Why Marketing Managers Struggle with AI Success (And How to Fix It)

It seams that Marketing Managers face a peculiar problem: they're more likely to distrust AI when it performs well than when it fails. Recent research reveals a surprising psychological bias that's sabotaging AI adoption in retail—and it's not what you think.

A comprehensive study by researchers at Poznan University of Economics revealed that business managers consistently attribute AI success to luck or external factors, while crediting human success to skill and competence. This bias creates a dangerous cycle: the better your AI performs, the less your team trusts it.

For e-commerce managers investing in AI-powered recommendation engines, chatbots, and inventory management systems, understanding this psychological trap is crucial for maximising ROI and team adoption.

The Hidden Psychology Behind AI Resistance

When AI Succeeds, Humans Take Credit

The research uncovered a striking pattern across three separate studies involving over 350 marketing managers. When AI systems delivered successful outcomes—whether recommending profitable products or optimising campaigns—managers attributed the success to external factors like market conditions or pure chance.

Meanwhile, when human colleagues achieved identical results, managers credited their skills, experience, and decision-making abilities. This "locus of causality" bias creates a fundamental problem: your team doesn't recognise AI's genuine contributions.

The Poznań researchers found that managers consistently view AI as "following pre-set rules rather than demonstrating individual competence." This perception undermines confidence in AI systems, even when they consistently outperform human alternatives.

The Satisfaction Paradox

Here's where it gets interesting: this attribution bias directly impacts satisfaction with AI tools. The study demonstrated that when managers attribute AI success to external factors, their satisfaction with the AI system drops—even when performance remains identical.

This creates a dangerous feedback loop. Your team uses AI tools, they work brilliantly, but because success is seen as coincidental, satisfaction decreases. Lower satisfaction leads to reduced adoption, and reduced adoption means missing out on AI's competitive advantages.

For e-commerce businesses, this translates to real costs. Teams may abandon effective AI tools, revert to manual processes, or fail to expand AI integration across departments—all because of a psychological bias they don't even recognise.

What This Means for E-commerce Operations

Customer Service and Chatbots

Your customer service team might resist AI chatbots not because they're ineffective, but because when chatbots successfully resolve customer issues, staff attribute the success to simple customer queries rather than the AI's sophistication.

There is a risk that if AI handles customer interactions successfully, managers assume the problems were "easy to solve anyway." This perception prevents teams from recognising AI's value in handling complex customer scenarios and emotional intelligence tasks.

Inventory Management and Forecasting

AI-powered inventory systems often deliver remarkable accuracy in demand forecasting and stock optimisation. However, when these systems prevent stockouts or reduce excess inventory, procurement teams may credit market stability, seasonal trends, or their own oversight rather than the AI's predictive capabilities.

This misattribution can lead to underinvestment in AI systems precisely when they're delivering the most value—during stable market conditions where their subtle optimisations create significant cost savings.

Product Recommendations and Personalisation

Recommendation engines that boost conversion rates and average order values face similar challenges. When personalised product suggestions drive sales increases, marketing teams might attribute success to improved product photography, seasonal demand, or promotional campaigns rather than AI-driven personalisation.

This bias prevents teams from fully leveraging recommendation systems or investing in more sophisticated personalisation tools that could further enhance customer experience and revenue.

Strategies to Overcome Attribution Bias

Make AI Contributions Transparent

The research suggests that transparency can help combat attribution bias. Instead of presenting AI recommendations as black-box outputs, provide clear explanations of how the AI reached its conclusions.

For example, rather than simply stating "AI recommends increasing inventory for Product X," explain: "Based on analysis of 50,000 historical transactions, seasonal patterns, and current market indicators, AI forecasts 40% increased demand for Product X in the next two weeks."

This approach helps teams recognise AI's analytical sophistication rather than dismissing recommendations as simple rule-following.

Implement Attribution Reflection Practices

Build regular review sessions where teams explicitly discuss what factors contributed to successful outcomes. Ask questions like:

• What role did AI insights play in this success?
• Which human decisions complemented the AI's recommendations?
• How might the outcome have differed without AI support?

These discussions help teams develop more balanced attribution patterns and recognise AI contributions they might otherwise overlook.

Create Human-AI Collaboration Frameworks

Position AI as a collaborative partner rather than a replacement tool. The research found that when managers view AI as a team member rather than a simple tool, they're more likely to credit it appropriately for successful outcomes.

Develop decision-making processes that explicitly outline human and AI contributions. For instance, "AI analyses customer behaviour patterns, humans interpret business context and make final strategic decisions." This framework helps teams understand complementary roles rather than viewing AI success as somehow diminishing human contributions.

