Investing in your existing people is one of the highest-return moves you can make in AI adoption, yet many organisations underinvest dramatically. Effective reskilling goes beyond generic "Introduction to AI" courses. It means identifying the specific capabilities different roles will need, designing learning pathways that connect to real work, and creating space for people to practice new skills in low-stakes environments before applying them to their actual jobs. The best programmes combine structured learning with hands-on projects, pairing people with mentors who've already made the transition. They recognise that different people start from different places and learn at different speeds, and they measure success by changed behaviour and business impact rather than course completion rates. Common pitfalls include making training optional (so only the already-enthusiastic participate), making it too theoretical (so people can't connect it to their work), or treating it as a one-off event rather than an ongoing investment. The organisations that get results commit to sustained programmes over months and years, adjust based on feedback, and make clear that developing AI skills is a valued and rewarded part of everyone's role.