Webinar hosted by :
Hamza Charrat
EPM transformation AdvisorFinance IT Project Manager
Youssef Sebbar
Senior Finance Transformation LeadProject manager EPM
Thierry Beaurepaire
Expert Anaplan, IBM Cognos, Pigment
IN BRIEF: WHAT YOU WILL LEARN
1EPM replaces Excel-based management with a single, shared model: its real gain is not reporting, it is the ability to simulate and make trade-offs during the meeting.
2An EPM project is not a tool project: data and master data, governance, user support and a step-by-step roadmap determine its success.
3V-model or agile: the method depends on the organization's maturity. For a first implementation, prototyping is often the best route.
4Pigment, Anaplan, Oracle: you don't choose the best tool on the market, but the best EPM tool for your profile, with 5 trade-offs, a demo on your own data and a 5-year TCO.
5AI is already in EPM tools, but 63% of finance organizations find its rollout slower than expected: it speeds up a healthy model, it does not fix poorly governed data.
EPM (Enterprise Performance Management) refers to the set of processes and tools that allow a company to plan, budget, forecast, consolidate and manage its performance. In practice, it is what lets a finance department move from Excel files consolidated by hand to a single model, shared with the business, on which scenarios can be simulated and decisions made.
Yet in many organizations the situation remains the same: closings that drag on, as many versions of the truth as there are files, forecasts that arrive after the decision they were meant to inform. And sometimes an EPM that is already in place… bypassed by the spreadsheet.
At the same time, artificial intelligence is arriving at every vendor. The question is therefore no longer just “which tool?”, but also “what will AI really change?”.
On October 15, 2026, FocusTribes brought together three EPM consultants from its community, Hamza Charrat, Youssef Sebbar and Thierry Beaurepaire, for the webinar “Deciding faster and smarter: EPM in the age of AI”. This article summarizes the key takeaways: what an EPM project really involves, how to choose an EPM tool, three field case studies and how much of the AI in EPM story is real.
Whether you are a CFO, management controller, FP&A manager or CIO, you will find what you need to scope your project with method and clear-sightedness.
To go further on the data foundations of such a project, see our complete guide to data governance, and for AI on the ERP side, our article on SAP Joule and SAP Business AI.
I- What EPM really changes for the finance department
EPM relieves six symptoms that every finance department describes, in different words. Its real gain is not reporting: it is the ability to simulate and make trade-offs during the meeting.
Excel, closings, master data: the symptoms of management running out of steam
Finance, IT, supply chain: the speakers see the same difficulties from one project to the next.
- Excel remains the real management tool: as many versions of the truth as there are files, versioning that is impossible to trace, multiple data sources.
- Closings take too long: the figure arrives after the decision it was meant to inform.
- No autonomy for the business: every change to the model goes back through IT.
- Non-harmonized master data: naming conventions differ from one entity to another.
- Impossible to replay a scenario: no simulation in the meeting, hence no supported trade-offs.
- An EPM in place… and Excel everywhere: the tool exists, but usage has stayed on the spreadsheet.
For Thierry Beaurepaire, the challenge is clear: moving from fragmented management to more collaborative, better integrated and forward-looking management.
From reporting to simulation: making trade-offs in the meeting
✕ Spreadsheet mode✓ With a well-run EPM
Consolidation
✕In weeks
→In days, continuously
Source of truth
✕One version per file
→A single source
Scope
✕Finance alone
→Finance, HR and supply chain plan together
Simulation
✕Impossible in the meeting
→Test a hypothesis in real time: “what if sales fall by 5%?”
Approval workflow
✕Emails and files
→A controlled, traceable workflow
The scope goes well beyond finance. Asked during the Q&A session, the speakers cited payroll valuation on the HR side, lease and land management in real estate, cost allocation and costing to establish the full cost of a service, or even merchandising, assortment and strategic planning in retail.
An EPM in place… and Excel everywhere: why the tool is not enough
Many organizations already have a legacy EPM tool, and yet their teams keep working in their spreadsheets. The cause is rarely technical: processes were not rethought, the master data was not cleaned up, users were not brought on board.
