Following last week's post, I reached out to people in my community to find out how they're actually leveraging AI at work. I thought about doing this as a survey as well but decided to kill two birds with one stone and instead have 1:1 responses as well as begin to practice my podcasting skills ahead of Season 2 of my podcast.
Listen to my podcast Bridging The Gap on Spotify or Apple Podcasts. Subscribe to be the first to know when new episodes drop.
Here's the thing: when AI advancement gets discussed globally, Africa is either absent from the conversation, mentioned in some distant futuristic context of how AI will power Agriculture, or worse - only brought up when talking about climate concerns and big tech opening data centers across the continent.
But the reality? Everyday professionals here are already using these tools in surprisingly embedded ways to me. I dusted off my mic after almost a year to record a new "A Thing in 5" episode (a spin-off podcast just for substack), speaking with lawyers, economists, and startup operators across Nigeria and the diaspora about how AI is actually being used in their work lives.
What I found was more interesting than I expected.
The conversations were revealing. Not because people are doing anything particularly revolutionary, but because of how practical and, frankly, human their approaches are. There's a gap between the AI hype and the AI reality that I think is worth exploring.
What People Are Actually Getting From AI
The most interesting pattern that emerged was how people are using AI as what one person called an "extended mind" - not to replace their thinking, but to amplify it.
Loye, a lawyer working between London and Dubai, uses AI to make his communication "tighter and punchier." His firm has embraced legal AI tools that handle confidentiality requirements, and he uses them for everything from refining emails to planning case strategies. What struck me was his clarity about the boundaries he respects when working with AI tools.
Pelumi, who leads people operations at an insurtech startup, used AI to handle recruitment. But it wasn't just about efficiency around candidate screening but in various aspects of the hiring lifecycle such as creating the tests, assessing and creating parameters to determine best fit in the company. In her personal life she even built a cute anniversary website for her partner using prompt engineering on the platform Bolt which she mentions several times.
Fadekemi, an economist, treats AI like "a very fast, slightly robotic research assistant that doesn't sleep." When the World Bank drops a 100-page report on Nigeria, she uses AI to get a first take on what's new, then builds her real analysis from there. It's AI as a starting point, not an ending point.
The common thread? People are using AI to cut through noise, handle the information heavy parts of their work, and free up mental space for the thinking that actually matters. It's less "AI is taking over" and more "AI is augemeing my work."
The Trust Issues Are Real
But let's talk about the elephant in the room - nobody fully trusts these tools yet, and for good reason.
Hallucination came up in every single conversation. Loye and our anonymous guest mentioned how AI sometimes gets things wrong, and they’re able to catch things only because they know the correct answer. Pelumi found that chatGPT assessed candidates who weren't even on interview calls, giving them random scores however, Claude performed better in her experience, but the underlying issue remains: you can't outsource your judgment.
Fadekemi put the other concern perfectly: "It's like someone giving you really confident advice but refusing to tell you where they heard it." The lack of transparency in how AI reaches conclusions makes it hard to trust, especially when working in business areas where context is everything. Nigeria in 2025 is not Nigeria in 2022, but AI doesn't always know that unless you take the time to drill down prompting as our anonymous Guest did.
The credibility question is huge. People are asking for verifiable sources, pushing back when AI gives lazy responses, and constantly double-checking outputs. The tools are helpful, but they require active management. You can't just set and forget.
What We Actually Want AI To Do
This is where the conversations got really interesting. When I asked people what they'd want an AI assistant to handle, the answers weren't about replacing human creativity or decision-making. They were about handling the administrative chaos of modern life.
Fadekemi wants an AI tool that can "run my entire personal logistics life". This includes plan her week, book workouts when meetings aren't back-to-back, suggest meals based on what's in the fridge, remind her to text friends back. As she put it, "being a millennial in Nigeria right now just feels like running a startup and you're the startup."
Loye wants AI tools that can plug into his bank account and auto-invest based on his goals, or analyse his direct debits and suggest which services to cancel or switch. Basic financial management, but intelligent which given where the global economy is is needed more than ever as people navigate this anxious socioeconomic and political times.
Fuad talked about wanting AI that could draw patterns across different aspects of life such as connecting sleep data from their watch to their mood patterns in voice journals. Imagine journaling for years, then having your journal start to speak back, helping you understand your own patterns.
The wish list isn't about replacing human connection or creativity. It's about handling the mundane so we can focus on what matters. It's about AI as a chief of staff for adulthood, not a replacement for human judgment.
In Summary
What strikes me most about these conversations is how grounded they are. People aren't waiting for AGI or getting caught up in dystopian fears. They're using the tools available today, understanding their limitations, and imagining futures that enhance rather than replace human capability.
But there's something interesting I noticed: most people are primarily using generative AI capabilities - asking ChatGPT to summarise documents, using Claude to assess candidates, getting AI to refine emails. Only a few are leveraging AI agents or no-code tools to make AI more proactive.
Take Pelumi, who built that anniversary website using prompt engineering with Bolt. That's closer to what we call "vibe coding" where you’re using AI to actually build and deploy functional tools rather than just generate text. Imagine if more people took this approach: instead of asking AI to organise your schedule and still having to implement it manually, you could build (or have AI build) automated systems that actually book your workouts, reschedule meetings based on your energy patterns, or manage your investment portfolio according to preset rules.
The potential is there. Tools like Zapier AI, n8n, or even advanced GPT/Claude and GCP Actions could help people move from "AI as consultant" to "AI as executor." But it requires a slight mindset shift from prompting to designing systems.
The real AI revolution in the workplace might not be about superintelligence or extensive job displacement. It might be about giving people back time and mental space to be more thoughtful, more creative, more human.
But we're not there yet. The tools are helpful but require constant supervision. The trust issues are real. AI is a powerful tool that requires wisdom to use well. It's not magic, it's not sentient, and it's not going to solve all our problems . And the gap between what we want AI to do and what it can actually do remains significant despite all the hype being pushed out by silicon valley.
Come back again next week for part 3 as I explore what the impact of AI is on creative roles.
Disclaimer: All individuals interviewed for this post spoke in their personal capacity and their views do not represent those of their employers or affiliated organisations.
Share this post