Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2026
|
| Online Access: | https://doi.org/10.5281/zenodo.20141801 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866902059092017152 |
|---|---|
| author | Azalia, Areta Vania Gunawan, Ali Mulyono |
| author_facet | Azalia, Areta Vania Gunawan, Ali Mulyono |
| contents | <p>In this paper, we carried out a systematic literature review on the evolution of artificial intelligence capabilities in e-commerce, from rule-based automation to agentic systems. Following the PRISMA 2020 framework, 371 peer-reviewed studies published between 2021 and Q1 2026 from Scopus and Web of Science were reviewed. The findings show that e-commerce AI development can be classified into four maturity stages: rule-based automation, machine learning augmentation, generative AI, and agentic systems. The review identifies 23 algorithm families and highlights two major stage transitions, namely the rise of LLM and generative AI capabilities from S2 to S3, and the emergence of multi-agent or agentic architectures from S3 to S4. However, as AI systems become more advanced, production readiness declines, with S4 agentic systems showing high reliance on simulation and limited real-world validation. This study proposes the AI-Commerce Maturity Model (ACMM) as a classification framework for understanding algorithmic paradigm, architectural pattern, and autonomy level in e-commerce AI. In general, the paper argues that the future of agentic commerce should focus not only on capability scaling, but also on verification infrastructure to ensure autonomous systems can operate reliably in real production environments.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_20141801 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | From Rule Engines to Agentic Systems: An AI Capability Maturity Framework for E-Commerce Azalia, Areta Vania Gunawan, Ali Mulyono <p>In this paper, we carried out a systematic literature review on the evolution of artificial intelligence capabilities in e-commerce, from rule-based automation to agentic systems. Following the PRISMA 2020 framework, 371 peer-reviewed studies published between 2021 and Q1 2026 from Scopus and Web of Science were reviewed. The findings show that e-commerce AI development can be classified into four maturity stages: rule-based automation, machine learning augmentation, generative AI, and agentic systems. The review identifies 23 algorithm families and highlights two major stage transitions, namely the rise of LLM and generative AI capabilities from S2 to S3, and the emergence of multi-agent or agentic architectures from S3 to S4. However, as AI systems become more advanced, production readiness declines, with S4 agentic systems showing high reliance on simulation and limited real-world validation. This study proposes the AI-Commerce Maturity Model (ACMM) as a classification framework for understanding algorithmic paradigm, architectural pattern, and autonomy level in e-commerce AI. In general, the paper argues that the future of agentic commerce should focus not only on capability scaling, but also on verification infrastructure to ensure autonomous systems can operate reliably in real production environments.</p> |
| title | From Rule Engines to Agentic Systems: An AI Capability Maturity Framework for E-Commerce |
| url | https://doi.org/10.5281/zenodo.20141801 |