Kenya's Online Workers and the AI Shift
For six years, Omondi earned his living from a small rented room in Kisumu, a second-hand laptop and a very loud fan. He transcribed interviews, wrote product descriptions and proofread essays for clients he had never met, in time zones he could not pronounce. Then, sometime last year, the transcription jobs started drying up. Clients had discovered that software could do a rough transcript in minutes. Omondi's story is one small chapter in a much bigger one: AI and online work in Kenya are now firmly tangled together.
Kenya has built a reputation as one of Africa's busiest hubs for online work. Freelance writers, virtual assistants, data labellers, transcribers, designers and developers all earn from clients around the world. Now that AI tools can draft, translate, transcribe and summarise, many of these workers are asking the same question: is this the end, or just a change of route?
The tasks under pressure
Let us be honest about what is shrinking. Tasks that are repetitive, generic and easy to check are the most exposed:
Basic transcription of clear audio
Generic blog posts and product descriptions written to a template
Simple translation of everyday text
Rewriting or spinning existing articles
Basic data entry from forms and documents
Clients who used to pay for these now often use AI for the first pass. Some still hire humans, but they pay for editing and checking rather than writing from scratch, and the rates can reflect that.
Where AI and online work in Kenya are growing together
At the same time, AI has created new types of work, and some old skills have become more valuable:
Reviewing and correcting AI output, where accuracy and judgement matter
Data labelling and quality checking, which helps train and evaluate AI systems
Specialist writing that needs real expertise, local knowledge or a distinctive voice
Transcription and translation of difficult audio, local languages and mixed-language speech
Virtual assistance that uses AI tools to handle more clients, faster
Building simple automations and workflows for small businesses
Notice the pattern. The work moving up in value is the work where a human adds something the machine cannot: context, accuracy, taste, trust or local understanding. A Kenyan who knows how a Nairobi customer actually talks, or how a Kenyan school or SACCO really operates, has an advantage no chatbot has.
How online workers can adapt
Panic is not a plan. Here is what we have seen work for people making the shift:
Learn the AI tools in your field properly, then sell speed and quality together
Move up the chain: from writer to editor, from transcriber to reviewer, from assistant to operations support
Specialise. A niche such as legal, medical, agriculture or finance is harder to automate
Build a portfolio that shows judgement, not just volume
Protect your reputation: clients notice unedited AI work, and they leave
Keep learning in small steps; even an hour a week adds up
What this means for Kenyan businesses
If you run a business, this shift is good news. Kenya has a large pool of digitally skilled people who already know how to work remotely, meet deadlines and use online tools. Many of them are now learning to work alongside AI. That means you can hire local talent to help with content, customer support, data cleanup and automation, often people who understand your customers far better than an overseas agency would.
It also means you should rethink what you pay for. Paying someone to type is becoming less sensible. Paying someone to check, improve, organise and take responsibility for the work is where the value sits.
Back to Omondi
Omondi did not quit. He taught himself to clean up AI transcripts quickly, then started offering Swahili and Dholuo transcription that the software kept getting wrong. Last month, he hired his cousin to help. The fan is still loud, but the room has a second desk now.
Your next step, whether you are a freelancer or an employer: list the tasks you do or pay for each week, mark the ones a machine could draft, and decide what human skill you will put on top. If you are curious how we think about people and technology working together, read a little more about Philsoft.