AI Transformation
Why the CFO Is Becoming the Architect of Enterprise AI Transformation
The transformation is owned by the business, not by IT.
Dan Mintz
·
The CFO is becoming the architect of enterprise AI transformation
At a glance
Sixty-two percent of CFOs surveyed by the IBM Institute for Business Value say their role has already expanded into enterprise technology or AI strategy leadership.
Only 6% describe their finance organization as transformation-ready, with AI embedded in workflows and decisions at scale.
The gap is a sequencing problem: companies start with AI tools before fixing data, governance and workflow design.
IBM's quote-to-cash redesign reached 90% touchless automation, a 60% productivity improvement and a 54% increase in the velocity of cash conversion.
The transformation is owned by the business, not by IT. The CFO is the natural owner of the investment thesis that connects the two.
By Dan Mintz. 3x founder, Wharton MBA, MIT MS in Machine Learning.
Leading enterprise AI transformation expert with experience in many projects, operating in the intersection of business strategy and AI technologies to drive impactful results.
The blueprint: business strategy should drive the AI transformation.
Intro
AI is changing what a CFO is responsible for. The job used to center on controllership, risk, fiduciary duty and cash management. A new study from the IBM Institute for Business Value, conducted with Oxford Economics, shows the remit moving into enterprise technology, capital reallocation and business-model design.
The survey covered 1,500 CFOs across 33 geographies and 26 industries in the first half of 2026. Its headline numbers tell two stories at once. Most CFOs are being pulled into AI strategy. Almost none of their finance organizations are ready for it.
For anyone leading or influencing an AI transformation at a large company, this matters. The person who controls the investment thesis, the capital and the measurement of value is now also being asked to shape the operating model. This article explains what the data shows, why so few finance organizations are ready, and what a business-led approach looks like in practice.
How much has the CFO role actually changed?
The CFO role has changed materially, and the survey data is specific. Sixty-two percent of CFOs say their role has already expanded into enterprise technology or AI strategy leadership. More than half report greater portfolio-management and capital-reallocation authority. More than half also report more responsibility for business-model or growth-strategy design.
The expectations for 2030 go further:
56% expect greater responsibility for financial and ethical guardrails around AI.
55% expect to help shape operating models, workforce strategies and organizational structure.
52% expect a bigger role in enterprise value creation and portfolio strategy.
IBM's own CFO, Jim Kavanaugh, describes the shift as moving from "guardian of stability" to "agent of transformation." He is clear that this does not replace other executives. CEOs set direction, technology leaders enable capabilities, COOs redesign workflows and CHROs shape talent and culture. The CFO connects those decisions to an investment thesis, an operating model and measurable value.
That framing is useful for decision makers. AI transformation is a cross-functional program. Someone has to hold the economics together, and finance is structurally placed to do it.
Why are only 6% of finance organizations transformation-ready?
Because most companies begin with the technology instead of the work. The study found that just 6% of CFOs describe their finance organization as transformation-ready, meaning AI is consistently embedded in workflows and decision-making at scale.
Kavanaugh calls this a sequencing problem. He identifies three prerequisites, in order:
Sound data architecture and governance.
Business-model and workflow redesign.
Technology deployment.
Most failed programs reverse this order. A team selects a model or a platform, runs a pilot on a convenient use case, and then discovers that the data is fragmented, the process is undefined and nobody owns the outcome. The pilot works in a demo and stalls in production.
I see the same pattern across enterprise AI programs. The technology is rarely the constraint. The constraint is that the workflow was never examined, so there is nothing stable for the AI to sit inside. A strong model on top of a broken process produces a faster broken process.
What does a business-first redesign look like in practice?
It starts by mapping how the work is really done, then redesigning it end to end before any automation is built. IBM's quote-to-cash process is the clearest example in the study.
Quote-to-cash spans customer quotes, contracting, billing, collections and cash receipt. When IBM examined it, the company found 364 different iterations across job roles and functions. The same process was being performed 364 different ways.
