A family-owned pharmacy in Pretoria introduces an AI system to manage stock, predict demand and flag possible drug interactions. Within months, stockouts fall, ordering takes half the time and the pharmacist catches two potentially serious interactions the system highlighted. It’s a clear success.
But the owner notices something else. The customers who keep coming back don’t talk about the efficient ordering. They talk about the pharmacist who remembers their children’s names, explains side effects patiently and notices when an elderly customer seems confused. The technology made the pharmacy more efficient. The people made it trusted.
Across almost every industry, AI is changing how work gets done: in retail, healthcare, finance, logistics, professional services and beyond. The businesses gaining most from it are rarely those that simply bolt it on. They tend to be the ones that rethink how work flows and keep people deliberately involved where judgement, care and trust matter.
One Like insight: AI can make a business more efficient. People still make it worth choosing.
What the data says
McKinsey & Company runs one of the largest regular surveys of how organisations use AI. In its State of AI survey published in November 2025, 88% of respondents said their organisations regularly used AI in at least one business function, up from 78% a year earlier. Adoption was spreading across industries, from technology and professional services to retail, healthcare and energy.
Yet broad use hadn’t translated into broad results. Only around a third of organisations had moved beyond experimenting and piloting to scaling AI across the business. McKinsey found that intentionally redesigning workflows was one of the strongest contributors to achieving meaningful business impact. In other words, value came from changing how people and technology work together.
Human oversight was a recurring theme. In McKinsey’s earlier 2025 survey, organisations varied widely in how much AI-generated content people reviewed before use: about 27% said employees reviewed all of it, while a similar share reviewed 20% or less. The latest research reported that the organisations getting the most from AI were far more likely to have defined processes for deciding when AI outputs need human validation.
The surveys mostly reflect larger organisations, and they’re self-reported. McKinsey also sells AI consulting. But the message is consistent with other research explored in this series: AI works best when people stay thoughtfully in the loop.
Keeping the human element
For small businesses adopting AI, keeping the human element is both a safeguard and a competitive advantage. A few principles help.
Decide where people must stay in charge. Write down the decisions and outputs that always need a person: anything affecting a customer’s health, money, safety or legal position; complaints; sensitive conversations; hiring. As explored in When Good AI Makes People Less Careful, oversight must be designed in, not hoped for.
Use AI to free time for people work. Let AI take on stock management, scheduling, data entry and first drafts, then deliberately reinvest the time in customer conversations, advice and care. As AI Can Help Your Newest Staff Work Like Your Best showed, AI is often most powerful when it supports people.
Redesign the work, not just the tool. Before adding AI, map how the task flows and who does what. Decide which parts AI handles and where people add judgement, as explored in Fix the Process Before You Automate It.
Be open with customers. Tell people when they’re dealing with AI and how to reach a person, as explored in Always Leave a Door to a Human.
Invest in your people’s skills. The businesses that benefit most will have staff who know how to work alongside AI and who are confident in their own judgement.
[Infographic placement] AI and people, side by side. Two overlapping circles. AI does well: forecasting, sorting, drafting, flagging patterns, routine tasks. People do well: empathy, judgement, trust, explaining, handling the unexpected. Overlap: Better outcomes. Beneath: McKinsey (2025): 88% of organisations use AI, but about a third have scaled it; redesigning workflows is a key driver of impact.
A human-first AI plan
Before introducing your next AI tool, answer three questions in writing. What will the AI do? What will people do with the time it saves? Where must a person always check, decide or speak to the customer? Share the answers with your team, and revisit them after a month of use. This simple plan keeps technology in service of the business, rather than the other way round.
A few questions can help:
- Which of our tasks are routine enough for AI, and which depend on human judgement or care?
- Where must a person always review or decide?
- How will we reinvest the time AI saves?
- Do our customers know when they’re dealing with AI, and how to reach a person?
Which brings us back to the Pretoria pharmacy and its newly efficient stockroom. The AI handled the ordering and flagged the risks; the pharmacist made the final call and kept the human relationships that customers valued. As AI reshapes industry after industry, that partnership, technology for efficiency and people for trust, is likely to be what separates the businesses customers merely use from the ones they choose.
Sources
- McKinsey & Company, The State of AI in 2025: Agents, Innovation, and Transformation, November 2025.
- McKinsey & Company, The State of AI: How Organizations Are Rewiring to Capture Value, March 2025, for the 27% review figure.
