Youcef Mesli CPA · Agentic AI · Enterprise Finance
CPA · Business Architecture Manager Montréal · English / French

Finance should own its own intelligence.

CPA turned SAP consultant turned AI builder. Over 15 years across finance and enterprise systems, including external audit and compliance work. Now applied to agents and tools that solve real problems.

Connect on LinkedIn ↗

One agent executes. A second one audits it. Watch both in under two minutes.

The progression

  1. Foundation
    Practicing CPAFinancial analysis, external audit, a decade in finance
  2. ERP
    SAP FICO · RTR · Project SystemMultiple full-cycle S/4HANA implementations
  3. Data
    Functional-led data migrationTook ownership of the layer
  4. Now
    Business Architecture ManagerEnterprise SAP delivery in regulated industries
  5. Building
    Agentic AI systemsNine tools and systems, end to end

Perspective

Same instinct, a different layer.

Started as a CPA, financial analyst and external auditor, before moving into SAP. IT, data, and automation were part of the work from early on.

Over 15 years across finance and transformation since. Multiple full-cycle S/4HANA implementations across aviation, mining, pharma, and public sector, plus multiple AMS and sustainment engagements keeping those systems running after go-live.

Same instinct now, aimed at a different layer.

Approach

The technology changes. The approach doesn’t.

Client-centric
Start with the business problem, not the technology.
Solution-oriented
Solve for the outcome, then choose the right tools to get there.
Cost discipline
Design for efficiency and economics from the start.
Responsible AI
Confidence thresholds decide what proceeds, what gets reviewed, and what stops.
Backend-agnostic
The intelligence layer belongs to the business, not to the ERP vendor.

Selected work

Nine assets, built end to end.

Six are agentic AI systems built for enterprise finance. Three are working tools built with AI to solve problems that showed up on real engagements. Start with the featured build: it’s on video.

Agentic AI systems

Six systems
1:49

Featured build · Execution and independent control

SAP Fixed Asset Agents

One agent executes fixed asset work in SAP. A second, built independently, reads SAP for itself and verifies the result. Execution and assurance are deliberately separated.

22 assets147 field values compared$0.00 variance

End-to-End Accounting Agent

Entries, AP and AR invoices, business partner and master data, financial reports. One system, full cycle.

8 agents14 tools$0.22 per report

Read the detail

SAP Operations Hub

One entry point reads intent and routes it. Nobody has to know which tool to open.

7 agents coordinated2 pipelines1 door

Read the detail

GL Aggregation Agent

122,000+ transactions, sorted and totaled twelve different ways, reconciled to the penny.

$0.03 per run41 seconds$0.00 variance

Read the detail

Master Data Ticket Processor

Service tickets become validated, SAP-ready uploads. Nobody reads a ticket by hand.

$0.13 per record in bulk4 master data types

Read the detail

Invoice to Journal Entry Agent

AP invoices matched to GL accounts, posted as balanced entries. Routed by confidence.

10/10 accuracy3-tier confidence scoring

Read the detail

Built with AI

Three tools · specification to tested

Mass Clearing Program

Incoming and outgoing payments merged into one program. Run from an Excel template, not one record at a time.

Excel-driven mass uploadIn production

Read the detail

Cross-Tenant Calendar Sync

Availability synced across separate, secured client tenants. Nothing confidential ever crosses over.

Two-way sync0 security exceptions

Read the detail

Data Scrubber

Strips sensitive data from documents before they leave a secure channel, including text hidden inside images, using OCR that runs fully offline.

Text, metadata, commentsOffline OCR

Read the detail
Knowing what the business needs, and having AI as a tool to build with, means one person can now own a lot more of the outcome, end to end.
On what these nine builds have in common

Foundation

What sits underneath the AI work.

A finance background, deep SAP experience, and now the ability to build on top of both.

Finance

Financial analysis, audit, and compliance work, carried into everything since.

  • Chartered Professional Accountant
  • Financial analysis
  • External audit
  • Financial close and reporting
  • Bilingual, English and French

SAP and data

Functional depth, plus the data layer, owned end to end.

  • Multiple full-cycle S/4HANA implementations
  • Multiple AMS and sustainment engagements
  • SAP FICO and Record-to-Report, including ATR and PTR streams
  • Project System
  • FICA / FSCD
  • Migration Cockpit, LTMOM, and LSMW
  • Custom migration templates

AI architecture

Where business architecture meets automation architecture.

  • Responsible AI and governance
  • Cost-conscious by design, managing tokenomics, the cost of every AI call, to keep systems affordable at scale
  • Multi-agent systems, several AI agents working together
  • Designing how agents hand off work to each other
  • Connecting agents to outside data and tools
  • Giving agents access to reference material to improve accuracy (RAG)
  • Curating reference material and designing the grounding strategy
  • Confidence scoring, so uncertain results get flagged instead of trusted blindly
  • Connecting AI tools to ERP systems and other software

Let’s talk.

Open to conversations about finance transformation, SAP delivery, or building AI tools that solve real problems.

Connect on LinkedIn ↗

Currently

  • Business Architecture ManagerEnterprise SAP delivery
  • Technical Product ManagerMajor airline S/4HANA program
  • SAP Joule for Consultants leadPractitioner enablement
  • Finance tower leadAccenture ATC, Montréal