“When a project goes off the rails, the tool isn't necessarily the cause. You have to look at the data scope, data governance and adoption.”
Hamza Charrat
II- Why an EPM project is not a tool project
An EPM project is a finance transformation: four workstreams come into play before the technology, the method depends on the organization's maturity, and the path is built in stages.
The 4 workstreams that come before the technology
“An EPM project is really a finance transformation that brings together several workstreams. It isn't pure technology: these are projects within the project.”
Hamza Charrat
01
Data and master data: clean up before tooling up
Harmonize naming conventions, data governance built around an MDM (Master Data Management), a data quality framework.
02
Governance and method
Who arbitrates, who validates, how often, with what project governance bodies.
03
Users
Awareness, training, business sponsors: all the more necessary when EPM maturity is low.
04
Roadmap and blueprint
A step-by-step path, aligned with the organization's actual maturity.
On the data workstream, read our case study on an MDM project run under constraints; on user support, the levers of change management at AXA and Parfums Christian Dior apply fully.
V-model or agile: maturity decides
Should an EPM project be run using a V-model or agile approach? For Hamza Charrat, the debate is framed badly. The right method depends on two criteria: the organization's EPM maturity and the nature of the project.
What is the organization's EPM maturity?
IT: already equipped, comfortable with multidimensional models
V-model
Specify → Build → Test → Deploy
Benefits: validated specifications, stable scope, clear milestones and workload commitment. Typically: a migration.
IT: first implementation, moving from Excel to EPM
Agile and prototyping
Prototype ↻ Validate in workshops ↻ Adjust
Benefits: no more tunnel effect, shorter testing phase, easier change management: the client learns by validating.
In a first implementation, the client discovers multidimensional logic: validating specifications written in a jargon they don't yet master is a tall order. Iterating with prototypes from the very first steps lets them build maturity and co-build the solution with the integrator.
“The client wasn't validating the functional specifications: too technical. We switched to prototyping directly on their Excel file.”
Hamza Charrat
Hamza Charrat's conclusion is unambiguous: there is no good or bad method, only the one best suited to the organization's context. To compare approaches more broadly, see our article on project management methods.
A roadmap in stages, not a big bang
The starting point is not the same everywhere, and the path must adapt to it. The speakers distinguish four stages:
STAGE 1
Everything in Excel
Make the master data reliable, scope the real need, prove the value through a targeted POC.
STAGE 2
A legacy tool
Don't migrate for the sake of migrating: rethink processes and calculation models before changing tools.
STAGE 3
A modern platform
Connected planning, driver-based planning, scenarios played in the meeting.
STAGE 4
Augmented management
Continuous forecasting, anomaly detection, AI agents, on a healthy foundation.
Maturity→
For Thierry Beaurepaire, the goal is to move toward a better-integrated platform, even if it means defining a common reference framework: a data hub that shares information between finance, HR and supply chain.
The three pitfalls to avoid: selling a stage 4 to an organization still at stage 1, changing vendor every two years thinking you are changing stage, and redoing in the EPM exactly what you did in Excel.
III- Pigment, Anaplan, Oracle: how to choose your EPM tool?
Functionally, EPM platforms converge: you don't choose the best tool on the market, but the best tool for your profile, based on five trade-offs and a demo on your own data.
Solutions overview: the tool was never the problem
The EPM market has expanded considerably, and the solutions are more alike than you might think. What sets them apart is their specialty: one excels at integration thanks to a built-in ETL, another stands out for its ease of use, a third is built for statutory consolidation.
Pigment
CONNECTED PLANNING
Finance, HR and supply chain on shared objects. French vendor.
Anaplan
MODELING
Powerful modeling and its own methodology, the Anaplan Way.
Oracle EPM
FULL SUITE
A very broad historic suite (Hyperion, Essbase, PBCS), with its own data integration layer.
