AI Can Reshape the Transformation Office, but Accountability Must Remain Human

Boston Consulting Group described on July 27 a transformation office that uses AI to manage large-scale change. According to the authors, traditional teams spend much of their time collecting reports, checking deadlines, preparing summaries, and chasing managers. A system able to identify deviations independently can perform these tasks continuously and alert a person before a delay or budget variance turns into a loss. BCG divides the work of transformation offices into three areas: programme management, tracking financial benefits, and change management. By 2030, systems in the first area are expected to monitor risks, dependencies, and unresolved decisions continuously. In the second, they should compare baselines with actual results. In the third, they may detect resistance, delayed training, or weak adoption of a new process and propose a targeted intervention. The authors reject the idea that automation should assume responsibility for change. “AI will not replace transformation-office teams,” BCG says. It should free their time for discussions with workstream leaders, dispute resolution, and reprioritisation. Research conducted with Harvard Business Review, according to the article, shows that using AI merely as a substitute for people can weaken accountability, reduce the quality of control, and damage trust. Operational value depends on data quality. A system cannot monitor a budget reliably if departments use different definitions of savings or report project status only once a month. Before deployment, the company must therefore standardise indicators, data ownership, and change rules. Automated alerts should be traceable: managers need to know which data produced them, how severe the system judged them to be, and who decided on the next step. The legal and organisational impact becomes visible when tasks are delegated. A system may draft a corrective request, change the order of controls, or recommend moving funds, but approval authority must follow internal rules. The company needs a list of permitted interventions, financial limits, and a record of every decision. Without these rules, faster management may create a dispute over whether an instruction came from an authorised person or an unverified tool. BCG expects continuous monitoring to reduce value leakage and speed up decisions. The metric should not, however, be the number of automated alerts. Management should track the time needed to remove an obstacle, the accuracy of financial forecasts, the share of timely decisions, and value actually delivered. Too many alerts can instead cause managers to ignore them. The system therefore requires regular review of both false alarms and missed cases. “The strongest offices will create learning loops between the stages of change,” the authors write. In practice, this means recording which intervention worked, in what type of project, and under what conditions. The company can use this to improve the next programme, but must also prevent a historical solution from being applied automatically in a different context. The decisive advantage remains an experienced person who can place data in its commercial and workforce context. Terms explained Transformation office: A function coordinating large-scale change and tracking its finances, deadlines, risks, and outcomes. Workstream: A distinct part of a large programme with defined tasks, an owner, deadlines, and an expected benefit. Value leakage: The gap between the planned and the actually achieved financial benefit of a transformation. Learning loop: A process in which an organisation evaluates an intervention and uses the lesson in later decisions.

AI Can Reshape the Transformation Office, but Accountability Must Remain Human

The full article is available to users registered at Hard Skills