Insight
Insight
The 12 Steps to Enterprise AI Recovery
14 Jun 2026
Matt Cull
Why the most successful AI transformations don’t start with AI
For more than eighty years, the Twelve Steps have provided a framework for recovery. Originally developed by Alcoholics Anonymous, they have since been adapted by organisations, communities and individuals facing every kind of complex change. Their enduring appeal lies in a simple truth: lasting transformation rarely begins with a solution. It begins with an honest acknowledgement that the way things are today is no longer sustainable.
Enterprise AI transformation is reaching a remarkably similar moment.
Over the last two years, organisations have invested billions in artificial intelligence. Generative AI has found its way into almost every boardroom agenda. Pilot projects have proliferated across marketing, finance, customer service and operations. Every major technology provider has introduced new copilots, assistants and intelligent agents, each promising to reshape the way organisations work.
Yet despite the excitement, many enterprise leaders are asking an uncomfortable question. Why, after so much investment and experimentation, does genuine transformation still feel out of reach?
The answer, in many cases, has very little to do with AI itself.
Across large organisations, the same underlying issues continue to surface. Critical business data remains fragmented across multiple systems. ERP platforms have evolved through years of acquisitions and regional variations. Finance teams continue to rely on manual reconciliations. Processes differ from one market to another, governance is inconsistent and departments optimise locally rather than collectively. Artificial intelligence hasn’t created these problems, but it has made them impossible to ignore.
At Vannda, we have seen this repeatedly through enterprise transformation programmes. Working alongside organisations such as WPP Media has reinforced a lesson that becomes clearer with every implementation: AI is rarely the beginning of transformation. More often, it is the catalyst that exposes whether the organisation has already done the difficult work of simplifying operations, modernising systems and creating the foundations required for intelligence to scale.
Perhaps what enterprise organisations need today is not another AI strategy.
Perhaps they need a recovery programme.
It is a programme we have begun to build. Not another system to bolt on, but a single governed layer that sits between artificial intelligence and the systems a finance team already relies upon. Automation introduced through it does not arrive all at once; it diffuses, spreading one process at a time as trust and evidence accumulate, while a person keeps every decision that matters. The twelve steps that follow are, in effect, how that layer earns its place.
Step One: Admit your operating model isn’t AI-ready
The first step in every recovery programme is admitting there is a problem that cannot be solved through determination alone. It requires people to stop pretending that small adjustments will be enough and instead acknowledge that meaningful change is necessary.
Many organisations are approaching AI from the opposite direction. They assume that deploying new tools will overcome longstanding operational challenges, when in reality those challenges often become more visible as AI is introduced. Intelligent systems depend on accurate data, consistent processes and clear governance. Where those foundations are weak, AI simply accelerates inconsistency rather than eliminating it.
This is why operational maturity has become such an important concept. An organisation can be digitally sophisticated while remaining operationally fragmented. It can deploy the latest AI technologies while still relying on manual workarounds, disconnected reporting and duplicated processes. Admitting that reality is not an admission of failure. It is the first meaningful step towards building an organisation that is genuinely ready for AI.
Steps Two, Three and Four: Believe change is possible, commit to it, and take an honest inventory
The next stages of recovery ask people to believe that change is possible, commit themselves to the process and honestly examine the behaviours that brought them to where they are today.
Those principles translate surprisingly well into enterprise transformation.
For many leadership teams, the greatest obstacle is not technology but inertia. Legacy processes have often existed for decades. Regional variations become accepted as inevitable. Manual interventions become embedded in everyday operations until they are no longer questioned. Over time, complexity becomes normal.
The organisations making the greatest progress are those willing to challenge those assumptions. Rather than asking how AI can automate existing processes, they begin by asking whether those processes should exist in their current form at all. They examine where data originates, how decisions are made, why duplicated systems continue to exist and where operational friction is preventing the organisation from moving at the speed the business now demands.
This is precisely the thinking that underpins successful ERP transformation. Replacing legacy platforms is rarely the objective in itself. The real objective is to establish a common operating model that creates consistency across finance, operations, people and technology. Modern enterprise platforms become valuable not because they introduce new functionality, but because they create a single, trusted foundation upon which automation and AI can be built with confidence.