IBM did not automate those variants. It redesigned the process as a single end-to-end, cross-functional workflow, and then used AI to automate many of the steps. The reported results:
90% touchless automation.
A 60% productivity improvement.
A 54% increase in the velocity of cash conversion.
IBM also reports $4.5 billion in productivity generated over the past three years, with a target of $5.5 billion in 2026. The company frames those gains as capacity to invest in growth, not only as cost reduction.
The lesson is about order of operations. The value came from collapsing 364 variants into one designed process. The AI made that design fast and scalable, but the design came first.
Why does cash conversion make such a strong first workflow?
Cash conversion is measurable, cross-functional and tied directly to the balance sheet. That combination makes it a strong candidate for an early, high-credibility transformation effort.
A workflow like quote-to-cash or collections has properties that suit AI well:
High volume of repetitive, rules-based decisions.
Clear outcome metrics, such as days sales outstanding and cash velocity.
Many handoffs between systems and teams, which is where delay and error accumulate.
A natural executive sponsor in the CFO.
In my own work on treasury and cash collection, including deduction dispute handling, the same logic applies. The value sits in the handoffs between agents and humans, in clear escalation rules and in the quality of the underlying data. The model choice is the smaller decision. The workflow design and data readiness are the larger ones.
This is also why the CFO's expanding remit matters. A transformation that touches billing, collections, contracts and cash crosses several functions. It needs someone with authority over all of them and with a financial lens on what counts as success.
What do AI-first CFOs do differently?
According to IBM, they lead across five disciplines: shaping advantage, governance, intelligence, capital motion and building optionality. The study reports that companies led by AI-first CFOs achieved revenue growth rates 23% higher than peer organizations from 2022 to 2024.
That figure should be read carefully. It shows correlation between a leadership profile and financial performance. It does not prove that appointing an AI-focused CFO causes growth. What it does suggest is that companies treating AI as a capital allocation and operating model question tend to outperform those treating it as a tooling question.
Three of the five disciplines map directly to decisions a transformation leader faces:
Governance. Who is accountable when an agent makes a wrong call, and what guardrails are financial, ethical and regulatory. 56% of CFOs expect to own more of this by 2030.
Capital motion. How investment moves between use cases as evidence comes in, instead of being locked into a single large bet.
Building optionality. Choosing architectures and vendors that keep future options open as models and costs change quickly.
Each of these is a business decision expressed through a technical choice. That is the intersection where transformation succeeds or fails.
Is AI a technology project or a business transformation?
It is a business transformation. Kavanaugh's view is that the defining mistake is treating AI as a technology project with a people component. The correct framing is a business transformation that combines human and digital work to create sustainable enterprise value.
This has practical consequences for how a program is structured:
Ownership sits with the business. A workflow owner, backed by the CFO or another executive, is accountable for the outcome. IT is a delivery partner.
Success is defined in business terms. Cash velocity, cost to serve, cycle time and error rates, not model accuracy in isolation.
The redesign precedes the build. The target workflow is agreed before architecture decisions are made.
Data work is funded as part of the program. It is not a hidden prerequisite that delays the pilot.
Human roles are designed on purpose. Where people review, approve or handle exceptions is a design decision, not an afterthought.
This is the blueprint I apply: business strategy should drive the AI transformation. Technology choices follow from the workflow, the data and the economics. When that order is respected, the architecture becomes easier to decide, because the requirements are clear.
What does this mean for the wider workforce?
The transformation reaches well beyond finance. McKinsey Global Institute's "Workforce in Motion" report finds that AI, automation and broader economic shifts could force the largest and most sustained workforce transformation in U.S. history. The economy could create more jobs than automation replaces over the next decade. Even so, more than 70% of workers could need some level of reinvention, because new tasks may take more than 15% of their time.
McKinsey also reports that demand for AI fluency has risen 11-fold since 2022. For a CFO who is now expected to help shape workforce strategy and organizational structure, this is a planning input, not background reading. Reskilling budgets, role redesign and hiring plans all belong in the same investment case as the technology spend.
How should a decision maker start?
