Four kinds of work, one way of working

Some of what a business loses goes through process — work done twice, errors found late, hours spent moving numbers between systems. Some of it goes through exposure nobody is watching. We work on both, we measure before we build, and we keep operating the result afterwards.

Four groups

Four kinds of work, and the symptoms that usually bring someone to each. Most engagements start in one and touch a second.

Custom management systems

Your operation runs on a process that no product you can buy actually fits, so it runs on spreadsheets and people remembering.

We build the internal platform around how the business already works, rather than bending the business to fit someone else's software. That means the process mapping happens before the data model, and the data model happens before anyone writes a screen — because a system that encodes the wrong process is more expensive than the spreadsheets it replaced.

  • Requirements and process mapping
  • Data model and workflow design
  • Role-based access and audit trails

You probably need this if

  • Nobody can close the month while a particular person is on leave
  • Two systems hold the same figure and they disagree
  • Renewals depend on someone remembering to check a list
  • Your member or customer count and your billing count do not match

Operational automation

Closing the month takes four days, and three of them are spent moving numbers between systems by hand.

We find where the numbers actually live, connect the systems that hold them, and build the reconciliation, reporting and handoff steps that someone is currently doing manually. Integrations get retries, a record of what synced and an alert when they stop — the person in the middle goes back to the job you hired them for.

  • Process instrumentation
  • Scheduled and event-driven jobs
  • Integration between systems that don't talk

You probably need this if

  • Someone rebuilds the same report every month from the same exports
  • A recurring calendar reminder exists purely to move data between two systems
  • Nobody can tell you when a sync last succeeded
  • The answer to 'how do we know that' is a person's name

Security and compliance engineering

You have logs nobody reads, a scan report nobody worked through, and a customer starting to ask for evidence.

We build and operate the controls, evidence and documentation that regulators and certification bodies actually ask for — monitoring that reports things worth reading, findings triaged by what they would cost if exploited, and a management system that produces the proof of its own operation. What you get is a defensible position you can explain to a bank, an insurer or a board, not a clean scorecard.

  • Management system design and implementation
  • Control and evidence automation
  • Regulatory mapping

You probably need this if

  • You would find out about a compromise from a customer
  • Alert email goes to a distribution list, not a person
  • A scan report exists that nobody has worked through
  • A customer or regulator has started asking for evidence

Digital presence and operations

The site was delivered, invoiced and never touched again — and nobody is certain who holds the domain.

We build and then keep running the public-facing systems: the site, the hosting, the certificates, the backups and the monitoring that tells someone when it stops. Maintenance is the part that is usually missing, so it is written into the arrangement rather than left to whoever notices first.

  • Site design and build
  • Hosting, monitoring and backups
  • Ongoing content and technical maintenance

You probably need this if

  • Nobody is sure who controls the domain or the DNS
  • The last content change required finding the person who built it
  • You would learn the site was down from a customer
  • There is no backup you have ever tested restoring

How we price

Every engagement is one of three shapes. Which one applies is usually clear after the first conversation.

Fixed-scope diagnostic

A defined piece of work with a fixed price and a written output: what the problem costs, what we would do about it, and what that would cost. It stands alone. If you take the findings elsewhere, that is a reasonable outcome.

Project build

Scoped from the diagnostic, priced as a project with agreed milestones. We do not quote a build before we have measured the thing it is meant to fix.

Monthly operations retainer

For work that has to keep happening — running the monitoring, triaging what it produces, maintaining a system after handover. Priced monthly, with an agreed scope and an agreed responder.

What we will not do

  • Build a system when a configuration change would do

    If the tool you already pay for can do it, we will show you how and invoice you for the afternoon rather than the quarter.

  • Take on monitoring with no agreement on who responds

    Detection without a responder is an expensive way to produce a record of things that went wrong. We agree the response path first or we do not start.

  • Quote a build before measuring the problem

    A number produced before the diagnostic is a guess wearing a proposal. We would rather lose the work than pretend otherwise.

  • Take work we cannot staff properly

    We are small. When something needs a team we do not have, we will tell you that at the first conversation rather than the third month.

Name the process that costs you the most

Tell us which one, and roughly what it costs you in hours or cedis. We'll come back with either three questions or a proposed 30-minute call.

Book a scoping call