Appreciative Inquiry

by Andy Smith


The final phase is about

  • Planning and forming action groups to carry forward the actions identified during the design phase
  • (Possibly) developing detailed action plans
  • Celebrating the learnings identified so far
  • And, most importantly, actually doing it and learning in the process.

An AI summit process

This can be done with action groups over weeks or months. In an AI summit process the delivery phase could include the following activities (some activities may be truncated or carried out outside the summit, depending on the time available):

Generate action plans

Groups of eight, self-organising around a particular provocative proposition, work on the question: ‘What specific actions or changes to processes will bring the ideas to life?’

Feedback to group, and offers, commitments, requests

Each group feeds back their outcomes for their provocative propositions. Individuals must be encouraged to declare their commitment to act to make the new processes and actions happen. These individuals can also appeal for help, and people can commit to help them.

Task groups

The organiser forms task groups that take responsibility for each specific action. Each group meets to review the themes and provocative propositions for their actions. They create an initial action plan for their tasks. Plans should include how progress will be measured, and how the organisation will learn from the experience of carrying out the changes.


The final session of the summit reviews the event and acknowledges the progress made, and the energy and commitment of everyone involved. Everyone should be clear about what happens next.

Take action

The bulk of the delivery stage takes place outside the AI summit, as the groups and individuals who have assumed responsibility for implementing changes carry them out.

Once actions have been taken, the AI cycle can start again with discovering what has been achieved and what has worked well in the delivery stage.

‘Delivery’ or ‘destiny’?

Some AI practitioners prefer the term ‘destiny’ to ‘delivery’. David Cooperrider says:

In the early years of AI work, we called the fourth D delivery, not destiny. (In the 4D model, the definition stage is not included as it has happened before the start of the process.) We emphasised action planning, developing implementation strategies, and dealing with conventional challenges of sustainability. But the word delivery did not go far enough. What we discovered... was that the momentum for change and long-term sustainability increased the more we abandoned delivery ideas of action planning, monitoring progress, and building implementation strategies.

What we did instead, in several of the most exciting cases, was to focus on giving AI away to everyone and then stepping back and letting the transformation emerge. Our experience suggests that organisational change needs to look a lot more like an inspired movement than a neatly packaged and engineered product.

In a meta-analysis of 20 published AI case studies, Gervase Bushe found that in most of the cases that resulted in transformational change, the Delivery or Destiny phase was best described as ‘improvisational’ – letting go of central command, control and monitoring, and supporting people in putting AI into practice for themselves.

For example, the telecoms company GTE (now Verizon) trained thousands of employees in AI and encouraged them to make it happen. Their organisational change programme won the American Society for Training and Development award for best organisational change program in 1997, assessed on measures such as stock price changes, union-management relations and employee satisfaction surveys.

Where the destiny/delivery phase reverted to centrally-dictated action plans, Bushe found that the results were comparable to those of other, more conventional, organisational change efforts.

Having said that, it seems that the term ‘delivery’ is coming back in vogue among the AI community. Delivery is certainly an easier term for potential clients new to AI to understand.