Legacy platforms
INSTALLED BASE
Board, Cognos Planning, Tagetik, on-premise Essbase: still very widely installed.
Gartner's Magic Quadrant is often the first reflex. Yet it measures the completeness of the offering more than its relevance for your organization, and Gartner itself advises against choosing only among the best-placed vendors.
“The Magic Quadrant is an excellent starting point: it answers the question ‘who are the best vendors right now?’. But it doesn't answer the question ‘what is the right tool for the company?’.”
Youssef Sebbar
The 5 trade-offs and the 4-step method for choosing well
Five structuring trade-offs make it possible to quickly rule out a large part of the market:
Scope
Planning only
Planning and statutory consolidation
Solution philosophy
Native, pre-built finance functionality
Flexible platform, to be built
Model governance
Centralized, led by the group
Decentralized, entity autonomy
Level of change ambition
Continuity: Excel keeps its place
Break: overhaul and digitalization of processes
Tool administration
Owned by IT
Owned by finance and the business
Once these trade-offs are settled, Youssef Sebbar recommends a four-step selection method:
1
Clarify the trade-offs
Short-list of 2 to 3 solutions.
2
Weight the criteria
Scoring grid: functional coverage, architecture, IT integration, security, AI roadmap.
3
Test in a scripted demo
On your data and your real use cases, not on the vendor's demo.
4
Validate cost and references
TCO (total cost of ownership) over 5 years, customers in production.
Calculating TCO is not a detail: each vendor has its own pricing grid, and an expensive tool that isn't adopted can lead to a return… to Excel. Youssef Sebbar also warns against choices made solely on the basis of a relationship with a vendor, a source of adoption problems.
“Changing tools without a precise framework is just moving the problem while paying more. We will repeat the mistakes of the past.”
Hamza Charrat
FP&A: what the tool changes, and what it doesn't
Everyone wants FP&A to become a business partner that informs company decisions. Why is it taking so long? The figures from the FP&A Trends Survey 2025 (459 finance professionals) are telling:
46%
of FP&A teams' time still goes into collecting and validating data.
31%
only is devoted to analysis and action.
77% vs 27%
of forecasts rated good or excellent with a driver-based model, versus basic models or no model.
“FP&A isn't lacking in ambition: it just lacks the time to devote to business analysis.”
Youssef Sebbar
An EPM changes four dimensions of the job. Only 3% of teams run real-time scenarios today:
✕ Yesterday✓ With an EPM
The question asked
✕“What happened?”
→“What happens if…?”
The pace
✕Fixed annual budget, reforecasts tied to the calendar
→Rolling forecast and replanning as soon as an event occurs
The model
✕Excel files consolidated by hand
→Single, driver-based model (volumes, prices, headcount), connected to the P&L
The team's role
✕Collect, reconcile, produce reporting
→Analyze and inform decisions
Some EPM tools go further, with predictive analytics based on time series, Hamza Charrat points out. Youssef Sebbar adds a caveat: these advanced features assume that teams have the necessary maturity. The tool does not make the shift on its own: processes, data and skills matter as much as the platform.
Preparing to choose or migrate your EPM tool?
The EPM consultants of the FocusTribes community support you, from scoping to solution selection.
Let's talk about your projectIV- Three EPM case studies: what made the difference
Three projects, three contexts and one lesson: the gain came from what was done before and around the tool.
Railway infrastructure: a budget cycle shortened by 15 days
TRANSPORT & INFRASTRUCTURE · THIERRY BEAUREPAIRE
Unifying the management of more than 15,000 investment projects
Context: The finance departments of a large railway infrastructure manager had several EPM tools and decided to unify them. The scope: more than 15,000 investment projects and more than €6 billion in multi-year forecasts.
Challenge: The existing tool satisfied the finance teams, but was reaching its volume limits given the growing number of projects to manage.
What was done: A benchmark of six vendors, then a POC targeted at investment management. Pigment was selected by the business. Two conditions got the project through: strong sponsorship from the finance department and business involvement, in particular from management control during testing.
