How AI Testing Changes the Implementation Playbook

Most SAP programmes prove a go-live the same way: a large test team, months of manual execution, and evidence assembled by hand at the end. With ECC mainstream maintenance ending in 2027, that model has run out of road. In this 60-minute session, Christophe Derdeyn (Managing Director of Services, Icon Consulting Group) is joined by Christophe le Mao (Chief Information Officer, ALEPH) to show what AI-enabled testing looks like in practice — on a live S/4HANA Public Cloud consolidation of 100+ entities across roughly 70 countries. Wave 5 of that programme was the first AI-enabled implementation: 9 entities, 7 countries, 8 finance process areas. One AI engineer and three part-time functional consultants turned user stories into 90 executable scripts in two weeks, ran 564 checks across every process and country in three days, and handed the business two weeks of validation working from evidence the runs produced rather than evidence it had to assemble. What we cover: · The constraints forcing the change — and which two never go away · Where AI actually augments the implementation lifecycle, and where it shouldn't · The Icon Test Agent Platform: run from a dashboard, or ask your AI assistant over MCP · What gets validated — inputs, the calculations the system makes on the way, and the posted outcome: G/L accounts, VAT, GST and withholding tax · The old testing process against the new one — team size, elapsed time, and who carries the load · ALEPH's CIO on what changed, and what was harder than expected Want to test the claim? Pick one process and one country. We stand the platform up against your tenant and return verified evidence for that scenario in two weeks, fixed price. cderdeyn@iconconsultinggroup.com iconconsultinggroup.com Icon Consulting Group — an augmented SAP boutique. Everyone talks about Cloud and Digital Transformation. We do it differently.