Start with one workflow that matters financially, redesign it, and measure it. A practical sequence:
Pick a cross-functional workflow with clear financial metrics, such as quote-to-cash, collections or procurement.
Map the real process, including the variants. If IBM found 364 versions in one process, assume yours has many too.
Assess data readiness and governance before choosing tools.
Redesign the workflow end to end, then decide where agents act and where humans decide.
Pilot with a baseline, a target and an owner. Scale only what the numbers support.
The 6% figure is the opportunity as much as the warning. Most competitors are not ready either. The organizations that fix sequencing first will move faster than the ones still buying tools.
Closing
The IBM data makes the direction clear. CFOs are being asked to co-architect AI strategy, and most finance organizations are not yet built for it. The gap is not about access to technology. It is about starting with the wrong thing.
Fix the data, redesign the work, then deploy the technology. That order is what turned 364 process variants into 90% touchless automation at IBM. It is also the order that separates transformation programs that scale from those that stall in pilot.
FAQ
1. What is an "AI-first CFO"?
In IBM's study, it is a CFO who leads across five disciplines: shaping advantage, governance, intelligence, capital motion and building optionality. Companies led by these CFOs reported revenue growth 23% higher than peers from 2022 to 2024.
2. What does "transformation-ready" mean for a finance organization?
It means AI is consistently embedded in workflows and decision-making at scale. Only 6% of CFOs in the IBM survey describe their finance function this way.
3. What is quote-to-cash?
It is the process spanning customer quotes, contracting, billing, collections and cash receipt. IBM redesigned it as one end-to-end workflow and reported 90% touchless automation.
4. Why do so many finance AI programs fail?
Kavanaugh points to sequencing. Companies start with AI technology before establishing data architecture and governance, and before redesigning the underlying workflow.
5. Why should business strategy drive an AI transformation?
Because value is created by changing how work and decisions happen, not by installing technology. Strategy defines which workflow matters, what success looks like and how much to invest, and architecture follows from those answers.
6. How does your background shape the way you approach these programs?
I have built and exited ventures, and I hold an MBA from Wharton and an MS in Machine Learning from MIT. That lets me work on the economics and the architecture in the same conversation, which is where most enterprise programs lose time.
7. Where should a large company begin?
With one cross-functional workflow that has clear financial metrics. Map the real process, assess data readiness, redesign end to end, then pilot with a baseline and an owner.
Sources
Keep reading
Previous article
Building an AI-Native Sales Org on the Stack You Already Own: Enterprise AI Transformation Lessons From Anthropic
Next article
Next article
Gemini 4 Argon and Enterprise AI Agents: Why the Model Is No Longer the Bottleneck
Gemini 4 Argon and Enterprise AI Agents: Why the Model Is No Longer the Bottleneck
AI Transformation
Why the CFO Is Becoming the Architect of Enterprise AI Transformation
The transformation is owned by the business, not by IT.
Dan Mintz
·
The CFO is becoming the architect of enterprise AI transformation
At a glance
Sixty-two percent of CFOs surveyed by the IBM Institute for Business Value say their role has already expanded into enterprise technology or AI strategy leadership.
Only 6% describe their finance organization as transformation-ready, with AI embedded in workflows and decisions at scale.
The gap is a sequencing problem: companies start with AI tools before fixing data, governance and workflow design.
IBM's quote-to-cash redesign reached 90% touchless automation, a 60% productivity improvement and a 54% increase in the velocity of cash conversion.
The transformation is owned by the business, not by IT. The CFO is the natural owner of the investment thesis that connects the two.
By Dan Mintz. 3x founder, Wharton MBA, MIT MS in Machine Learning.
Leading enterprise AI transformation expert with experience in many projects, operating in the intersection of business strategy and AI technologies to drive impactful results.
The blueprint: business strategy should drive the AI transformation.
Intro
AI is changing what a CFO is responsible for. The job used to center on controllership, risk, fiduciary duty and cash management. A new study from the IBM Institute for Business Value, conducted with Oxford Economics, shows the remit moving into enterprise technology, capital reallocation and business-model design.
