INTERNATIONAL RESEARCH / PEER-REVIEWED PAPER

How should your marketing team work with AI?

Giving a marketing team an AI tool does not automatically improve its work. This study examines how employees work with AI and how that relates to their task performance. Interviews and surveys find a positive relationship, but do not prove that AI alone causes better performance. Our marketing recommendation at DijitalPi is to agree what AI will do, who will check its output and who will make the final decision when preparing an advertisement. Then track both time and errors.

ACADEMIC WORK REVIEWED

Dancing with AI: how human-AI interaction affects employee task performance

ResearchersYepeng Wu, Yuanyuan Jiao, Ping Li, Yujie Liang.

Humanities and Social Sciences Communications · 9 September 2026 · 13:1578 · Peer-reviewed; version of record 17 September 2026 · Source language: English

Open the original publication Find on Google Scholar

ENGLISH REVIEW AND COMMENTARY: DIJITALPI

We explain the paper in plain English, discuss what it may mean for organisations and identify our own examples separately.

Read the explanation Go to 3 business examples

English review: 21 September 2026 · The source date appears in the citation above.

Editorial review draft

The source has been checked; the review still carries its original editorial status. DijitalPi did not conduct a customer experiment for this paper.

LET'S READ THE RESEARCH TOGETHER

We first explain the researchers' question, method and findings. We then discuss how to interpret the results, clearly separating DijitalPi's commentary from the source.

01 / WHAT DID THE RESEARCHERS WANT TO UNDERSTAND?

How can workplace AI interaction be measured?

The paper studies employee–AI interaction and task performance.

Explaining the context · DijitalPi commentary

DijitalPi commentary: Providing a tool does not automatically allocate responsibility. A team may know who generates a draft while remaining unsure who corrects an inaccurate product claim. Treat adoption as workflow design as well as tool training.

02 / HOW WAS THE RESEARCH CONDUCTED?

How did they test the question?

Interviews and surveys in China develop a 19-item scale; there is no randomized experiment.

Understanding the method · DijitalPi commentary

Understanding employee experience through a survey does not establish why an advertising campaign performed better. In your pilot, keep staff feedback separate from task records such as delivery time, corrections and accepted output.

03 / RESEARCH FINDINGS

Association does not establish causation.

A positive association is reported, alongside role identity and self-efficacy.

Source: Original paper: methods, sample and limitations ↗

04 / DIJITALPI'S EXPLANATION

How should we interpret these findings?

DijitalPi commentary: Make acceptance criteria concrete. Replace “good copy” with checks for product facts, verifiable sources and appropriate brand tone.

Low tool usage is not automatically poor performance. Preparation may dominate one task and review another. Avoid turning usage counts into performance ratings without considering responsibility and output quality.

05 / CONCLUSION AND OPEN QUESTIONS

What did we learn, and what do we still not know?

The peer-reviewed paper does not prove a fixed productivity gain for every marketing team.

Published 9 September 2026; interviews occurred in July 2024. Our question is which current workflow needs fewer corrections in your team. DijitalPi has not run this pilot.

Return to the researchers' original publication

Original paper: methods, sample and limitations

DIJITALPI'S APPLICATION COMMENTARY

How can we use this in marketing?

We created these scenarios to make the topic concrete. They are not cases from the paper or measured client results.

EXAMPLE 01

Content: everyone drafts, who approves?

AI drafts multiply while the editorial review queue grows.

Does draft volume measure team productivity?

Open the recommendation for example 01

Measure accepted output and total effort.

Assign a writer, source checker and final approver. Record accepted pieces, revision rounds and active work time. Note differences in task difficulty and avoid assuming that faster first drafts mean faster delivery.

EXAMPLE 02

Advertising: advice or automatic action?

AI suggests budget changes from a weekly report, but nobody owns the decision.

Can responsibility be assigned to the tool?

Open the recommendation for example 02

Keep budget decisions owned and documented.

Check the data period and missing inputs. Record who accepted or rejected each suggestion and why. Track false alarms, decision time and the target metric after implementation. Do not turn the pilot into unrestricted spending changes.

EXAMPLE 03

Customer communication: fluent but unchecked

The assistant writes a confident answer about campaign conditions and the team sends it without review.

Has fluency become a substitute for correctness?

Open the recommendation for example 03

Define when to hand the task back to a person.

Verify dates, stock and returns against approved information. Specify the owner for missing facts. Review wrong answers, unnecessary handovers and correction time using example records. Do not evaluate staff solely by how often they use the tool.

TRY IT WITH YOUR TEAM

Make responsibility visible in one workflow.

Document AI contribution, human checks and the final decision for one marketing task. Ask staff which step remains unclear. Review time, errors and accepted output together. DijitalPi has not yet run this proposed pilot.

Discuss your team’s AI workflow

Source and ownership

Yepeng Wu, Yuanyuan Jiao, Ping Li, Yujie Liang.
Hubei University of Technology; Hubei Digital Industrial Economy Development Research Center; Nankai University; China University of Geosciences.

Humanities and Social Sciences Communications · 9 September 2026 · 13:1578 · Peer-reviewed; version of record 17 September 2026. Open the original publication · Find on Google Scholar

The academic work belongs to the researchers named above. This page contains DijitalPi's explanatory review and original business examples; it is not a full translation of the paper.

← Return to all research articles