When a company rolls out a shiny new AI tool like Baidu’s ERNIE chatbot or plugs advanced analytics into its supply chain, the pitch to executives usually goes something like this: the technology will make workers smarter, faster, and more creative. But anyone who has watched a promising software rollout fizzle knows that fancy tools do not automatically produce inventive employees. So what actually converts a data-rich workplace into a creative one?
A study published in the Journal of Business-to-Business Marketing takes on that question by examining Chinese business-to-business (B2B) firms adopting generative AI tools. The researchers find that big data-driven supply chains do help employees develop creative skills, but the effect runs through two very human ingredients: motivation and resilience.
The question behind the research
Yongming Tao of Dongbei University of Finance and Economics and colleagues from Xi’an Jiaotong University, the University of Bradford, and the Getulio Vargas Foundation wanted to understand what happens inside the heads of workers when their companies build sophisticated, AI-augmented supply chains. The technology part is well documented. What is less well understood is how everyday employees translate all that data into fresh ideas.
The team drew on social learning theory, a framework developed by psychologist Albert Bandura in the late 1960s. The core idea is straightforward: people pick up new behaviors by watching others, imitating what works, and adjusting based on feedback. The researchers argue that an AI-enhanced workplace is essentially a giant social learning environment, where employees observe how colleagues and AI systems tackle problems and then borrow those approaches.
But observation alone does not produce creativity. The researchers propose that two psychological factors do the heavy lifting. The first is what they call hybrid task motivation: a mix of internal drive (finding the work genuinely interesting) and external rewards (recognition, better performance metrics) that keeps employees engaged with both traditional tasks and AI tools. The second is hybrid resilience: the capacity to keep going when the AI spits out something confusing, when systems glitch, or when the pace of technological change gets overwhelming.
How the study worked
The researchers gathered survey data from 580 senior employees across 116 Chinese B2B firms that were using generative AI tools, particularly Baidu’s ERNIE chatbot. The companies came from three provinces chosen for their heavy involvement in AI adoption: Guangdong, Liaoning, and Shandong. The firms worked in information technology, chemicals, electronics, and plastics and rubber manufacturing.
To reduce the risk that a single survey moment would distort the results, the team collected data at three points in time, spaced roughly a week apart, between November 2023 and March 2024. A team of 22 interviewers, including professors and graduate students, provided on-site support. Employees rated statements about their firms’ data-driven supply chain practices, their own motivation and resilience in working with AI tools, and their creative output on a seven-point scale.
The researchers then used statistical modeling to test whether the pieces fit together the way their theory predicted.
What the analysis revealed
The first finding was fairly direct: employees who saw their firms as having strong big data-driven supply chains reported stronger creative skills. In practical terms, working in an AI-augmented, data-rich environment was associated with generating more novel ideas and suggesting new ways to improve work.
But the more interesting result concerned the pathway. Big data-driven supply chains were linked to higher hybrid task motivation, and higher motivation was in turn linked to more creativity. The researchers interpret this as evidence that the technology itself does not directly make people creative. Instead, working in a data-rich environment tends to energize employees, and that energy is what fuels inventive thinking. Without the motivational spark, the data insights sit there dormant.
The third piece involved resilience. The connection between data-driven supply chains and motivation was stronger for employees who scored high on hybrid resilience. Workers who bounced back from setbacks, treated challenges as learning opportunities, and adapted to unexpected situations got more motivational lift from their AI-enhanced environments. Less resilient employees appeared to get bogged down by the complexity and uncertainty that AI tools can introduce, weakening the chain from technology to motivation to creativity.
The researchers frame this as a layered process. Technology sets the stage. Resilience determines how well an employee can absorb what that stage offers. Motivation converts that absorption into creative action.
What it means for managers
The authors argue that firms hoping to squeeze creative performance out of their AI investments cannot stop at buying the software. They recommend pairing technology rollouts with programs that build employee motivation and resilience.
On the motivation side, this might mean role-specific AI training, transparent communication about how AI tools are meant to support rather than replace workers, and opportunities for employees to experiment with tools like ERNIE in low-stakes ways. On the resilience side, the authors suggest creating psychologically safe environments where experimentation and failure are treated as normal parts of learning, along with structured feedback loops, peer mentoring, and scenario-based training that exposes employees to shifting decision-making conditions.
The researchers also note that one size does not fit all. Smaller firms with limited digital infrastructure might focus on foundational AI training, while larger firms can invest in analytics-based resilience programs. Manufacturing firms might emphasize operational resilience, while service-oriented firms might prioritize flexibility in digital communication.
Caveats worth keeping in mind
A few limitations shape how the findings should be read. The data come entirely from Chinese B2B firms in four industries, so the patterns might look different in service industries, in Western firms, or in cross-cultural contexts. The study relied on employee surveys, which capture perceptions rather than direct observation of creative behavior, and self-report data always carries the risk that respondents describe themselves more favorably than reality warrants.
The design was also correlational rather than experimental. While the researchers used time-lagged surveys to strengthen their inferences, they cannot definitively prove that big data-driven supply chains cause motivation, or that motivation causes creativity. The relationships are associations that are consistent with the proposed model, not experimental proof of causation.
The measurement of big data-driven supply chains was also based on individual employees’ perceptions of their firms rather than on independent firm-level metrics, which the authors acknowledge as a constraint. They flag multi-firm, multilevel studies as a direction for future work, along with qualitative research that could observe how employees actually interact with tools like ERNIE in daily practice.
