Every finance transformation conversation we have these days starts with AI. Clients arrive at the table with a list of agentic capabilities they want to deploy and an assumption that technology is the answer.
That assumption is half right. AI is one of the most consequential forces reshaping finance in a generation, and the future-state operating models we are designing today look fundamentally different as a result. But AI alone is not a strategic finance roadmap. The return comes from sequencing AI with the right process redesign, technology architecture and data foundation, with each piece compounding the next.
For decades, CFOs treated transformation like a renovation, managing tasks room by room. They executed separate projects to update the close, replace an ERP and layer better reporting on the same operating model. AI changes that blueprint. For the first time, we can take the house down to the studs and rebuild around capabilities that did not exist three years ago. That is a different kind of project, and it requires a different kind of leader. AI has done something that previous waves of transformation never managed to do. It has brought the right stakeholders into the conversation and forced finance and IT to address the same problem simultaneously. That side effect alone is reshaping how the next phase of transformation gets built.
This is the part of the conversation IT leaders need to understand.
For most of the last decade, finance transformation was launched in response to a single trigger. A cost-out program. A post-acquisition integration. An ERP that aged out. The CFO owned the initiative, the CIO got pulled in to handle the technology and the rest of the business found out when the new system went live. The AI moment has changed that pattern and forced a new conversation. When you are taking the house down to the studs rather than swapping out fixtures, the work cannot be done by Finance or IT alone. Stakeholder alignment becomes a precondition for the rebuild, not a nice-to-have. AI is the attraction that pulls everyone into the room.
When a client walks in asking about agentic finance, the CFO, CIO and heads of accounting, FP&A and controllership all show up for the same assessment. Leaders who have historically run parallel agendas now talk about the same problem at the same time. By the time we get to solution design, the cross-functional alignment that used to require months has already happened. Issues get raised early. The transformation moves faster because everyone agreed on the destination before we picked the route.
Maybe we can thank AI for bringing people together who otherwise wouldn’t have been collaborating. That is the reason the next phase of finance transformation looks structurally different from the last.
A strategic finance roadmap is not a project plan. It is more like an architect’s drawings, providing a multi-year design that connects business strategy to a future-state operating model, sequenced so each step compounds rather than starts over.
Every roadmap has a North Star. That future-state vision almost always has an AI component. A multi-agent solution orchestrated on a platform. Continuous close. Predictive FP&A. Real-time controls. Humans in the Loop. A streamlined org chart. That vision, eventually, becomes the destination.
But the destination is not today. What matters today is that the work between the as-is and future-state is cumulative, not throwaway. The data foundation organized for one transformation becomes the foundation for the next. Process improvements made now will accelerate the AI capabilities deployed two years from now.
You do not need an architect for a renovation. You hire a contractor, hand them a task list and inspect the results. You need an architect when you take the house down to the studs. The role most transformation programs are missing right now is not another builder. There are plenty of developers, system implementers and business integrators in any given program. What is missing is the architect.
The architect designs how the parts come together before anyone starts building. That work runs side by side with the client, connecting the dots between the technology, the problems, the data and the functions inside Finance and IT. The architect’s job is to create the vision and the blueprint and to translate between groups that do not naturally speak the same language. Done well, the result is a structure that is fundamentally sound and built to expand as the client’s appetite grows. Done poorly or skipped, the result is a house that cannot accommodate what comes next.
The traditional transformation pitch promises ROI at the end. Wait three years and the numbers will look great. That pitch is structurally fragile, and every CIO who has watched a transformation budget get cut in year two knows why.
Stakeholder fatigue is real. Budget uncertainty is real. In private equity-backed companies, the three-to-five-year hold period is real. Any program that has not produced measurable value in the first phase loses political support before the second phase is built.
The architectural alternative is to design the program so the first floor is occupied while the second and third floors are still being completed. The first process redesign produces value while the next is being scoped. The first technology deployment generates a return while the next is being implemented. The roadmap is not a march toward a distant payoff. It is a sequence of compounding wins.
The materiality of the return matters more than the size. A small process improvement or quick win that delivers a meaningful percentage gain in week ten is worth more than a large transformation that promises a bigger gain in year three. Quick wins prove that the blueprint is working. That proof builds the confidence stakeholders need to fund the next phase and the one after that. Investors want immediate ROI. Boards want immediate ROI. The strategic finance roadmap should be designed to deliver it.
Picture data as the electricity and water in a building: every room needs it, and the pipes and wires that carry it are what make the structure livable. Consolidating, warehousing and improving data quality is how you run those lines. It strengthens every element of the roadmap, not just the AI.
That has implications for how the stack gets selected. Most clients end up with an ecosystem rather than a single platform. ERP, EPM, close management, procurement, reporting and an emerging set of agentic capabilities, all integrated against a shared data foundation. The boundaries between those tools matter less than the data architecture that connects them.
It also has implications for what the CIO needs to be doing right now, even on transformations not yet formally launched. The data work pays dividends regardless of which solutions eventually get built. Cybersecurity, controls and business continuity are not bolted on at the end. They are design constraints embedded in the architecture from day one. SOX compliance is easier to build in than to retrofit, as is every other control discipline that lands on the CIO’s desk.
The most concrete way to think about where this is going is to look at the org chart.
The finance org chart five years from now is not going to look like the org chart today. Where there used to be five controllers, there may be one human controller and three E-controllers, with agents sitting alongside them on the chart. The remaining human roles will be split between onshore and offshore in ways that look unfamiliar from where we sit now. The balance and location of every component will shift.
The strategic finance roadmap is the artifact that builds toward that end state, and the roadmap itself will evolve as the work progresses. Not a destination, but a continuous design exercise.
AI is rebuilding finance from the studs up. The strategic finance roadmap is how we make sure the new house is structurally sound, produces returns from the first floor and reflects the vision and the priorities of the people who will live in it. That work belongs to the CFO and the CIO together, or it does not become livable.
This article is published as part of the Foundry Expert Contributor Network.Want to join?