadsheets.
Mobile-friendlyPower BI, Metabase or Looker Studio built to be readable on a phone during a board meeting.

Capability 03

Workflow automation

Quiet automations between the tools you already use. No system replacement and no migration project required.

Sheets to actionTrigger flows from Google Sheets to M-PESA, WhatsApp, your CRM, your warehouse or your finance system.
Document automationAuto-generated invoices, statements, contracts, fee notes and purchase orders from live data, not stale templates.
Approval chainsRouted sign-offs with audit trails for finance, operations, admissions and procurement teams.
Scheduled tasksDaily exports, weekly reports and monthly reconciliations running on their own without anyone pressing a button.

Capability 04

AI-assisted processes

Language-model-powered triage and document work, hardened with guardrails, logs and fallbacks to a human.

Inbox triageClassify and route incoming WhatsApp and email enquiries to the right human, with confidence scoring on every classification.
Document extractionPull structured data from invoices, IDs, exam results and unstructured PDFs. Validated, not blind.
SummarisationDaily operations reports, customer-feedback summaries and call notes generated overnight for the morning meeting.
GuardrailsLogging, fallback to humans and audit trails built for accountability, not novelty or demo magic.

Capability 05

Reporting & alerts

Numbers in front of the right person at the right time. Not buried inside a dashboard nobody opens until quarterly review week.

Email reportsDaily, weekly and monthly digests delivered to leadership inboxes on a fixed schedule everyone learns to expect.
Slack & WhatsApp alertsReal-time exception alerts when something drifts beyond an agreed threshold, with the responsible person tagged.
Board packsQuarterly board-ready reporting with consistent narrative and on-brand design, ready to print or present.
Self-serve explorationLight analyst-mode access for the curious people in your team, scoped so they cannot break the model.

Capability 06

Operational analytics

Metrics that move decisions, rather than vanity charts in a quarterly slide deck that nobody actually actions.

Lead-time trackingOrder to fulfilment, ticket to resolution, lead to quote, and admission application to acceptance, measured and trended.
Churn & retentionCustomer, member or beneficiary cohort analysis with predictive signals where the data genuinely supports them.
Marketing attributionFirst-touch, last-touch and multi-touch attribution with the underlying assumptions documented honestly.
SLA trackingService-level reporting tied to client contracts, donor reporting requirements and internal performance targets.

Stack

PythonSQLdbtMetabasePower BILooker StudioPandasAirflow

What changes

Outcomes our clients see.

1 model

Everyone agrees on

Sales, finance and ops finally seeing the same number for the same word.

3 dashboards

Not thirty

One per role. Read in five seconds. Actioned the same day.

10–20 hrs

Saved per week

Typical recovery from removing manual report-building in the first month.

How we work

Four steps. Same rhythm every engagement.

Source audit

Every system that produces data is catalogued and scored for trust, completeness and freshness.

One model

A dbt or Python project that produces the metrics everyone in the business agrees on.

Dashboards

Three dashboards, not thirty, each one for a specific role and a specific decision.

Monthly review

One hour each month to tune the model based on what is actually being used and what is gathering dust.

Where this lives

Real examples from the market.

Retail

Daily stock-and-margin dashboard for a Nairobi supermarket chain, pulling from POS, suppliers and accounting.

Logistics

Fleet utilisation and fuel-cost analytics for a cross-border haulier serving the Northern Corridor.

Healthcare

Patient-throughput and revenue-cycle dashboard for a private clinic group across three counties.

Finance

Loan portfolio quality, arrears and PAR-30 tracking for a SACCO with 40,000 members.

Common questions

Things buyers usually ask before signing.

Do we need to move all our data to a warehouse?

Usually not. We pick the smallest infrastructure that works, sometimes a single Postgres database with views is plenty. Warehouses earn their keep above a certain volume; below that they are overkill.

Can you automate without replacing our existing software?

Yes. Most of our automation work sits between the tools you already use, Sheets, Sage, your CRM, M-PESA, WhatsApp. We rarely need to replace the underlying systems.

What about data security?

We sign NDAs, we work in your environment when you prefer, and we follow the Data Protection Act 2019 for any personal data. We can also align to your ISO or SOC posture if you have one.

Ready to start? Let's talk.

One business day. Written plan. Fixed budget. No PDFs in a fortnight, no exploratory calls in disguise.

Start an analytics project ← All services