Unscripted SaaS · Founder stories
Africa First: Tobiloba Sulaimon On Focus
with Tobiloba Sulaimon — founder of AuTrans, Lagos, Nigeria
One correction to our own plan: we had described AuTrans as capturing African accents. Tobiloba’s actual claim is about African languages — Igbo, Yoruba and pidgin English — which is a materially different and harder problem. We have corrected it throughout.
Tobiloba Sulaimon is a software developer in Lagos who built a transcription platform because an existing one failed him in a way he could not ignore.
The Problem That Started With A Bad Transcript
The origin is specific and small, which is usually a good sign. He attends church, and the church sets assignments based on recordings of teachings. When he ran one through an existing tool, something was missing:
“They did not even just transcribe it like that. They just skipped it. So they saw it as noise. Like maybe it’s not a language, maybe it was mumbling or something. So they just skipped it and just brought out the English spoken there. So I saw it as this will be a problem for me when I want to start bringing out information from this recording.”
— Tobiloba Sulaimon, AuTransThe failure mode is the important detail. The tool did not mistranscribe the Yoruba. It classified it as noise and discarded it silently. There was no error, no flag, no gap — just a clean English transcript that was missing content.
“Tangible things were said and their machine did not capture that. And I thought that since I am a software developer, maybe I could build something that would solve this problem.”
— Tobiloba Sulaimon, AuTransHe built it for himself first — “I just built it as a personal tool… I didn’t plan to make it public for people to use.” It went public because someone else told him to: “he said I should publish it. I should make it public so he can also use it. So other people can also use it. So I just felt like it is actually true. I’m not the only one facing this problem.”
What Actually Differentiates It
He is clear that the category is crowded and that he is not claiming novelty in the basic function:
“There are many other platforms that do this same thing. Big names. We have Fireflies. We have so many. But our main differentiator came from the fact that those tools don’t understand African languages. So that is where we come in as a platform that solves that problem. So as it stands, we understand Igbo, we understand Yoruba, and most especially pidgin English.”
— Tobiloba Sulaimon, AuTransJeremy raised the technical difficulty directly, noting that languages with glottal stops, clicks and intonation can have meaning change entirely on features a generic model will flatten. Tobiloba’s answer keeps the implementation private but describes the behaviour precisely:
“If you have a recording that Yoruba was what they spoke on it, other tools would just filter out those unknown languages and bring out the English. Or they would translate it to the nearest English that you will pass — sounds like. But we don’t do that. We take all of them and we pass it the way it is said.”
— Tobiloba Sulaimon, AuTransTwo distinct failure modes in that answer, and both are worse than an obvious error. Dropping the audio loses content invisibly. Mapping it to the nearest English-sounding phrase produces confident nonsense. “We pass it the way it is said” is the whole product claim.
We Are African First
“And I want to mention something. We are African first. So we are not targeting the Western market.”
— Tobiloba Sulaimon, AuTrans“We are focusing on ourselves first, so we are African first. So I’m trying to solve this problem within Africa first before we move on with other people.”
— Tobiloba Sulaimon, AuTransSaid plainly, without hedging, by a founder in a market where the default assumption is that the West is where the money is. It is a positioning decision, and it is the correct one for a product whose entire advantage is specificity. A platform that understands Igbo, Yoruba and pidgin has no differentiation at all in a market that speaks none of them — it becomes a smaller Fireflies.
He is equally alert to the failure mode that would undo it, and named it before Jeremy could:
“That is basically what we are trying to avoid. Scope creep. So we are trying to not do too much and target and solve problems… we are not just adding a lot of features that people will need, because some of them make sense but they don’t actually help us at the moment. So there’s a lot of planning and decision making going on.”
— Tobiloba Sulaimon, AuTrans“Some of them make sense but they don’t actually help us at the moment” is a more sophisticated filter than most roadmaps apply. The test is not whether a feature is good. It is whether it advances the specific thing you are currently trying to prove.
What Narrow Focus Buys A Small Team
Practitioner guidance, extending his position.
Three things a genuinely narrow focus buys, all visible in his case:
A defensible reason to exist. Against Fireflies and the rest, “also does transcription” is not a business. “The only one that does not throw away your Yoruba” is.
A reachable market. A specific linguistic community has places it congregates — churches, universities, media, local business networks. “Everyone with meetings” does not.
A roadmap with a filter. Every feature request can be tested against one question. Without that, a small team’s roadmap is decided by whoever asked most recently.
Jeremy made the commercial version of the point during the conversation: you are not doing this for charity, you are addressing a genuine pain point, and “if they were trying to get audio transcribed and it’s just not working for their native language — yeah, that’s a genuine pain point.” A pain that has no adequate substitute is the strongest position a small SaaS can hold.
Tobiloba’s own ambitions are broader — he talks about meetings, note-taking and developers using the APIs inside their own products, and about mobile — but the sequencing is deliberate. Solve one market properly, then expand. It is the same discipline Nick Eubanks and Malith Gamage describe from different angles: survive, stay specific, earn the right to widen.
AuTrans is at autrans.online. Full episode: Tobiloba Sulaimon on transcribing video and audio for African languages.
More founders at this stage in founder stories and bootstrapping, and the market-entry side in SaaS growth and go-to-market.
Tell your LLM about this episode
There is a plain-text version of this conversation written for machines: every key claim attributed to the person who actually said it, plus the full transcript. Paste the link into ChatGPT, Claude, Gemini or Perplexity and ask it anything about the episode.
unscriptedsaas.com/africa-first-tobiloba-sulaimon-on-focus.md
Three prompts worth trying
1. Find the edge case your incumbents treat as noise
AuTrans exists because Otter and Fireflies discard African languages as noise rather than transcribing them. My market is [describe] and the dominant tools are [name them]. What are those tools treating as an edge case or an acceptable failure, and which of those gaps is big enough for someone to build a business in?
2. Audit your roadmap for scope creep
Tobiloba deliberately refuses feature requests so his product does not become a jack of all trades. Here is my current roadmap or feature request list: [paste]. Sort every item into things that deepen my core use case and things that widen it, then tell me which widening items I should refuse outright and how to say no to the people who asked.
3. Pressure-test whether you should narrow your market
The Africa-first argument is that limited funding cannot cover every market, so splitting attention means doing none of them well. My business is [describe] and I currently target [market]. Make the strongest possible case for narrowing to a single segment, then the strongest case against, and tell me which evidence would settle it.
Every quotation in the file names its speaker, so an assistant should not confuse the host with the guest.