Steps Five to Eight: Let go of what no longer serves you
The middle stages of the original recovery programme focus on letting go of behaviours that are no longer helpful, accepting support from others and recognising the wider impact those behaviours have had. Recovery is not simply about introducing something new; it is about having the courage to stop doing what no longer works.
Enterprise transformation demands exactly the same discipline.
One of the most common mistakes organisations make is attempting to automate unnecessary complexity. Every redundant approval process, duplicated report, disconnected workflow and legacy application adds friction that AI will faithfully reproduce unless it is first removed. Technology is exceptionally good at scaling whatever already exists. If inefficiency remains embedded within the operating model, AI simply allows that inefficiency to operate more quickly.
This is the discipline the layer is built around. It will not automate a process before that process has been simplified, and it advances each one only by evidenced, approved steps rather than in a single leap. Diffusing automation this way removes friction on the way in, instead of faithfully reproducing it.
The most successful transformation programmes therefore spend as much time simplifying as they do implementing. They remove unnecessary processes before introducing automation. They establish clear ownership of data before deploying intelligent agents. Most importantly, they bring together leaders from finance, operations, HR, technology and commercial functions to create a shared vision for how the organisation should operate. AI cannot become an enterprise capability if transformation remains confined to a single department.
This philosophy has shaped Vannda’s work with WPP Media. The transformation programme centred on Microsoft Dynamics 365 Finance was never intended to be a technology replacement exercise. It was designed to create a consistent global operating model, strengthen governance and establish the trusted data foundation required to support future innovation. In doing so, it demonstrated that ERP transformation has evolved from a back-office initiative into one of the most important enablers of enterprise AI.
Steps Nine to Twelve: Keep learning, keep improving and help others do the same
The final stages of recovery are perhaps the most important because they recognise that transformation is never complete. Progress depends upon continual reflection, regular course correction and a willingness to share what has been learned with others.
This is where many organisations still think about AI in the wrong way.
Transformation is too often viewed as a programme with a defined end point. There is a business case, a delivery phase, a go-live date and, eventually, attention shifts towards the next strategic priority. AI does not operate within those boundaries. The technology evolves continuously, organisational expectations continue to change and competitive advantage increasingly belongs to businesses that learn faster than those around them.
Success therefore depends less on implementing AI than on creating an organisation capable of adapting alongside it. Governance must evolve. Processes must continue to be refined. Data quality must be continually improved. Leaders must develop new ways of measuring success that extend beyond technical implementation to include adoption, behavioural change and business outcomes.
A programme built this way is never truly finished, and it is not meant to be. The layer keeps pace as the business changes, taking on a little more only once it has earned the confidence to do so, and always leaving people in charge of the decisions that matter. Improving over time is not a phase of the work; it is simply how the business now runs.
Perhaps most importantly, organisations need to cultivate a culture in which learning is shared rather than isolated. The most mature enterprises do not simply improve their own operations; they create ecosystems in which partners, clients and employees benefit from the knowledge they have accumulated. That is how transformation moves beyond technology and becomes part of organisational culture.
Recovery is a commitment, not a destination
The enduring power of the Twelve Steps lies in the recognition that recovery is not something people complete. It is a discipline that requires honesty, humility and a commitment to continual improvement. Success comes not from a single breakthrough but from consistently making better decisions over time.
Enterprise AI transformation deserves to be viewed through the same lens.
The organisations creating lasting value from AI will not necessarily be those announcing the most ambitious pilot programmes or adopting every new model that reaches the market. They will be the organisations prepared to confront uncomfortable truths about the way they operate, modernise the systems that underpin their business and continuously refine the relationship between people, processes and technology.
At Vannda, we believe this is why ERP transformation has become one of the defining strategic investments of the AI era. It is no longer simply about replacing legacy software. It is about creating the operational foundations that allow intelligence to scale safely, consistently and with measurable business value.
AI may be the catalyst that is reshaping enterprise organisations, but it is operational discipline that determines whether that change becomes transformational or merely another technology initiative. Like every successful recovery programme, the journey begins with an honest acknowledgement of where you are today – and a commitment to building something better tomorrow.
For us, building that better tomorrow is what the layer is for: a recovery programme made practical, so that intelligence can scale safely and consistently, with a person always keeping the gates that matter.