Result: a project delivered in 10 months, a budget cycle shortened by 15 days, users who are autonomous at both central and regional levels. Next step: new dashboards, then AI to analyze the calculation model.
Real estate, international group: a budget in 1 month instead of 3
FINANCE TRANSFORMATION · HAMZA CHARRAT
Making the reporting of figures between subsidiaries and headquarters reliable
Context: A group with subsidiaries abroad, which has already chosen its EPM tool and is entering the scoping phase.
Challenge: Figures were coming in by email, in several versions of Excel files in the middle of the budget exercise. For the group, the answer remained the tool: the process itself had never been questioned.
What was done: Mapping of processes, stakeholders and data flows, and flattening of business rules. A bottleneck emerged in the workshop: master data, with several naming conventions for the same indicator. Three decisions followed: harmonize business rules wherever possible, set up a systematic arbitration meeting with a regional manager, and bring the budget calendar closer to the accounting close.
Result: a budget cycle, from subsidiary input to group consolidation, cut from 3 months to 1 month. More reliable data, fewer back-and-forths, and exchanges tracked in the EPM workflow rather than in email chains.
“The group was right to equip itself with an EPM. But EPM doesn't necessarily make data consistent: it assumes it already is.”
Hamza Charrat
International assistance: recharges simulated several times a year
FINANCE TRANSFORMATION · YOUSSEF SEBBAR
Equipping the recharging of central function costs
Context: In a large international assistance group, central functions (IT, support functions) bear shared costs, recharged to local entities through allocation keys. These recharges weigh directly on each country's P&L and margin.
Challenge: A very complex Excel file, in multiple versions, with no simulation or traceability. The exercise was only run once a year, whereas the group wanted to run it two to three times.
What was done: Scoping with central management control and the entities, mapping of flows and cost types, formalization of the keys, then design of an Anaplan model with scenarios, versions and workflow. The challenge: getting common keys accepted across countries, a sensitive topic since it touches local margins, and finding the right balance between calculation accuracy and model simplicity.
Result: a tooled, shared and transparent process, run several times a year, that feeds directly into the P&L: this is the principle of connected planning, with a single source of truth for finance and HR.
The common lesson: in all three cases, the tool only delivered its effects once the processes, business rules and master data had been reset, with the business involved from start to finish.
V- EPM and AI: what AI will make possible, and what it won't
AI is already present at almost every EPM vendor, but the value has not yet followed: you have to sort out what is mature, what needs to be challenged and what remains a promise.
3 out of 4 CFOs
expect their IT budget to rise in 2026 (Gartner).
63%
of finance organizations found AI implementation slower than expected in 2025 (Gartner).
40%+
of agentic AI projects will be abandoned by the end of 2027 (Gartner).
The problem is not the absence of AI: it is already in the tools. It is knowing what really works.
“We're not here to replace the management controller, who knows their figures perfectly. AI is there to help them understand their model, or even define it.”
Thierry Beaurepaire
What is mature, what to challenge, what remains a promise
Mature
Predictive forecasting (machine learning) as a baseline, based on history and external data (Anaplan, OneStream).
Querying your data in natural language (Anaplan CoPlanner, CCH Tagetik in Excel).
Key takeaway: a starting point for the controller, not a replacement. Requires a well-structured model.
To challenge
AI-assisted model building.
Variance analysis and generated commentary.
Key takeaway: useful for a POC or a first draft, to be reviewed: the model builder remains accountable for the model.
Still a promise
Planning without human intervention.
Proven ROI at scale.
Key takeaway: even vendors put their agents under human validation.
Thierry Beaurepaire gives a concrete example of a short-term use: asking an agent a business question in natural language, such as “which Q3 capital expenditures exceed budget by more than 10%? Find the corresponding projects and analyze the variances.” In a few minutes, the agent scans the model and answers precisely. Provided the model is clean:
“If you have poorly named dimensions, the AI will give you off-target answers.”
Youssef Sebbar
The agentic shift: what level of autonomy?
