AI Programme & Product Lead

AI Programme & Product Lead Function: Big Data, AI & Intelligent Automation

Reports to Managing Executive: Big Data, AI & Intelligent Automation Direct reports Senior Business Analyst Location: Midrand

Role purpose A dual-mandate role at the centre of the AI function. As Programme Lead, it runs the operating system behind the AI strategy: converting strategic intent into a governed, sequenced, funded roadmap, then holding the organisation to it. It owns the single source of truth for what is being built, what it is worth, when it lands, and whether the value materialised.

As Product Owner, it takes personal delivery ownership of the function's flagship agentic products, from problem definition through agent design, build, deployment, adoption and measured impact. The two halves are deliberately connected. Portfolio credibility depends on knowing what delivery costs, and product decisions are better when made by someone who can see the whole portfolio. Indicative split of effort: approximately 60% programme and portfolio, 40% hands-on product ownership. This is not a PMO administration role, and it is not a backlog-management role. The successful candidate will think like an owner, challenge weak business cases, ship product, and be credible in front of both engineering teams and executive stakeholders.

Part A: Programme, portfolio and strategy Strategy support and execution

• Partner with the executive on drafting, refining and cascading the AI and Big Data strategy across its customer experience, monetisation and productivity pillars. • Translate strategy into an executable multi-quarter roadmap with named owners, sequencing logic, dependencies and funding requirements. • Prepare executive and governance forum material: narrative, evidence, decision asks. Argument construction, not slide production. • Maintain the pipeline of strategic asks covering platform, environments, headcount and procurement, and drive each to a decision against a deadline. C2 General Portfolio and roadmap tracking

• Own the consolidated view of the AI and automation use-case portfolio: stage, owner, blockers, dependencies and delivery confidence.
• Run intake and prioritisation: business case quality gates, effort-versus-value scoring, and kill criteria for use cases that stall.
• Manage graduation of use cases through the environment lifecycle to production against agreed criteria, and hold the line where criteria are not met.
• Drive delivery cadence: weekly delivery review, monthly portfolio review, quarterly re-plan.

Value tracking and benefits realisation

• Own the value ledger for the portfolio, tracking each use case from claimed, to modelled, to independently measured, to finance-verified. • Set and enforce measurement standards with the data science team: control group design, incrementality, pre-registered measurement. No value is recognised without a defensible counterfactual. • Reconcile AI-attributed value with Finance and commercial stakeholders and resolve double count risk across teams. • Produce a monthly value pack that withstands line-by-line executive scrutiny. Delivery management across the portfolio

• Unblock delivery across Technology, Data, Commercial, Legal and Risk, clearing dependencies before they become escalations. • Manage cloud provider and vendor engagements: scope, milestones, commercial hygiene, consumption management. • Track spend against platform and model-consumption budgets, flagging drift early. • Own programme risk, issue and decision logs. Escalate with a recommendation, never just a problem. Forward-thinking programme shaping

• Maintain a rolling 12-18 month view of where the AI and Big Data programme should go: capability gaps, platform bets, organisational implications, emerging technique adoption. • Lead the shift from reporting and point-solution machine learning toward agentic, data-driven use cases, including what must change in governance, measurement, skills and platform to support it. • Benchmark against telecommunications and cross-industry peers, bringing options and a point of view rather than a market summary. Part B: Product ownership Product vision and backlog

• Define and maintain the product vision, roadmap and success metrics for the function's flagship agentic products: conversational and service agents, customer-value agents, and the analytical products that feed them. • Own a prioritised, refined backlog with clear acceptance criteria, and make the trade-off calls on scope, sequence and technical debt. • Establish the problem statement before the solution. Remove features that do not move the metric. C2 General Agentic solution design

• Design agent architectures with the engineering team: task decomposition, orchestration, tool and data grounding, guardrails, human-in-the-loop points, escalation and fallback behaviour. • Define the data contracts and feature requirements each agent depends on, and work with data engineering to secure them. • Own evaluation design: what constitutes a good output, how it is scored, and how regressions are caught before customers experience them. Delivery

• Run the delivery cadence with the engineering team: sprint planning, refinement, reviews, release readiness. • Be the decision-maker in the room so engineers are never blocked awaiting business input. • Manage MVP scope tightly. Ship narrow and real, then expand. Adoption and lifecycle

• Own adoption. Work with business, marketing, care and channel teams on rollout, training, communication and feedback loops. A product with no users has no value.
• Establish support, monitoring and incident process with engineering, and own operational health metrics.
• Run structured experimentation after launch. Treat go-live as the start of the value curve, not the end of the project.
• Manage the roadmap from MVP through to scaled, multi-journey deployment.

Cross-cutting accountabilities

• Governance, risk and compliance • Ensure every use case clears data privacy (POPIA), consent, model governance, security and AI risk requirements before production. • Define decision authority and controls for agentic use cases: what an agent may decide autonomously, what requires human approval, and what the audit trail must capture. • Maintain the AI use-case register and audit trail to a standard that survives external review. Stakeholder management

• Act as the interface between the AI function and business owners, translating in both directions. • Present portfolio performance, product outcomes, roadmap and decision requests to senior stakeholders with evidence. Leadership

• Line-manage and develop the Senior Business Analyst and set the standard for use-case shaping and requirements quality across the function.

Scope boundaries Deliberate limits on the role, to protect objectivity and capacity: C2 General

• Value verification is not self-certified. Measurement design is reviewed by data science, and outcomes are verified with Finance. The role owns the ledger and the standard; it does not sign off its own numbers.
• Product ownership is scoped, not universal. The role personally owns a defined set of flagship agentic products. Other use cases in the portfolio carry business-side or delegated owners.
• Shaping and requirements sit with the Senior Business Analyst, freeing this role to focus on direction, prioritisation and unblocking.

Success measures: first 12 months Programme and portfolio

• The value ledger is trusted and reconciles to Finance without material dispute. • Portfolio financial commitments are met or exceeded, with the independently measured share of value rising each quarter. • Delivery predictability improves quarter on quarter against agreed milestones. • No production deployment bypasses governance or graduation criteria. • A credible, funded 18-month agentic AI roadmap exists and is being executed against. Product

• Flagship products are live in production, in daily use, with adoption tracked and growing