Using AI on e-commerce analytics: Data quality, availability remain critical obstacles | e27

E-commerce analytics is one of the key areas where artificial intelligence (AI) can be implemented in Southeast Asia. Still, businesses face several key challenges in implementing AI in their e-commerce analytics processes.

“Implementing AI in e-commerce analytics is not without its challenges. Data quality and availability remain critical obstacles. Data accuracy, cleanliness, and consistency are essential for building reliable AI models. Additionally, harmonising data from disparate sources into a cohesive dataset across different platforms remains complex and time-consuming, and AI still can’t help much,” says Annie Yao, Head of Growth, Market Intelligence at e-commerce analytics SaaS platform Flywheel, in an email to e27.

“Another significant challenge is model explainability. As AI models become more complex, understanding how they arrive at their conclusions becomes increasingly important for building trust and ensuring regulatory compliance.”

Yao believes the initial investment in AI technology, infrastructure, and talent “can be substantial.”

“Finding the right partner that can work with you to demonstrate a clear return on investment is essential to secure ongoing support for AI initiatives,” she stresses.

Also Read: Infrastructure, talents are some of the challenges finance industry faces in adopting AI: Provenir

In this interview, she explains how AI is changing e-commerce analytics and what other changes we can expect.

The following is the edited excerpt of the conversation:

What are the most significant ways AI is transforming e-commerce analytics today?

AI is reshaping the e-commerce analytics landscape. One of the most impactful areas is data processing and analysis. AI’s ability to process vast datasets at unprecedented speeds and identify patterns is revolutionising how businesses extract insights.

Additionally, natural language processing (NLP) powered by AI transforms how we derive meaning from textual data. This includes analysing PDPs, customer reviews and social media sentiment to extract valuable insights into areas such as brand communication and consumer perceptions. Moreover, AI’s capacity to process visual information through image and video analysis unlocks new avenues for understanding consumer behaviour and product performance.

Ironically, an increasing amount of this data is now generated by AI. Being able to identify AI-generated content, if you want to exclude it from your analysis, will become increasingly important.

Finally, with the right inputs, companies can leverage AI to build sophisticated models to estimate e-commerce sales, inventory levels, customer behaviour, and market trends based on past signals and outcomes.

Can you share specific examples of how AI-driven analytics have improved business outcomes for e-commerce companies?

Our experience lies primarily in the application of AI, specifically large language models, to enhance natural language processing capabilities. By leveraging AI, we’ve significantly improved the efficiency and effectiveness of any text-based analysis for our clients.

Also Read: How art consultant, online gallery The Artling uses AI to pick the best art pieces for your space

For example, we’ve achieved a 30-40 per cent reduction in project costs and accelerated project timelines by two to four weeks compared to before. Moreover, LLM’s strong knowledge base has brought a fresh perspective to our analysis, uncovering insights that would have been difficult to identify using human-driven methods.

Beyond text analysis, we see immense potential in applying AI to internal knowledge management. Many organisations struggle to leverage their data effectively—there is too much data and too little time. AI can play a pivotal role in transforming vast amounts of information into actionable insights. By automating or simplifying data analysis, summarisation, and presentation, AI can empower employees to make data-driven decisions more efficiently.

How can businesses overcome data privacy and security concerns associated with AI in e-commerce analytics?

Addressing data privacy and security concerns is vital for businesses implementing AI in e-commerce analytics. While e-commerce analytics often involves non-personal data, adopting robust data protection measures when dealing with sensitive information is essential.

Key strategies include data minimisation, collecting only the necessary data, and data anonymisation to protect personal information. Implementing strict access controls and conducting regular security audits are also crucial. By adhering to these practices, businesses should be able to mitigate most privacy risks.

Looking ahead, what emerging trends in AI do you believe will further revolutionise e-commerce analytics in the next five years?

The future of AI in e-commerce analytics holds immense promise. We anticipate significant advancements in AI’s ability to generate insights and recommendations without explicit programming or prompting. AI can help identify issues to be fixed, tell you how to fix them, and, if the systems are connected, fix them for you.

Also Read: One-third of Singaporeans never used AI tools in their workplaces: Survey finds

However, developing the necessary capability to harness AI’s full potential remains challenging. Businesses need talents and partners with a deep understanding of data and, more importantly, their businesses to unleash AI’s full potential.

Additionally, the increasing complexity of the business environment, with many factors influencing outcomes, makes such tasks harder – especially when everyone, including their competitors, uses AI to make competitive moves.

The AI-assisted business wars will only intensify, and those who are not investing enough will be left behind.

—

Image Credit: Flywheel

This article was first published on July 23, 2024

The post Using AI on e-commerce analytics: Data quality, availability remain critical obstacles appeared first on e27.

Source link