From pilot to orchestrated AI: Solving legacy issues across ANZ enterprises

Industries across the world are recognising the cruciality of intelligent orchestration. It helps to coordinate people, artificial intelligence (AI) agents and workflows in real time to ensure the right people are handling the right task, at the right time. The need for orchestration is highlighted by a report of a top-ranked private research university that states: the vast majority of AI pilots across the world have not delivered tangible outcomes.

In Australia and New Zealand's (ANZ) banking and insurance, telecom and retail sectors, the impact of these unachieved outcomes has led to financial and other losses. Surprisingly, the main culprit hasn’t been technology but decades of accumulated process debt. The debt is not limited to processes, but extends to data debt and organisational debt. It highlights the absence of intelligent orchestration across enterprises.

The three kinds of organisational debts

Adoption of AI at a pilot scale is often successful. But when organisations move towards scale, the oh-so-successful AI is bound to amplify the existing debt. If the debt has been left unaddressed, AI just makes the cracks wider, which then translates into vicious cycles of lag and failures.

ANZ's legacy systems impose a brutal innovation tax

In ANZ, the high failure-rate challenge is intensified by the weight of legacy infrastructure. A leading global network of services firm’s report states that the Big 4 banks now spend a combined A$8.9 billion, annually, on technology, yet much of this goes toward maintaining legacy infrastructure. The telecom sector faces structurally similar challenges.
Here’s the foundational challenge: Driving tangible value out of AI is possible only if the technology is enabling intelligent automation and orchestration across critical operations.

Why are enterprises unable to create sustainable value?

The evidence overwhelmingly supports the thesis that the bottleneck is orchestration.
The 10–20–70 principle as stated in a playbook published by a leading global venture firm, quantifies that AI success is 10% algorithms, 20% data and technology and 70% people, processes, and cultural transformation.

The 10-20-70 principle

Most organisations aren’t able to correctly utilise data and technology. They invest in AI as a technology deployment rather than an operational and digital transformation. The companies getting AI right are focusing on orchestrating how to deliver outcomes with AI. Their success rests on understanding that existing workflows aren’t meant for AI.

Also read: Is using AI in customer service a smart choice

Here’s the blueprint organisations should follow:

1. Workflow redesign: Fundamentally restructuring how work flows through the organisation
2. Role redefinition: Creating new roles like "M-shaped supervisors" and "T-shaped experts"
3. Governance architecture: Establishing centralised AI governance with clear accountability

The new shift: From cost-arbitrage to AI-orchestrated business outcomes

The enterprise AI value proposition is fundamentally transitioning from labour cost reduction to outcome-based transformation.

There are three structural forces driving this shift:

• Software is becoming labour as AI turns service businesses into scalable software
• Per-seat pricing is dissolving as the natural metric becomes successful outcomes
• Variable costs have become unpredictable because AI inference creates 50–60% gross margins versus 80–90% for traditional software as a service (SaaS)

This has direct implications for the ANZ outsourcing and digital transformation market. Clients are pushing technology vendors away from billing by headcount and hours worked towards billing linked to outcomes and measurable business gains. That change strikes at the commercial core of the labour-arbitrage model. Moreover, the shift to outcome-based models requires continuous AI lifecycle management, which the industry calls machine learning operations (MLOps) maturity.

Intelligent orchestration is the competitive moat

The evidence from ANZ's banking, telecom and retail sectors points to a single, consistent conclusion: the organisations capturing AI value are the ones that invested in orchestration infrastructure before scaling pilots.

This is precisely what TP in ANZ is solving for enterprises. By bringing the right combination of people, processes, and technology, and orchestrating it with intelligence, TP has built the TP.ai FAB framework. It essentially reflects TP’s deep know-how of processes mixed with proprietary tools that can turn the complexity associated with AI into clarity and tangible business outcomes.

Connect with our experts to know more