Evidence-Aware Framework

Human\u2013AI Collaboration in Accounting

Two complementary frameworks \u2014 HACAS and ATLAS \u2014 for understanding how AI transforms accounting practice, and how ApeX aligns with the evidence.

Validated Instrument

The HACAS Scale

The Human\u2013AI Collaboration in Accounting Scale (HACAS) is a validated 20-item instrument designed to measure the quality and impact of human\u2013AI collaboration specifically within accounting practice.

Collaboration

Measures the nature and extent of interaction between the accountant and the AI system — how actively the human engages with, guides, and responds to AI outputs.

How would you rate the quality of your interaction with ApeX AI across your daily accounting workflows?

Professional Judgement

Assesses how the collaboration affects the accountant’s ability to apply professional judgment — the contextual reasoning that standards and ethics demand.

Does working alongside ApeX AI sharpen or dull your professional judgment on ambiguous matters?

Accounting Performance

Evaluates the impact on task performance and efficiency — speed, accuracy, and throughput of accounting work.

How has ApeX AI changed the speed and accuracy of your core accounting tasks?

Decision Quality

Measures the effect on the quality of decisions made — from tax treatment choices to advisory recommendations.

Are the decisions you reach with ApeX AI better-reasoned than those you would reach alone?

Psychometric Properties

287
practising accountants
75.7%
variance explained (EFA)
CFI 0.96
model fit (CFA)
\u03b1 0.908 – 0.929
Cronbach\u2019s alpha range

Model fit: \u03c7\u00b2/df = 1.84, RMSEA = 0.054

Nomological Model \u2014 How Collaboration Influences Outcomes

Collaboration\u2192Professional Judgement\u03b2 = 0.444
Collaboration\u2192Accounting Performance\u03b2 = 0.467
Collaboration\u2192Decision Quality\u03b2 = 0.59

The effects on performance and decision quality were partially mediated by professional judgement, suggesting collaboration improves outcomes partly by sharpening judgment.

The full 20-item HACAS instrument is embedded in the original publication and its specific item wording is not reproduced here. Only the four confirmed dimension names and independently written reflective prompts are used. The exact Zenodo record for this scale was not independently locatable through public search at time of writing; the values above are presented as reported in the source briefing. The closest related peer-reviewed study (Shaikh & Asif, RELATE, 2026) examines human–AI collaboration in accounting with a sample of 277 accountants and reports a positive correlation between augmentation-oriented collaboration and decision-quality (β = 0.41, p < .001).

Source References

Human–AI Collaboration in Accounting: A Quantitative Analysis of Generative AI’s Effects on Professional Judgement, Decision-Making, and Financial Reporting Quality

Shaikh, F., & Asif, M. (2026). RELATE Journal.

Augmentation-oriented collaboration positively correlates with decision-quality (β = 0.41); substitution-oriented collaboration does not (β = 0.07). Automation bias negatively moderates professional scepticism.

DOI: 10.26710/relate.v8i2.3955 \u2192

Educational content. The HACAS scale and ATLAS framework are academic instruments; ApeX interpretations are hypotheses unless explicitly labeled research-grounded. Not professional tax or accounting advice.