Youssef Sebbar suggests imagining an AI agent as a new hire on the team: in the first month it answers questions, then it makes proposals, then it acts. Four levels of autonomy emerge:
1
Answer
To a question asked in natural language.
Available2
Alert and recommend
Spot variances, propose actions (Oracle Planning Agent, Pigment Planner).
Available3
Act with validation
E.g.: CCH Tagetik agents on intercompany reconciliations.
Available4
Act autonomously
Reserved tomorrow for low-risk tasks.
Not yetAgent autonomy→
Vendors are also opening studios for building your own agents. But caution is in order, as Gartner's forecast on the abandonment of many agentic AI projects reminds us, for lack of demonstrated value or control over costs and risks.
What AI won't solve for you
AI changes the speed of construction, not the quality of the foundations.
What AI speeds up
✓Building the data model and business rules: a prototype put together in a few clicks
✓Forecasting and anomaly detection
✓Team autonomy over their own analyses
What it doesn't solve
✕Master data that lives in Excel
✕Non-harmonized naming conventions
✕Missing governance and teams left without support
For Youssef Sebbar, three conditions set apart the projects that move forward:
1Targeted, measurable use cases, identified in advance so that a return on investment can be calculated.
2A controlled model and data foundation, with processes in place.
3Finance–IT governance set from the start to frame agentic uses.
These lessons echo those of other large groups: read the case studies on AI integration at PepsiCo and L’Oréal.
THE 4 KEY TAKEAWAYS FROM THE WEBINAR
1An EPM project is won upstream: data, master data, governance and users before the technology.
2The method adapts to maturity: V-model for mature organizations, prototyping for a first implementation.
3Choose for your profile, not for the ranking: 5 trade-offs, a demo on your data, a 5-year TCO.
4AI amplifies, it doesn't repair: start with targeted, measurable use cases, on a healthy foundation.
VI- FAQ: EPM project and artificial intelligence
Q. What is an EPM tool and what is it used for?
An EPM (Enterprise Performance Management) tool is used to plan, budget, forecast, consolidate and manage company performance in a single model. It replaces Excel files consolidated by hand with a source of truth shared between finance and the business, and lets you simulate scenarios to make trade-offs during the meeting itself.
Q. What is the difference between EPM, ERP and BI?
The ERP records the company's transactions (accounting, purchasing, sales). BI analyzes and reports data, mostly historical, in dashboards. EPM looks ahead: it plans, budgets, forecasts and simulates, drawing on ERP data. The three are complementary. To dig deeper into BI, see our key stages of a BI project.
Q. How long does an EPM project take?
The duration depends on the scope, the quality of the data and the organization's maturity. For example, the EPM migration of a large railway infrastructure manager, covering more than 15,000 investment projects, was completed in 10 months. A first implementation is best approached in stages, starting with a targeted POC.
Q. How do you choose between Pigment, Anaplan and Oracle EPM?
There is no such thing as the best tool in absolute terms. Start by settling five trade-offs (scope, solution philosophy, model governance, level of change ambition, tool administration), draw up a short-list of 2 to 3 solutions, test them in a demo on your own data, then compare their 5-year TCO and their production references.
Q. Will AI replace management controllers and FP&A teams?
No. The AI built into EPM tools provides baseline forecasts, answers questions in natural language and offers first variance analyses, but human validation remains essential. Above all, it frees up time spent on data collection and reconciliation, so that teams can focus on analysis and on informing decisions.
An EPM does not make an organization more agile by itself: it makes visible, and fast, what the organization had structured before it.
AI will accelerate this movement, provided the foundations are there. The question is therefore not “when?”, but “are we ready?”. To review the discussion in full, watch the webinar replay and download the ebook “EPM in the Age of AI”, which includes the complete project checklist.
An EPM project to scope?
Let's spend 30 minutes on your context, your maturity and the profiles to mobilize: FocusTribes is a selective community of more than 1,200 accredited independent consultants.
Let's talk about your project