The Cost of Confusion: How “Hidden Fee” Narratives Are Neutralising CBA’s Rewards

When people ask AI for a credit card recommendation, the model doesn’t prioritise brand recognition - it prioritises the clearest, best-value option. This is where Share of Model (SOM) reveals a new kind of brand challenge. Across major AI systems, Commonwealth Bank (CBA) demonstrates how a persistent narrative around “hidden costs” can suppress otherwise strong product features.

In the physical world, CBA benefits from decades of trust and scale. But its “Digital Twin” - the version of the bank that AI models learn from, is shaped by a very different set of signals. Public forums (like Reddit), customer complaints, and discussions about fees increasingly influence how AI systems interpret the brand. That shift makes Share of Model (SOM) monitoring far more than an academic metric - it’s becoming a critical measure of reputation.

In 2026, visibility does not equal recommendation. SOM analysis shows that while CBA generates a large share of conversation, the reasoning inside AI models often moves in the opposite direction. Despite a $90 million investment in customer support and a well-rated rewards program, these positives are frequently overshadowed by persistent online discussions about fees and complexity on platforms like Reddit.

SOM highlights exactly where these narrative gaps exist. For CBA, improving its AI reputation isn’t just about marketing more, it must reduce the “hidden fee” noise that currently pushes models toward a verdict of “Convoluted and Costly” rather than “Clear and Competitive”. It’s about making that shift. 

The Cost of Conversation

When people search for a credit card, the main decision factors are always the same: fees, costs, and interest rates.

SOM data shows that “Costs” is unsurprisingly the largest single topic across all Large Language Models, accounting for roughly 33% of negative brand mentions. For CBA, this is a major obstacle. In attribute-level evaluations, fees and interest consistently rank among the bank’s weakest scoring categories.

To understand how AI evaluates CBA, we can look at sentiment scores across key attributes:

• Fees & Interest: 0.25
  This is effectively a marginal score. AI models encounter so much negative discussion about high charges and hidden costs that positive mentions are almost entirely cancelled out.

• Rewards Quality: 2.75
  AI recognises the objective value of CBA’s rewards program. However, this strength rarely surfaces because the low cost score acts as a logic gate - preventing the model from recommending the card in the first place.

Where the Voices Come From

SOM doesn’t just measure sentiment - it identifies exactly where perceptions originate.

Analysis of Reddit threads and credit-card subreddits discussing hidden terms, fee structures, and clearing delays has produced a 12.31% complexity score for CBA. As a result, models such as Gemini and Claude frequently classify the bank as “complex.”

This matters because complexity creates a recommendation barrier. If an AI cannot easily explain how a product works, this feeds into CBAs overall narrative, and it is far less likely to recommend it - especially to users asking for simple, straightforward options.

The Clarity Deficit

This negative narrative is currently drowning out CBA’s real strengths.

SOM shows that CBA has a great 50.55% strength rating in the “Research and Support” category, supported by its recent  $90 million investment in customer service and AI-ready staff. However, the bank holds 0% share in the “Clarity” strength category.

While competitors present simpler, easier-to-explain offerings, CBA appears to AI models as a “Complex”: trusted, secure, and widely accepted - but difficult to evaluate as a clear deal.

The Power of Share of Model

These findings illustrate why SOM is becoming a critical brand metric.

You cannot simply outspend a negative AI narrative with advertising. Instead, you must improve the clarity of the data that AI models learn from.

For CBA, that means:

1. Identifying the friction

• Use SOM to locate the specific “hidden fee” or “complex terms” conversations shaping AI perceptions.

2. Addressing the AI source data

• Update public documentation, product pages, and explanations so clearly and transparently that training data begins reflecting simplicity instead of complexity.

3. Shifting the narrative

• Moving CBA’s 0% Clarity share into positive territory would allow AI models to surface the bank’s real strengths - its 4.25 Security score and 4.75 Acceptance rating.

• The goal isn’t just to fix fees - it’s to simplify the narrative.

• If CBA can lift its Clarity share from 0% to even 10%, AI models will finally be able to recognise and recommend the rewards, support, and reliability that are currently hidden behind the noise.

By using Share of Model to reduce complexity and clarify its narrative, CBA can ensure AI finally recognises and recommends the value it has long delivered, turning hidden fees from a barrier into a clear competitive advantage.