The survey covered 1,500 CFOs across 33 geographies and 26 industries in the first half of 2026. Its headline numbers tell two stories at once. Most CFOs are being pulled into AI strategy. Almost none of their finance organizations are ready for it.
For anyone leading or influencing an AI transformation at a large company, this matters. The person who controls the investment thesis, the capital and the measurement of value is now also being asked to shape the operating model. This article explains what the data shows, why so few finance organizations are ready, and what a business-led approach looks like in practice.
How much has the CFO role actually changed?
The CFO role has changed materially, and the survey data is specific. Sixty-two percent of CFOs say their role has already expanded into enterprise technology or AI strategy leadership. More than half report greater portfolio-management and capital-reallocation authority. More than half also report more responsibility for business-model or growth-strategy design.
The expectations for 2030 go further:
56% expect greater responsibility for financial and ethical guardrails around AI.
55% expect to help shape operating models, workforce strategies and organizational structure.
52% expect a bigger role in enterprise value creation and portfolio strategy.
IBM's own CFO, Jim Kavanaugh, describes the shift as moving from "guardian of stability" to "agent of transformation." He is clear that this does not replace other executives. CEOs set direction, technology leaders enable capabilities, COOs redesign workflows and CHROs shape talent and culture. The CFO connects those decisions to an investment thesis, an operating model and measurable value.
That framing is useful for decision makers. AI transformation is a cross-functional program. Someone has to hold the economics together, and finance is structurally placed to do it.
Why are only 6% of finance organizations transformation-ready?
Because most companies begin with the technology instead of the work. The study found that just 6% of CFOs describe their finance organization as transformation-ready, meaning AI is consistently embedded in workflows and decision-making at scale.
Kavanaugh calls this a sequencing problem. He identifies three prerequisites, in order:
Sound data architecture and governance.
Business-model and workflow redesign.
Technology deployment.
Most failed programs reverse this order. A team selects a model or a platform, runs a pilot on a convenient use case, and then discovers that the data is fragmented, the process is undefined and nobody owns the outcome. The pilot works in a demo and stalls in production.
I see the same pattern across enterprise AI programs. The technology is rarely the constraint. The constraint is that the workflow was never examined, so there is nothing stable for the AI to sit inside. A strong model on top of a broken process produces a faster broken process.
What does a business-first redesign look like in practice?
It starts by mapping how the work is really done, then redesigning it end to end before any automation is built. IBM's quote-to-cash process is the clearest example in the study.
Quote-to-cash spans customer quotes, contracting, billing, collections and cash receipt. When IBM examined it, the company found 364 different iterations across job roles and functions. The same process was being performed 364 different ways.
IBM did not automate those variants. It redesigned the process as a single end-to-end, cross-functional workflow, and then used AI to automate many of the steps. The reported results:
90% touchless automation.
A 60% productivity improvement.
A 54% increase in the velocity of cash conversion.
IBM also reports $4.5 billion in productivity generated over the past three years, with a target of $5.5 billion in 2026. The company frames those gains as capacity to invest in growth, not only as cost reduction.
The lesson is about order of operations. The value came from collapsing 364 variants into one designed process. The AI made that design fast and scalable, but the design came first.
Why does cash conversion make such a strong first workflow?
Cash conversion is measurable, cross-functional and tied directly to the balance sheet. That combination makes it a strong candidate for an early, high-credibility transformation effort.
A workflow like quote-to-cash or collections has properties that suit AI well:
High volume of repetitive, rules-based decisions.
Clear outcome metrics, such as days sales outstanding and cash velocity.
Many handoffs between systems and teams, which is where delay and error accumulate.
A natural executive sponsor in the CFO.
In my own work on treasury and cash collection, including deduction dispute handling, the same logic applies. The value sits in the handoffs between agents and humans, in clear escalation rules and in the quality of the underlying data. The model choice is the smaller decision. The workflow design and data readiness are the larger ones.
