She Trained the AI Chatbot That Fired Her: Why Banks Are Learning the Hard Way

After 25 years at Commonwealth Bank, Kathryn Sullivan trained the AI chatbot that replaced her job at age 63. Her story reveals why AI-only customer service fails—and what companies get wrong about automation.

YEET MAGAZINE

20 Sep 2025 • 3 min read

By YEET MAGAZINE

"Machines might mimic scripts, but they cannot replace compassion, judgment, or the trust built by real people." – Project Nightfall

"After cutting staff, the bank admitted customer calls only increased." – ABC News Australia

Here's what happened: Kathryn Sullivan gave 25 years to Commonwealth Bank of Australia. She trained staff. She built customer relationships. She even helped teach their AI chatbot how to answer questions. Then at 63, the bank said she wasn't needed anymore. The same chatbot she trained replaced her.

This isn't a hypothetical. This is what happens when companies prioritize cost-cutting over actually understanding how AI fits into real work.

Why The Algorithm Can't Feel Your Pain

The bank's math was simple: AI is cheaper than payroll. Deploy the chatbot, cut staff, watch profits rise. Except it didn't work that way.

Customer complaints went up, not down. Calls increased. People weren't satisfied with robotic responses to complex problems. A chatbot can follow decision trees. It cannot understand frustration, build trust, or make judgment calls that require actual empathy.

Kathryn felt "discarded like a number." She'd given decades to the organization. The algorithm gave her a termination notice.

The AI Paradox Nobody Talks About

Here's what's wild: companies automating customer service don't realize they're removing the one thing that keeps customers loyal—actual human connection.

The data shows it. When automation replaces humans without a hybrid approach, customer satisfaction tanks. Retention drops. The algorithm optimizes for cost, not for outcomes that matter.

Kathryn's story isn't unique. It's a pattern playing out across banking, retail, and call centers. Someone teaches the AI how to do their job. The AI gets deployed. The human gets laid off. The company realizes too late that efficiency ≠ effectiveness.

What Should Actually Happen

Smart companies are doing this differently. They use AI to augment workers, not replace them. A chatbot handles simple FAQs. A human steps in for complex issues. The technology supports the person; the person provides the judgment.

This approach costs more upfront. It also costs less in the long run because you keep customers, retain institutional knowledge, and avoid the PR disaster of being the company that fired the person who trained your chatbot.

The future of work isn't AI or humans. It's AI and humans, designed intentionally.

Common Questions About AI Job Displacement

Can an AI chatbot actually replace a customer service worker?
Technically, yes—for simple, scripted interactions. But not for building trust, solving novel problems, or handling frustrated customers. Banks are discovering this the hard way.

Why do companies keep making this mistake?
Short-term cost savings look good on quarterly reports. Long-term customer satisfaction and retention don't show up until it's too late to fix.

What jobs are safest from AI automation?
Roles requiring empathy, complex judgment, relationship-building, and creative problem-solving. Jobs that are purely rule-based and repetitive are most at risk.

Could Kathryn have kept her job if she hadn't trained the AI?
Maybe not—but the irony reveals how backwards this approach is. The people with the most institutional knowledge are often the first casualties.

What does this mean for the future of banking?
Banks that figure out human-AI collaboration will win. Banks that go all-in on chatbots will lose customers to competitors with better service.

Is AI inherently bad for workers?
No. Badly implemented AI is bad for workers. Well-designed automation that enhances jobs instead of eliminating them creates better outcomes for everyone.

```