Skip to content
Jisan Nuren Siam
Jisan Nuren Siam

Data · Research · Products

All work

Data Science & AnalyticsPython · Research support

AI Adoption in Marketing

The Mediating Role of Perception

A research analytics collaboration: statistical analysis in Python for a study of what drives AI adoption among marketers, and how perception mediates the path from awareness to adoption.

Overview graphic: 250 responses from marketers, a Python workflow, regression and mediation analysis, and a conceptual model with four perception mediators and four external and internal factors leading to AI adoption.
Fig. 01 — Project overview
Role
Research analytics collaboration: resources, analytical support and statistical assistance. The associated report was written by another researcher.
Type
Professional work
Status
Completed analysis
Tools
Python · pandas · statsmodels · Pingouin · matplotlib · seaborn
Respondents
250 marketers
Methods
OLS · mediation
Language
Python
Project type
Research support

Case study

01Question

Drivers, and the role of perception

Two questions: which organizational and market factors predict AI adoption, and do perceptions of AI (usefulness, ease of use, trust, risk) mediate the relationship between awareness and adoption?

02Workflow

A reproducible Python pipeline

  • Data preparation and profiling with pandas
  • Descriptive statistics and normality checks
  • Correlation analysis and scatterplots
  • OLS regression with statsmodels
  • Mediation analysis with Pingouin
  • Visualization with matplotlib and seaborn
Six-step Python analysis workflow from data loading and profiling to descriptive statistics, assumption checks, multiple regression and mediation analysis, with the respondent profile and library stack.
Fig. 02 — Analysis workflow

View full size: Analysis workflow (opens in a new tab)

03Regression

Customer expectations led

The model explained 64.7% of the variance in adoption (R² = .647, adjusted R² = .641; F = 112.3; Durbin–Watson = 1.751). Customer expectations (β = .548) were the strongest predictor, followed by market competition (β = .218) and implementation cost (β = .124), all p ≤ .004. The regulatory and legal environment was not significant (β = −.039, p = .422).

Regression summary with R², adjusted R², F-statistic and Durbin–Watson, and a bar chart of standardized coefficients: customer expectations strongest, regulatory and legal environment not significant.
Fig. 03 — Regression drivers

View full size: Regression drivers (opens in a new tab)

04Mediation

Usefulness and trust carried the effect

Perceived usefulness (indirect effect .102, about 21.6% of the total) and trust (.131, about 27.8%) mediated the awareness–adoption relationship. Perceived ease of use (.072) and risk (−.058) did not. The total indirect effect was about .247 against a direct effect of about .225.

Mediation model table showing indirect effects of four perception mediators; perceived usefulness and perceived trust supported, ease of use and risk not supported, with total indirect and direct effects.
Fig. 04 — Mediation analysis

View full size: Mediation analysis (opens in a new tab)

05Implications

What the evidence supports

Among these marketers, adoption tracked competitive pressure, customer expectations and positive perceptions of AI more than regulation. The data are cross-sectional, so the results describe associations rather than causes.

Executive findings summary listing key results, supported and unsupported hypotheses, and the practical implication for marketers.
Fig. 05 — Executive findings

View full size: Executive findings (opens in a new tab)