Why the most successful AI transformations don’t start with AI
For more than eighty years, the Twelve Steps have provided a framework for recovery. Originally developed by Alcoholics Anonymous, they have since been adapted by organisations, communities and individuals facing every kind of complex change. Their enduring appeal lies in a simple truth: lasting transformation rarely begins with a solution. It begins with an honest acknowledgement that the way things are today is no longer sustainable.
Enterprise AI transformation is reaching a remarkably similar moment.
Over the last two years, organisations have invested billions in artificial intelligence. Generative AI has found its way into almost every boardroom agenda. Pilot projects have proliferated across marketing, finance, customer service and operations. Every major technology provider has introduced new copilots, assistants and intelligent agents, each promising to reshape the way organisations work.
Yet despite the excitement, many enterprise leaders are asking an uncomfortable question. Why, after so much investment and experimentation, does genuine transformation still feel out of reach?
The answer, in many cases, has very little to do with AI itself.
Across large organisations, the same underlying issues continue to surface. Critical business data remains fragmented across multiple systems. ERP platforms have evolved through years of acquisitions and regional variations. Finance teams continue to rely on manual reconciliations. Processes differ from one market to another, governance is inconsistent and departments optimise locally rather than collectively. Artificial intelligence hasn’t created these problems, but it has made them impossible to ignore.
At Vannda, we have seen this repeatedly through enterprise transformation programmes. Working alongside organisations such as WPP Media has reinforced a lesson that becomes clearer with every implementation: AI is rarely the beginning of transformation. More often, it is the catalyst that exposes whether the organisation has already done the difficult work of simplifying operations, modernising systems and creating the foundations required for intelligence to scale.
Perhaps what enterprise organisations need today is not another AI strategy.
Perhaps they need a recovery programme.
It is a programme we have begun to build. Not another system to bolt on, but a single governed layer that sits between artificial intelligence and the systems a finance team already relies upon. Automation introduced through it does not arrive all at once; it diffuses, spreading one process at a time as trust and evidence accumulate, while a person keeps every decision that matters. The twelve steps that follow are, in effect, how that layer earns its place.
Step One: Admit your operating model isn’t AI-ready
The first step in every recovery programme is admitting there is a problem that cannot be solved through determination alone. It requires people to stop pretending that small adjustments will be enough and instead acknowledge that meaningful change is necessary.
Many organisations are approaching AI from the opposite direction. They assume that deploying new tools will overcome longstanding operational challenges, when in reality those challenges often become more visible as AI is introduced. Intelligent systems depend on accurate data, consistent processes and clear governance. Where those foundations are weak, AI simply accelerates inconsistency rather than eliminating it.
This is why operational maturity has become such an important concept. An organisation can be digitally sophisticated while remaining operationally fragmented. It can deploy the latest AI technologies while still relying on manual workarounds, disconnected reporting and duplicated processes. Admitting that reality is not an admission of failure. It is the first meaningful step towards building an organisation that is genuinely ready for AI.
Steps Two, Three and Four: Believe change is possible, commit to it, and take an honest inventory
The next stages of recovery ask people to believe that change is possible, commit themselves to the process and honestly examine the behaviours that brought them to where they are today.
Those principles translate surprisingly well into enterprise transformation.
For many leadership teams, the greatest obstacle is not technology but inertia. Legacy processes have often existed for decades. Regional variations become accepted as inevitable. Manual interventions become embedded in everyday operations until they are no longer questioned. Over time, complexity becomes normal.
The organisations making the greatest progress are those willing to challenge those assumptions. Rather than asking how AI can automate existing processes, they begin by asking whether those processes should exist in their current form at all. They examine where data originates, how decisions are made, why duplicated systems continue to exist and where operational friction is preventing the organisation from moving at the speed the business now demands.
This is precisely the thinking that underpins successful ERP transformation. Replacing legacy platforms is rarely the objective in itself. The real objective is to establish a common operating model that creates consistency across finance, operations, people and technology. Modern enterprise platforms become valuable not because they introduce new functionality, but because they create a single, trusted foundation upon which automation and AI can be built with confidence.
