Proof over Promise Insights on Real World AI Adoption from 2025 MINDS Organizations 2026

Amplifying specialized capabilities through human-AI collaboration Leading MINDS organizations are using AI to enhance specialized expertise rather than replace it. Instead of confining AI to technical or data science teams, they are redesigning roles and workflows so experts across functions can use AI assistants at key decision points. These organizations are drawing on the full spectrum of AI technologies (from ML to generative and agentic systems) to extend human capability in complex, knowledge-intensive domains such as biotechnology, healthcare and finance. AI enables specialists to explore more options, simulate complex scenarios and scale insights that were previously unattainable (see Spotlight 4). This shift reflects a deeper maturity in AI adoption. Human expertise becomes the differentiator, and AI the force multiplier that unlocks new levels of precision, creativity and innovation. Instead of limiting AI to specialist teams, these organizations are redesigning roles, adding AI assistants for workers and embedding assistance into existing tools at key steps.SPOTLIGHT 3 Cambridge Industries embeds AI into daily work to empower mid-skill workers Cambridge Industries is taking a people-centred approach to AI adoption, empowering mid-skill workers involved in infrastructure management across African cities. Their approach is specifically aimed at workers who have traditionally lacked access to advanced engineering tools or decision-making roles. Cambridge Industries equipped workers with mobile-first AI applications designed for accessibility and scale even in areas with constrained connectivity, budgets and technical capacity. The approach allows workers to apply AI in routine tasks without adding complexity to their workflows. For road maintenance, inspectors capture routine images that are analysed by AI to detect surface damage, predict deterioration and generate heatmaps used to prioritize repairs, enabling real-time infrastructure monitoring with minimal technical overhead. In construction, autonomous drones monitor active sites, flagging safety risks and translating complex safety manuals into actionable site- specific guidance delivered via WhatsApp or dashboards. This transformation succeeds by bridging high-level engineering knowledge and everyday operational tasks. By embedding AI into the hands of local inspectors, technicians and site coordinators, it reaches the people who are closest to the challenge being addressed but who’re often excluded from AI innovation. Impact: Cambridge Industries cut emergency road repair costs by 40% within six months, issued 3,000 AI-generated safety alerts and reduced site safety incidents by 50%. These results highlight how embedding AI into frontline workflows, with accessible design and real-time insights, can drive operational efficiency, improve compliance and create safer, more resilient systems. SPOTLIGHT 4 Phagos combines AI and human biologics expertise to scale innovation in animal health and reduce antibiotic dependence Phagos is addressing one of agriculture’s most pressing challenges: antimicrobial resistance driven by widespread use of antibiotics in livestock. Traditional development of bacteriophage (phage)-based therapies, a promising alternative to antibiotics, has been slow and resource- intensive. Phagos developed an AI-powered platform that integrates generative AI and ML models to predict phage– bacteria interactions directly from genomic sequences and rank optimal therapeutic combinations, drastically reducing development time. The solution combines deep domain expertise held by humans with AI to address a challenge that has historically resisted scalable innovation. Expert microbiologists use the system daily to guide experimental design and interpret results, creating a feedback loop in which human expertise refines AI predictions and AI expands the scope of scientific exploration. Impact: Phage therapy development timelines have dropped from 2 years to 2 months, and testing cycles from days to minutes. A pilot programme has shown 100% clinical efficacy against E. coli in live chickens. Proof over Promise: Insights on Real-World AI Adoption from 2025 MINDS Organizations 15