What We Think · Editorial
The practitioners who run AI deployments inside SAP, Oracle, and Salesforce have a distinct view of what is working and what is not. This is where they write it down.
Elixir Perspectives is the editorial channel for practitioner thinking from the Elixir Global group. Points of view on agentic AI, enterprise transformation, finance operations, procurement, and HR. Not marketing copy. Not vendor positions. The view from inside a production deployment.
Manifesto · Featured
The Elixir AI-First Manifesto: Five operating principles for enterprises that are done running pilots
After two years of AI investment, most enterprise programmes have not reached production. This manifesto identifies the five structural decisions that separate the firms generating returns from those still in proof-of-concept. Every Elixir engagement runs on these principles. They are not internal guidelines. They are public commitments.
Read the manifesto5
Operating
Principles
Recent perspectives
Thinking from the Elixir practice teams.
Point of View
The dawn of the agentic enterprise: Why AI that recommends is no longer enough
The next phase of enterprise AI is not about AI that surfaces insights. It is about AI that acts. The structural difference between a recommending AI and an acting AI is not a technology decision. It is an operating model decision.
Research
SAP S/4HANA migration failures: The three decisions that determine whether a programme succeeds
Most SAP S/4HANA programmes that fail do so for predictable reasons. Data readiness, scope management, and the absence of a named delivery lead with accountability for the go-live date. This is an analysis of where the pattern breaks down.
Perspective
Agentic AI in finance operations: What a CFO should demand before the build begins
A CFO approving an AI deployment in the finance function has the right to ask five specific questions before signing off. This perspective sets out those questions and explains why the answers determine whether the programme reaches production.
Research
Healthcare revenue cycle in 2026: The operations still running on manual processes and what it costs them
Claim denial rates across Gulf and US healthcare enterprises have not materially improved despite two years of RCM investment. This research identifies the process design decisions that explain why, and what the AI-native operations model looks like.
Point of View
Why most enterprise procurement still takes weeks to process a purchase order when the technology to fix it has existed for four years
The SAP Ariba implementation is live. The procurement cycle still takes three weeks. This is not an Ariba problem. It is a work management problem that Ariba does not natively solve. The Procurement Control Tower was built to solve it.
Podcast
Finance leaders on what changed when AI ran the close: Three CFO conversations from 2026
Three finance leaders who have run AI-assisted close cycles share the same observation: the month-end sprint did not disappear when AI arrived. The work that remained after automation is the work that was always the hardest to automate.
Browse by topic
All perspectives, organised by practice area.
01/ 04
Agentic AI Framework. Permanent Operations. AI-First Manifesto. Production outcomes across SAP, Oracle, and Salesforce.
S/4HANA transformation. SAP Ariba procurement. BTP development. ERP modernisation decisions.
AI in finance operations. Close cycle automation. CFO decisions on AI investment. GST and compliance.
Revenue cycle management. Claims AI. Insurance operations. Finance operations in regulated sectors.
Thinking worth reading because it comes from people who have done it.
Subscribe to The Signal for a fortnightly briefing from the Elixir practice teams. The articles that matter, with context from the engagements that produced them.
Editorial types
Points of ViewResearchPerspectivesManifestosPractice areas
Agentic AISAP and ERPFinance and AccountingHealthcare and BFSIRelated
Research and ReportsPodcast: Elixir AI DispatchNewsletter: The SignalAnnual programmes
Annual AI OutlookIndustry BenchmarkingSignature Series