This is also why the CFO's expanding remit matters. A transformation that touches billing, collections, contracts and cash crosses several functions. It needs someone with authority over all of them and with a financial lens on what counts as success.
What do AI-first CFOs do differently?
According to IBM, they lead across five disciplines: shaping advantage, governance, intelligence, capital motion and building optionality. The study reports that companies led by AI-first CFOs achieved revenue growth rates 23% higher than peer organizations from 2022 to 2024.
That figure should be read carefully. It shows correlation between a leadership profile and financial performance. It does not prove that appointing an AI-focused CFO causes growth. What it does suggest is that companies treating AI as a capital allocation and operating model question tend to outperform those treating it as a tooling question.
Three of the five disciplines map directly to decisions a transformation leader faces:
Governance. Who is accountable when an agent makes a wrong call, and what guardrails are financial, ethical and regulatory. 56% of CFOs expect to own more of this by 2030.
Capital motion. How investment moves between use cases as evidence comes in, instead of being locked into a single large bet.
Building optionality. Choosing architectures and vendors that keep future options open as models and costs change quickly.
Each of these is a business decision expressed through a technical choice. That is the intersection where transformation succeeds or fails.
Is AI a technology project or a business transformation?
It is a business transformation. Kavanaugh's view is that the defining mistake is treating AI as a technology project with a people component. The correct framing is a business transformation that combines human and digital work to create sustainable enterprise value.
This has practical consequences for how a program is structured:
Ownership sits with the business. A workflow owner, backed by the CFO or another executive, is accountable for the outcome. IT is a delivery partner.
Success is defined in business terms. Cash velocity, cost to serve, cycle time and error rates, not model accuracy in isolation.
The redesign precedes the build. The target workflow is agreed before architecture decisions are made.
Data work is funded as part of the program. It is not a hidden prerequisite that delays the pilot.
Human roles are designed on purpose. Where people review, approve or handle exceptions is a design decision, not an afterthought.
This is the blueprint I apply: business strategy should drive the AI transformation. Technology choices follow from the workflow, the data and the economics. When that order is respected, the architecture becomes easier to decide, because the requirements are clear.
What does this mean for the wider workforce?
The transformation reaches well beyond finance. McKinsey Global Institute's "Workforce in Motion" report finds that AI, automation and broader economic shifts could force the largest and most sustained workforce transformation in U.S. history. The economy could create more jobs than automation replaces over the next decade. Even so, more than 70% of workers could need some level of reinvention, because new tasks may take more than 15% of their time.
McKinsey also reports that demand for AI fluency has risen 11-fold since 2022. For a CFO who is now expected to help shape workforce strategy and organizational structure, this is a planning input, not background reading. Reskilling budgets, role redesign and hiring plans all belong in the same investment case as the technology spend.
How should a decision maker start?
Start with one workflow that matters financially, redesign it, and measure it. A practical sequence:
Pick a cross-functional workflow with clear financial metrics, such as quote-to-cash, collections or procurement.
Map the real process, including the variants. If IBM found 364 versions in one process, assume yours has many too.
Assess data readiness and governance before choosing tools.
Redesign the workflow end to end, then decide where agents act and where humans decide.
Pilot with a baseline, a target and an owner. Scale only what the numbers support.
The 6% figure is the opportunity as much as the warning. Most competitors are not ready either. The organizations that fix sequencing first will move faster than the ones still buying tools.
Closing
The IBM data makes the direction clear. CFOs are being asked to co-architect AI strategy, and most finance organizations are not yet built for it. The gap is not about access to technology. It is about starting with the wrong thing.
Fix the data, redesign the work, then deploy the technology. That order is what turned 364 process variants into 90% touchless automation at IBM. It is also the order that separates transformation programs that scale from those that stall in pilot.
FAQ
1. What is an "AI-first CFO"?
In IBM's study, it is a CFO who leads across five disciplines: shaping advantage, governance, intelligence, capital motion and building optionality. Companies led by these CFOs reported revenue growth 23% higher than peers from 2022 to 2024.
