Situational Leadership starts with a simple question: what does this person or team need from me for this task now? Hersey and Blanchard (1977) describe four broad responses: directing, coaching, supporting and delegating. Leaders choose the response by considering competence, confidence, commitment and risk. Readiness is task-specific. For example, a skilled colleague may still need close support when using a new AI tool or making a decision with significant consequences.
Emotional intelligence helps leaders make better choices about how to adapt. Goleman (1998) links emotional intelligence to self-awareness, self-regulation, motivation, empathy and social skill. In practice, this means noticing when pressure, frustration or overconfidence might affect your judgement. It also means listening for concerns about workload, fairness, job redesign or trust. Leaders should use the model as a conversation tool, not as a fixed label for people, because confidence and capability can change quickly (Graeff, 1997).
Timescales matter. If there is a safety, security or service risk, leaders may need to give clear direction, set boundaries and check progress frequently. If the team is still learning, coaching over several weeks may be more useful. As confidence, skill and controls improve, leaders can move towards support and delegation. Review dates help prevent leaders from staying too directive for too long or delegating before people are ready.
AI-powered analytics can help leaders understand patterns such as adoption levels, workload pressures, errors or exceptions. This can support better questions and better decisions. However, data can be incomplete, biased or out of date. Leaders should explain what the evidence shows, what it does not show and where human judgement is still needed. Data can inform a conversation, but it should not decide a person’s readiness, motivation or value (Cabinet Office and DSIT, 2025; NIST, 2023).
Negotiation is stronger when evidence is combined with emotional intelligence. Leaders should distinguish stated positions from underlying interests, listen for concerns and seek options that protect shared outcomes (Fisher, Ury and Patton, 2011).
Storytelling can then connect data to human meaning: why the change matters, what a better future looks like, what will remain human-led and how people will be supported. A credible story acknowledges uncertainty and trade-offs rather than presenting AI as inevitable.
Difficult conversations require the same balance of clarity and empathy. Scott (2002) argues for addressing the real issue, listening carefully and ensuring the conversation leads to action. A leader can state the observed behaviour or risk, explain its impact, ask for the other person’s perspective and agree specific next steps. Emotional intelligence helps manage defensiveness without avoiding accountability.
Digital collaboration should support, not weaken, this process: use suitable channels, inclusive meetings, concise written summaries, shared decision logs and clear ownership. Asynchronous tools can widen participation, while live discussion is better for ambiguity, emotion or conflict. Leaders should check understanding rather than assuming that a message has been read or accepted. Sensitive or complex matters should not be reduced to an automated message.
Incremental pilots are often preferable to a “big bang” rollout because they create evidence, learning and visible safeguards (CMI, 2026). During a pilot, leaders may direct on scope, data use and stop rules; coach people through new workflows; support experimentation and feedback; and delegate only when competence, confidence and controls are demonstrated.
Resilience means treating setbacks as information while remaining responsible for consequences. A positive and adaptable leader separates a failed test from personal failure, protects psychological safety and still addresses repeated unsafe behaviour. The leader actively requests feedback from users and affected stakeholders, changes course when necessary and ensures a named human owns the final decision to scale, modify, pause or stop.
Action Point
Choose one AI-enabled change you are leading or supporting. Assess the team’s task competence, confidence, concerns and decision risk. Decide whether your leadership response should be more directive, coaching, supportive or delegated. Set a review date, identify who owns the final human decision and plan one conversation that combines evidence with empathy.