Steps Five to Eight: Let go of what no longer serves you
The middle stages of the original recovery programme focus on letting go of behaviours that are no longer helpful, accepting support from others and recognising the wider impact those behaviours have had. Recovery is not simply about introducing something new; it is about having the courage to stop doing what no longer works.
Enterprise transformation demands exactly the same discipline.
One of the most common mistakes organisations make is attempting to automate unnecessary complexity. Every redundant approval process, duplicated report, disconnected workflow and legacy application adds friction that AI will faithfully reproduce unless it is first removed. Technology is exceptionally good at scaling whatever already exists. If inefficiency remains embedded within the operating model, AI simply allows that inefficiency to operate more quickly.
This is the discipline the layer is built around. It will not automate a process before that process has been simplified, and it advances each one only by evidenced, approved steps rather than in a single leap. Diffusing automation this way removes friction on the way in, instead of faithfully reproducing it.
The most successful transformation programmes therefore spend as much time simplifying as they do implementing. They remove unnecessary processes before introducing automation. They establish clear ownership of data before deploying intelligent agents. Most importantly, they bring together leaders from finance, operations, HR, technology and commercial functions to create a shared vision for how the organisation should operate. AI cannot become an enterprise capability if transformation remains confined to a single department.
This philosophy has shaped Vannda’s work with WPP Media. The transformation programme centred on Microsoft Dynamics 365 Finance was never intended to be a technology replacement exercise. It was designed to create a consistent global operating model, strengthen governance and establish the trusted data foundation required to support future innovation. In doing so, it demonstrated that ERP transformation has evolved from a back-office initiative into one of the most important enablers of enterprise AI.
Steps Nine to Twelve: Keep learning, keep improving and help others do the same
The final stages of recovery are perhaps the most important because they recognise that transformation is never complete. Progress depends upon continual reflection, regular course correction and a willingness to share what has been learned with others.
This is where many organisations still think about AI in the wrong way.
Transformation is too often viewed as a programme with a defined end point. There is a business case, a delivery phase, a go-live date and, eventually, attention shifts towards the next strategic priority. AI does not operate within those boundaries. The technology evolves continuously, organisational expectations continue to change and competitive advantage increasingly belongs to businesses that learn faster than those around them.
Success therefore depends less on implementing AI than on creating an organisation capable of adapting alongside it. Governance must evolve. Processes must continue to be refined. Data quality must be continually improved. Leaders must develop new ways of measuring success that extend beyond technical implementation to include adoption, behavioural change and business outcomes.
A programme built this way is never truly finished, and it is not meant to be. The layer keeps pace as the business changes, taking on a little more only once it has earned the confidence to do so, and always leaving people in charge of the decisions that matter. Improving over time is not a phase of the work; it is simply how the business now runs.
Perhaps most importantly, organisations need to cultivate a culture in which learning is shared rather than isolated. The most mature enterprises do not simply improve their own operations; they create ecosystems in which partners, clients and employees benefit from the knowledge they have accumulated. That is how transformation moves beyond technology and becomes part of organisational culture.
Recovery is a commitment, not a destination
The enduring power of the Twelve Steps lies in the recognition that recovery is not something people complete. It is a discipline that requires honesty, humility and a commitment to continual improvement. Success comes not from a single breakthrough but from consistently making better decisions over time.
Enterprise AI transformation deserves to be viewed through the same lens.
The organisations creating lasting value from AI will not necessarily be those announcing the most ambitious pilot programmes or adopting every new model that reaches the market. They will be the organisations prepared to confront uncomfortable truths about the way they operate, modernise the systems that underpin their business and continuously refine the relationship between people, processes and technology.
At Vannda, we believe this is why ERP transformation has become one of the defining strategic investments of the AI era. It is no longer simply about replacing legacy software. It is about creating the operational foundations that allow intelligence to scale safely, consistently and with measurable business value.
AI may be the catalyst that is reshaping enterprise organisations, but it is operational discipline that determines whether that change becomes transformational or merely another technology initiative. Like every successful recovery programme, the journey begins with an honest acknowledgement of where you are today – and a commitment to building something better tomorrow.
For us, building that better tomorrow is what the layer is for: a recovery programme made practical, so that intelligence can scale safely and consistently, with a person always keeping the gates that matter.