2. What does "transformation-ready" mean for a finance organization?
It means AI is consistently embedded in workflows and decision-making at scale. Only 6% of CFOs in the IBM survey describe their finance function this way.
3. What is quote-to-cash?
It is the process spanning customer quotes, contracting, billing, collections and cash receipt. IBM redesigned it as one end-to-end workflow and reported 90% touchless automation.
4. Why do so many finance AI programs fail?
Kavanaugh points to sequencing. Companies start with AI technology before establishing data architecture and governance, and before redesigning the underlying workflow.
5. Why should business strategy drive an AI transformation?
Because value is created by changing how work and decisions happen, not by installing technology. Strategy defines which workflow matters, what success looks like and how much to invest, and architecture follows from those answers.
6. How does your background shape the way you approach these programs?
I have built and exited ventures, and I hold an MBA from Wharton and an MS in Machine Learning from MIT. That lets me work on the economics and the architecture in the same conversation, which is where most enterprise programs lose time.
7. Where should a large company begin?
With one cross-functional workflow that has clear financial metrics. Map the real process, assess data readiness, redesign end to end, then pilot with a baseline and an owner.
Sources
Keep reading
Previous article
Building an AI-Native Sales Org on the Stack You Already Own: Enterprise AI Transformation Lessons From Anthropic
Next article
Next article
Gemini 4 Argon and Enterprise AI Agents: Why the Model Is No Longer the Bottleneck
Gemini 4 Argon and Enterprise AI Agents: Why the Model Is No Longer the Bottleneck
AI Transformation
Why the CFO Is Becoming the Architect of Enterprise AI Transformation
The transformation is owned by the business, not by IT.
Dan Mintz
·
The CFO is becoming the architect of enterprise AI transformation
At a glance
Sixty-two percent of CFOs surveyed by the IBM Institute for Business Value say their role has already expanded into enterprise technology or AI strategy leadership.
Only 6% describe their finance organization as transformation-ready, with AI embedded in workflows and decisions at scale.
The gap is a sequencing problem: companies start with AI tools before fixing data, governance and workflow design.
IBM's quote-to-cash redesign reached 90% touchless automation, a 60% productivity improvement and a 54% increase in the velocity of cash conversion.
The transformation is owned by the business, not by IT. The CFO is the natural owner of the investment thesis that connects the two.
By Dan Mintz. 3x founder, Wharton MBA, MIT MS in Machine Learning.
Leading enterprise AI transformation expert with experience in many projects, operating in the intersection of business strategy and AI technologies to drive impactful results.
The blueprint: business strategy should drive the AI transformation.
Intro
AI is changing what a CFO is responsible for. The job used to center on controllership, risk, fiduciary duty and cash management. A new study from the IBM Institute for Business Value, conducted with Oxford Economics, shows the remit moving into enterprise technology, capital reallocation and business-model design.
The survey covered 1,500 CFOs across 33 geographies and 26 industries in the first half of 2026. Its headline numbers tell two stories at once. Most CFOs are being pulled into AI strategy. Almost none of their finance organizations are ready for it.
For anyone leading or influencing an AI transformation at a large company, this matters. The person who controls the investment thesis, the capital and the measurement of value is now also being asked to shape the operating model. This article explains what the data shows, why so few finance organizations are ready, and what a business-led approach looks like in practice.
How much has the CFO role actually changed?
The CFO role has changed materially, and the survey data is specific. Sixty-two percent of CFOs say their role has already expanded into enterprise technology or AI strategy leadership. More than half report greater portfolio-management and capital-reallocation authority. More than half also report more responsibility for business-model or growth-strategy design.
The expectations for 2030 go further:
56% expect greater responsibility for financial and ethical guardrails around AI.
55% expect to help shape operating models, workforce strategies and organizational structure.
52% expect a bigger role in enterprise value creation and portfolio strategy.
IBM's own CFO, Jim Kavanaugh, describes the shift as moving from "guardian of stability" to "agent of transformation." He is clear that this does not replace other executives. CEOs set direction, technology leaders enable capabilities, COOs redesign workflows and CHROs shape talent and culture. The CFO connects those decisions to an investment thesis, an operating model and measurable value.
