Strategic collaboration targets regulatory complexity using IBM’s watsonx Orchestrate
In sum – what we know:
- AI for governance – e& and IBM announced a collaboration at the World Economic Forum to deploy enterprise-grade agentic AI for governance, risk, and compliance (GRC).
- The tech stack – The solution utilizes IBM watsonx Orchestrate and watsonx.governance to create action-oriented agents that integrate with enterprise systems.
- A benefit to telcos – The initiative addresses high regulatory complexity and operational scale across e&’s markets in the Middle East and Africa.
Governance, risk, and compliance have never been glamorous work, but for telcos operating across dozens of jurisdictions, they’ve become inescapable. Regulatory requirements keep multiplying, operational complexity keeps growing, and the case for automation becomes increasingly tempting. A new partnership from e& and IBM, however, aims to solve that. The announcement involves deploying “agentic AI” to tackle compliance workflows at enterprise scale. The pitch is compelling — involving AI that doesn’t just field questions but actually reasons through regulatory tasks and executes, though obviously within defined guardrails.
The question, of course, is whether the tech can actually deliver on its promises, or if it’s a little more ambitious than it should be. GRC is exactly where AI failures could hurt most. Missed deadlines, botched policy interpretations, decisions that can’t be explained to regulators could expose a major operator to serious liability. The technology’s sophistication matters far less than whether IBM and e& have built guardrails capable of containing the risks that come with letting AI systems make decisions that actually count.
The announcement
The partnership was announced at the World Economic Forum Annual Meeting in Davos, and sees e&, the UAE-based technology group that used to operate as Etisalat, teaming up with IBM and regional partner Gulf Business Machines, to roll out what both sides describe as one of the Middle East’s first enterprise-grade agentic AI deployments. e& operates in 38 markets, and has over 200 million customers.
The core ambition here is moving past the chatbot paradigm that’s become table stakes in enterprise settings. Traditional NLP tools can answer questions about compliance policies well enough. Agentic AI is meant to take that to the next level, with the ability to reason through complex tasks, orchestrate actions across enterprise systems, and actually manage compliance workflows rather than just retrieve information about them.
Hatem Dowidar, e& Group CEO, positioned the initiative as a broad shift: “Our ambition is to move beyond isolated AI use cases toward enterprise-scale agentic AI that is trusted, governed, and deeply integrated into how the organization operates. By collaborating with IBM, we are embedding intelligence directly into our risk and compliance processes, enabling faster decisions, consistent policy interpretation, and a foundation for broader agentic AI adoption across the enterprise.”
AI governance
IBM’s watsonx Orchestrate platform forms the technical backbone, bringing more than 500 tools and customizable domain-specific agents to the table. The platform ties into IBM OpenPages and the broader watsonx portfolio, including watsonx.governance, which e& had already adopted before this announcement.
As part of the hybrid model, large language models can run on customer-managed infrastructure rather than defaulting to IBM’s cloud environment. For a telecom operator juggling sensitive regulatory data across multiple national jurisdictions, that flexibility directly addresses real concerns around data sovereignty and security controls.
IBM is emphasizing what it calls “compliance by design” principles throughout the deployment. Every AI-driven action and recommendation is built to be traceable, auditable, and explainable. Ana Paula Assis, IBM’s SVP and Chair for Europe, the Middle East, Africa, and Asia Pacific, acknowledged the stakes directly: “As organizations move from experimenting with AI to embedding it into the fabric of how they operate, governance and accountability become just as important as intelligence. Through our collaboration with e&, this proof of concept intends to demonstrate how agentic AI can be designed and validated for enterprise-scale use.”
IBM’s Client Engineering team, working alongside GBM and e&, delivered the proof of concept in eight weeks. The speed is noteworthy, though it does raise questions about how thoroughly the system has been tested against the edge cases and adversarial inputs that compliance environments inevitably surface.
Crucially, the system is designed to support human-led decision making rather than autonomous AI actions. For high-stakes governance applications where errors carry severe consequences and regulators expect human accountability, this distinction matters enormously. Whether the human-in-the-loop approach survives the inevitable pressure to automate more aggressively as the system matures is another question.
Useful for telcos?
The compliance burden facing telecom operators is substantial and shows no signs of easing. Companies like e& operate across the Middle East and Africa, navigating distinct regulatory frameworks, reporting requirements, and enforcement regimes in each market. Managing that manually demands significant headcount and creates persistent risk of inconsistent policy interpretation across the organization.
The deployment targets faster response times for policy and regulatory inquiries, with 24/7 self-service capabilities positioned as especially valuable for large-scale operations where compliance questions don’t respect business hours and regulatory delays can mean penalties.
There’s a competitive dimension here too. e& is staking out ground as an early adopter of advanced AI governance in the region. In an industry where digital transformation has become a key differentiator, and where regulatory relationships often shape market access and licensing outcomes, demonstrating sophisticated, responsible AI deployment may carry value well beyond operational efficiency gains.
The initiative also reflects a broader shift in how legacy telecom operators are approaching AI. Many of them are starting by treating it as a customer-facing tool, however slowly, they’re moving into implementing it as an automation tool too. That’s a more mature approach to enterprise AI adoption — and a more consequential one, especially when mistakes are made.
Still, the telecom industry has seen plenty of ambitious technology partnerships announced at high-profile venues that ultimately delivered less than advertised. A proof of concept is a long way from enterprise-scale production deployment, and an eight-week development timeline, however impressive, leaves fundamental questions about long-term reliability and edge case handling unanswered.
Deploying AI in these domains potentially creates new failure modes even as it eliminates others. IBM and e& have clearly wrestled with this challenge, building in explainability, auditability, and human oversight.
