Technology consulting, building from Africa
We build the systems organisations run on.
Kathalysth designs, builds and runs data platforms, software, automation and cloud infrastructure for banks, governments, energy companies and hotels. We also help them put AI to work, starting where it clearly pays.
What we do
Six practices, one team.
Most projects draw on more than one. You work with a single team that is accountable for the result.
- 01Data and analyticsWe build data warehouses and pipelines that bring your records into one place, so teams can report on them, and so they are ready when you want to use AI.
- 02Software engineeringWe build and modernise web, mobile and back-office applications, and add AI features where they save your users real time.
- 03Process automationWe find the repetitive work in your operations and automate it, using AI to read documents and route requests where fixed rules fall short.
- 04Applied AIWe help you find the work where AI pays off, build it into the systems you already use, and keep it accurate after launch.
- 05Cloud and infrastructureWe plan and run moves to the cloud, and can host your systems and AI models in the country where your data has to stay.
- 06Research and developmentOur R&D team tests new technology before our clients have to. We run proofs of concept, publish what we learn, and turn what works into products.
Applied AI
AI where it pays, not everywhere.
Many AI projects stop after the demo. We start with one task, prove the value on your data, then build it properly.
- 01
Answers from your documents
Staff ask questions in plain language and get an answer with the source document attached.
- 02
Documents read at volume
Forms, IDs and invoices turned into structured data, with people checking the uncertain cases.
- 03
Forecasts you can act on
Demand, risk and maintenance predictions fed into the decisions your teams already make.
- 04
Data kept where it must stay
Models and data hosted in-country when regulation or policy requires it.
How we deliver AI
Kathalysth Catalyst
Our four-step method for taking an AI idea from a question to a system people use every day. Kathalysth is a member of NVIDIA Inception.
- 01
Discover
We agree the decision or task to improve, and check you have the data to do it.
- 02
Prove
We build a working version on your own data, usually within a few weeks, so you can judge it before investing more.
- 03
Build
We connect it to your systems, test it, and add controls, including human review where mistakes would be costly.
- 04
Run
We monitor quality and cost after launch, and improve or retire the model as your needs change.
Kathalysth is a member of NVIDIA Inception.
BuildReach
Our initiative for people starting in technology: cohorts that build real products, mentoring from working engineers, and help into a first role.
Tell us what needs fixing.
Describe the problem. We will tell you plainly whether we can help, and what it would take.