That framing is useful for decision makers. AI transformation is a cross-functional program. Someone has to hold the economics together, and finance is structurally placed to do it.
Why are only 6% of finance organizations transformation-ready?
Because most companies begin with the technology instead of the work. The study found that just 6% of CFOs describe their finance organization as transformation-ready, meaning AI is consistently embedded in workflows and decision-making at scale.
Kavanaugh calls this a sequencing problem. He identifies three prerequisites, in order:
Sound data architecture and governance.
Business-model and workflow redesign.
Technology deployment.
Most failed programs reverse this order. A team selects a model or a platform, runs a pilot on a convenient use case, and then discovers that the data is fragmented, the process is undefined and nobody owns the outcome. The pilot works in a demo and stalls in production.
I see the same pattern across enterprise AI programs. The technology is rarely the constraint. The constraint is that the workflow was never examined, so there is nothing stable for the AI to sit inside. A strong model on top of a broken process produces a faster broken process.
What does a business-first redesign look like in practice?
It starts by mapping how the work is really done, then redesigning it end to end before any automation is built. IBM's quote-to-cash process is the clearest example in the study.
Quote-to-cash spans customer quotes, contracting, billing, collections and cash receipt. When IBM examined it, the company found 364 different iterations across job roles and functions. The same process was being performed 364 different ways.
IBM did not automate those variants. It redesigned the process as a single end-to-end, cross-functional workflow, and then used AI to automate many of the steps. The reported results:
90% touchless automation.
A 60% productivity improvement.
A 54% increase in the velocity of cash conversion.
IBM also reports $4.5 billion in productivity generated over the past three years, with a target of $5.5 billion in 2026. The company frames those gains as capacity to invest in growth, not only as cost reduction.
The lesson is about order of operations. The value came from collapsing 364 variants into one designed process. The AI made that design fast and scalable, but the design came first.
Why does cash conversion make such a strong first workflow?
Cash conversion is measurable, cross-functional and tied directly to the balance sheet. That combination makes it a strong candidate for an early, high-credibility transformation effort.
A workflow like quote-to-cash or collections has properties that suit AI well:
High volume of repetitive, rules-based decisions.
Clear outcome metrics, such as days sales outstanding and cash velocity.
Many handoffs between systems and teams, which is where delay and error accumulate.
A natural executive sponsor in the CFO.
In my own work on treasury and cash collection, including deduction dispute handling, the same logic applies. The value sits in the handoffs between agents and humans, in clear escalation rules and in the quality of the underlying data. The model choice is the smaller decision. The workflow design and data readiness are the larger ones.
This is also why the CFO's expanding remit matters. A transformation that touches billing, collections, contracts and cash crosses several functions. It needs someone with authority over all of them and with a financial lens on what counts as success.
What do AI-first CFOs do differently?
According to IBM, they lead across five disciplines: shaping advantage, governance, intelligence, capital motion and building optionality. The study reports that companies led by AI-first CFOs achieved revenue growth rates 23% higher than peer organizations from 2022 to 2024.
That figure should be read carefully. It shows correlation between a leadership profile and financial performance. It does not prove that appointing an AI-focused CFO causes growth. What it does suggest is that companies treating AI as a capital allocation and operating model question tend to outperform those treating it as a tooling question.
Three of the five disciplines map directly to decisions a transformation leader faces:
Governance. Who is accountable when an agent makes a wrong call, and what guardrails are financial, ethical and regulatory. 56% of CFOs expect to own more of this by 2030.
Capital motion. How investment moves between use cases as evidence comes in, instead of being locked into a single large bet.
Building optionality. Choosing architectures and vendors that keep future options open as models and costs change quickly.
Each of these is a business decision expressed through a technical choice. That is the intersection where transformation succeeds or fails.
Is AI a technology project or a business transformation?
It is a business transformation. Kavanaugh's view is that the defining mistake is treating AI as a technology project with a people component. The correct framing is a business transformation that combines human and digital work to create sustainable enterprise value.