Track and Communicate AI Impact

Develop metrics that specifically highlight AI contributions to business outcomes. Create dashboards that show:

• Revenue increases attributable to AI recommendations
• Cost savings from AI-optimised inventory management
• Customer satisfaction improvements from AI-powered support

Regular reporting on these metrics helps combat the tendency to overlook AI's role in successful outcomes.

Building Trust Through Balanced Expectations

Address Both Success and Failure

The research revealed that managers readily blame AI for failures while crediting themselves for successes. Combat this double standard by maintaining balanced perspectives on both positive and negative outcomes.

When AI systems underperform, conduct thorough analyses that consider data quality, implementation factors, and market conditions rather than simply blaming the technology. Similarly, when AI succeeds, ensure teams recognise the system's genuine contributions.

Establish Clear Responsibility Frameworks

The study highlighted a critical "responsibility gap" where accountability becomes unclear in human-AI collaborations. Prevent this by establishing explicit frameworks that define:

• Which decisions AI systems can make independently
• When human oversight is required
• How responsibility is shared for different types of outcomes
• Clear escalation procedures when AI confidence drops below set thresholds

Train Teams on Attribution Awareness

Implement training programmes that help staff recognise and overcome attribution biases. The research suggests that awareness alone can help reduce these psychological traps.

Include modules on:

• How attribution biases affect AI perception
• Techniques for objective performance evaluation
• Methods for recognising AI contributions
• Strategies for balanced human-AI collaboration

Measuring Success in Human-AI Partnerships

Beyond Traditional Metrics

Traditional performance metrics often fail to capture the nuanced value that AI brings to e-commerce operations. Develop measurement frameworks that account for AI's unique contributions:

Process Improvement Metrics:

• Time saved through AI automation
• Decision speed improvements
• Consistency in routine tasks
• Reduction in human error rates

Strategic Enhancement Metrics:

• Quality of insights generated by AI analysis
• Complexity of problems AI helps solve
• Innovation enabled through AI-freed human capacity
• Long-term learning and improvement rates

Satisfaction and Adoption Tracking

The research emphasised that satisfaction with AI systems directly impacts adoption and long-term success. Regularly survey team members on:

• Perceived value of AI contributions
• Confidence in AI recommendations
• Willingness to expand AI use
• Understanding of AI capabilities and limitations

Use this feedback to identify attribution bias patterns and adjust training or communication strategies accordingly.

The Future of E-commerce AI Integration

Moving Beyond Tool Mentality

The most successful e-commerce organisations will be those that view AI as genuine collaborative partners rather than sophisticated tools. This shift requires addressing the psychological barriers that the research identified while building systems that truly enhance human decision-making.

As AI capabilities continue advancing, the attribution challenges will become more complex. Teams must develop frameworks now for recognising and crediting AI contributions appropriately, or risk underutilising increasingly powerful technologies.

Building Competitive Advantage

Companies that successfully overcome attribution bias will gain significant competitive advantages. They'll be able to:

• Maximise ROI from AI investments through higher adoption rates
• Accelerate AI deployment across additional business functions
• Attract and retain talent comfortable with AI collaboration
• Make faster, more informed decisions through effective human-AI partnerships

Conclusion: Practical Steps for Implementation

Start addressing attribution bias in your organisation today with these concrete actions:

1. Audit current AI systems to identify where attribution bias might be occurring
2. Implement transparency measures that clearly explain AI decision-making processes
3. Train teams on recognition and mitigation of attribution biases
4. Establish clear responsibility frameworks for human-AI collaboration
5. Develop balanced metrics that capture both human and AI contributions
6. Create regular review processes that explicitly discuss AI's role in outcomes

The research makes clear that technical AI implementation is only half the battle. The other half—managing human psychology and attribution patterns—determines whether your AI investments deliver their full potential.

By understanding and addressing these psychological challenges, e-commerce managers can build more effective human-AI partnerships, maximise satisfaction with AI systems, and ultimately drive better business outcomes through intelligent technology adoption.

Remember: the goal isn't to eliminate human judgment, but to create balanced partnerships where both human insight and AI capabilities are recognised, valued, and optimally utilised.

Source:

How locus of causality shapes human-AI decision-making Piotr Gaczek, Grzegorz Leszczyński and Mateusz Kot Department of Marketing Strategies, Poznań University of Economics and Business, Poznań, Poland, and Rumen Pozharliev School of Management, Luiss Guido Carli University, Rome, Italy