AI augmentation in 2025: what Accenture's layoffs and a Nigerian app founder reveal about the future of work
The question isn't whether you're augmented. It's whether you can deliver.
Videos have been circulating on social media showing job candidates using AI tools to generate real-time interview responses. The clips, seemingly spawned by an Instagram account called newschoolboosted, feature someone with an earpiece receiving AI-whispered answers while speaking to interviewers on camera.
The account is run by someone who goes by Kagehiro Mitsuyami; describing himself as CEO of Lockedin AI, the very tool being demonstrated. The content is essentially promotional material dressed up as candid footage, which adds its own layer to the authenticity question.
Luiza Jarovsky, co-founder of the AI, Tech & Privacy Academy, flagged one such video, arguing that HR departments are unprepared for this kind of augmentation. Fair point. But I found myself asking a different question: if a candidate can genuinely deliver the work using whatever AI tools they have at hand, and if such use is disclosed and professionally acceptable, does that shift the practice from questionable to merely pragmatic?
The ethics crowd wants to debate whether it's cheating. The market, meanwhile, is busy answering a more brutal question.
The Accenture signal
In late September, Accenture announced it would cut approximately 11,000 staff as part of an USD 865 million business optimisation programme. CEO Julie Sweet was blunt about the rationale: the company is "exiting, on a compressed timeline, people where re-skilling is not a viable path for the skills we need."
By early December, Accenture had announced expanded partnerships with both OpenAI and Anthropic. The OpenAI deal, announced December 1, rolls out ChatGPT Enterprise across consulting, operations and delivery work. The Anthropic partnership, announced December 9, will train 30,000 employees on Claude-powered solutions and represents Anthropic's largest-ever deployment of Claude Code.
The sequence tells a story. First, exit those who cannot be reskilled for AI-augmented work. Then, double down on AI tooling for those who remain. The professional services industry, which has long sold human expertise by the billable hour, is recalibrating what that expertise actually means. Those who can work effectively with AI tools stay. Those who cannot, or will not, are being shown the door.
Andile Masuku
The solopreneur test
Nearly a year ago, I wrote about the tantalising prospect of AI-powered solopreneurship: the idea that generative AI tools could enable individuals to operate with the output capacity of small teams. I've spent the intervening months testing that thesis through my own journalism and independent consulting practice.
The verdict? AI augmentation is real, but it amplifies what's already there rather than conjuring capability from nothing. It accelerates research, sharpens drafts, and compresses workflows. It does not substitute for domain knowledge, editorial judgment, or the relationships that generate opportunities in the first place. The tools are force multipliers, not replacement forces.
Which brings me to Kelechi Onyeama.
From Lagos to acquisition
In a recent podcast interview, 22-year-old Nigerian app developer Kelechi Onyeama detailed how he built Social Wizard, a consumer app that achieved over 500,000 downloads and tens of thousands in monthly revenue, while effectively homeless in the United States. His story offers a masterclass in resourceful execution under constraint.
Onyeama reportedly taught himself to code, built his first app on a broken Dell laptop in Nigeria, then scraped together roughly USD 100 to fly to America on a tourist visa to give his venture one last push. When family arrangements collapsed, he found himself bouncing between relatives' homes, eventually landing with a five-month deadline to make something work or return home.
His breakthrough came from timing and iteration. He built Social Wizard alongside the release of GPT-4 Vision's API, creating functionality that didn't previously exist. When early marketing attempts flopped, he analysed which content formats showed any traction, then partnered with micro-creators willing to work for USD 120 per video. One clip hit two million views. Revenue went from zero to USD 25,000 in weeks.
Andile Masuku
The product itself (an AI tool designed to help users craft messages in dating contexts) raises its own questions about authentic human connection. Call me a stick in the mud or rigidly wedded to my integrity values, but Onyeama's ingenuity, resilience and resourcefulness notwithstanding, even if I were single and ready to mingle, Social Wizard wouldn't be something I'd use to augment my social skills. Highlighting his prowess is not an endorsement of his product.
But his process was authentically scrappy: identifying a problem, timing his entry to technological capability, iterating relentlessly on distribution, and converting early revenue into stability.
His trajectory illustrates a framework I've been developing called This Is Connectivity. The model isn't linear: skills, expertise, resources and connectivity interact dynamically, with connectivity itself something that can be acquired, inherited, endowed, nurtured or invested in throughout a career.
Onyeama learned to code, built products that generated revenue, then connected with established app developers in San Francisco. But those relationships weren't merely the outcome of his earlier efforts; they accelerated everything that followed, unlocking knowledge transfer, credibility and eventually an acquisition of his second app. Within months, he had secured an O-1 visa for individuals of extraordinary ability.
The real question
The debate about AI-assisted interviews is a proxy war for a larger anxiety: that the rules of professional performance are shifting beneath our feet. The discomfort is understandable. But the hand-wringing obscures what's actually happening.
Accenture isn't asking whether AI augmentation is ethical. It's asking whether employees can deliver value with these tools, and restructuring accordingly. Onyeama didn't wait for permission to leverage emerging AI capabilities; he built a business on them. The question isn't whether augmentation is fair. It's whether you can deliver.
That said, disclosure does matter. Using AI tools without transparency is a different proposition from using them openly. The professionals I respect most are explicit about their workflows, including their use of AI. Integrity isn't about refusing augmentation; it's about being honest about how you work.
As I close out my TechTides columns for 2025, I'm thinking about how to systematise and share the lessons from this year of live-testing AI-augmented solopreneurship.
One of my mentors, Thomas Debass, who has spent years building frameworks for public-private partnerships at the US State Department, has inspired me to start fleshing out a fellowship-style approach to knowledge transfer. Watch this space.
For now, the message is simple. The ethics debates will continue. The market won't wait. Augment or exit.
Editorial Note: A version of this opinion editorial was first published by Business Report on 16 December 2025.