This has practical consequences for how a program is structured:
Ownership sits with the business. A workflow owner, backed by the CFO or another executive, is accountable for the outcome. IT is a delivery partner.
Success is defined in business terms. Cash velocity, cost to serve, cycle time and error rates, not model accuracy in isolation.
The redesign precedes the build. The target workflow is agreed before architecture decisions are made.
Data work is funded as part of the program. It is not a hidden prerequisite that delays the pilot.
Human roles are designed on purpose. Where people review, approve or handle exceptions is a design decision, not an afterthought.
This is the blueprint I apply: business strategy should drive the AI transformation. Technology choices follow from the workflow, the data and the economics. When that order is respected, the architecture becomes easier to decide, because the requirements are clear.
What does this mean for the wider workforce?
The transformation reaches well beyond finance. McKinsey Global Institute's "Workforce in Motion" report finds that AI, automation and broader economic shifts could force the largest and most sustained workforce transformation in U.S. history. The economy could create more jobs than automation replaces over the next decade. Even so, more than 70% of workers could need some level of reinvention, because new tasks may take more than 15% of their time.
McKinsey also reports that demand for AI fluency has risen 11-fold since 2022. For a CFO who is now expected to help shape workforce strategy and organizational structure, this is a planning input, not background reading. Reskilling budgets, role redesign and hiring plans all belong in the same investment case as the technology spend.
How should a decision maker start?
Start with one workflow that matters financially, redesign it, and measure it. A practical sequence:
Pick a cross-functional workflow with clear financial metrics, such as quote-to-cash, collections or procurement.
Map the real process, including the variants. If IBM found 364 versions in one process, assume yours has many too.
Assess data readiness and governance before choosing tools.
Redesign the workflow end to end, then decide where agents act and where humans decide.
Pilot with a baseline, a target and an owner. Scale only what the numbers support.
The 6% figure is the opportunity as much as the warning. Most competitors are not ready either. The organizations that fix sequencing first will move faster than the ones still buying tools.
Closing
The IBM data makes the direction clear. CFOs are being asked to co-architect AI strategy, and most finance organizations are not yet built for it. The gap is not about access to technology. It is about starting with the wrong thing.
Fix the data, redesign the work, then deploy the technology. That order is what turned 364 process variants into 90% touchless automation at IBM. It is also the order that separates transformation programs that scale from those that stall in pilot.
FAQ
1. What is an "AI-first CFO"?
In IBM's study, it is a CFO who leads across five disciplines: shaping advantage, governance, intelligence, capital motion and building optionality. Companies led by these CFOs reported revenue growth 23% higher than peers from 2022 to 2024.
2. What does "transformation-ready" mean for a finance organization?
It means AI is consistently embedded in workflows and decision-making at scale. Only 6% of CFOs in the IBM survey describe their finance function this way.
3. What is quote-to-cash?
It is the process spanning customer quotes, contracting, billing, collections and cash receipt. IBM redesigned it as one end-to-end workflow and reported 90% touchless automation.
4. Why do so many finance AI programs fail?
Kavanaugh points to sequencing. Companies start with AI technology before establishing data architecture and governance, and before redesigning the underlying workflow.
5. Why should business strategy drive an AI transformation?
Because value is created by changing how work and decisions happen, not by installing technology. Strategy defines which workflow matters, what success looks like and how much to invest, and architecture follows from those answers.
6. How does your background shape the way you approach these programs?
I have built and exited ventures, and I hold an MBA from Wharton and an MS in Machine Learning from MIT. That lets me work on the economics and the architecture in the same conversation, which is where most enterprise programs lose time.
7. Where should a large company begin?
With one cross-functional workflow that has clear financial metrics. Map the real process, assess data readiness, redesign end to end, then pilot with a baseline and an owner.
Sources
Keep reading
Previous article
Building an AI-Native Sales Org on the Stack You Already Own: Enterprise AI Transformation Lessons From Anthropic
Next article
Next article