TikTok Shop’s recent retreat from Indonesia puts the spotlight on the fast-evolving e-commerce landscape in Southeast Asia. As a generative AI company deeply embedded in Southeast Asia and partnered with regional e-commerce players, we have a unique vantage point on disruptive technologies reshaping digital commerce.
While TikTok Shop’s setback highlights regulatory hurdles, Southeast Asia’s e-commerce sector still presents vast growth opportunities for e-commerce platforms and sellers. Statista forecasts the SEA’s e-commerce market to exceed USD 210 billion by 2025. Yet seizing this opportunity is not easy. The end-to-end journey faced lots of challenges for the platforms, the sellers and the consumers. With the accelerating numbers of new sellers, the e-commerce platforms usually lack resources to onboard and support their needs. Moreover, Cash-on-Delivery (COD) parcels, despite their prevalence in SEA markets, have a fulfillment rate of only around 70%. Additionally, while Southeast Asian consumers’ enthusiasm for online shopping is high, their dissatisfaction with extended customer service wait times is a significant concern.
In an industry moving at fast speed and with customer expectations ever-evolving, businesses need to be able to keep up in order to succeed. Generative AI-powered tools, although nascent, have emerged an important force and would be indispensable tools for eCommerce businesses to streamline platform operations, accelerate seller growth and improve buyer experience.
Revolutionizing e-Commerce seller operations and support
Over 95% of SEA e-commerce sellers are underserved
Major e-commerce platforms in SEA typically host 200K-600K sellers. Only the top 1-5% sellers receive personalized assistance from account managers. This leaves over 95% of SMB sellers with minimum support.
These underserved sellers are grappling with diverse challenges. For example, new sellers may struggle with navigating onboarding and understanding the platform’s features. Similarly, cross-border sellers may find it challenging to venture into unfamiliar overseas markets. Furthermore, sellers looking to scale their operations often find themselves at a loss, unsure of how to systematically grow their business.
While platforms usually offer self-served text guides crafted by e-commerce professionals, these generic resources can overwhelm sellers. Outsourcing seller operations and support to BPO (Business Process Outsourcing) companies is another common approach. However, the extensive seller base and limited manpower make it challenging for sellers to obtain immediate assistance.
Generative AI for personalized account management with long-tail sellers
Generative AI solutions can revolutionize seller operations by automating customized support across the seller journey, using industry best practices to guide previously underserved sellers. With generative AI-powered voicebots, it is akin to every seller having their own personal account manager, efficiently addressing their specific concerns. This eradicates the need to sift through massive information or endure prolonged waiting times to reach a human agent.
Voicebots handle tasks like calling new sellers and guiding them through the onboarding process, providing inventory management strategies based on trend forecasts, and promoting sales acceleration through platform growth mechanisms(free shipping, cashback), mega sales campaigns and advertising programs. The systematic guidance facilitates sellers’ growth from beginners into mature sellers, aiding them in brand establishment and sales acceleration.
Real-world impact: empowering 35X more sellers and driving sales up by 15%
A top ASEAN e-commerce platform struggled to support their massive seller base with BPO human agents. They partnered with WIZ.AI, and deployed the company’s gen AI-powered voice AI solution, WIZ Talkbot, across three critical seller interaction stages. Talkbots speak local languages and dialects to sellers in diverse markets of the region, such as Bahasa Indonesia, Thai and Tagalog. During the onboarding phase, Talkbots guided new sellers about platform navigation and encourage SKU additions. In the seller growth phase, Talkbots educated existing sellers on new platform features to foster engagement. And ahead of key seasonal promotional campaigns, Talkbots were scaled up swiftly to reach out to massive sellers and encourage participation.
The results were extraordinary: Compared to human agents, Talkbots served 35X more sellers, achieved a more than 30% higher conversion rate for seller participation in mega sales campaigns. These mega sales campaigns led participating sellers to witness an average 15% surge in sales and a 10% rise in basket size. Talkbots create a win-win situation fostering mutual growth for both the platform and sellers alike.
Boosting parcel fulfillment rate with optimized customer engagement
The challenge of high COD delivery failure rate
In Southeast Asia, COD (Cash-on-Delivery) is a dominant payment method. Millions of COD parcels wait for delivery daily. However, the challenge arises from the high failure rate of COD deliveries, which can reach up to 30%. This results in logistics companies bearing the burden of $0.30 per failed delivery.
The root causes of high fulfillment failure rates include package receipt decline, inaccuracies in addresses, and customer unavailability. Therefore, confirming with each customer before parcel delivery becomes indispensable. Equally important is the need to engage with customers after every failed delivery to understand the reasons and decide the next steps.
Logistics firms commonly rely on Interactive Voice Response(IVR) solutions, emails or texts to engage with customers at the two critical moments mentioned above, but the response rate is extremely low, leaving companies in the dark about the reasons for failed delivery.
Enhancing COD delivery success with scalable and interactive voicebots
This is where the scalable and interactive voicebots come into play. These generative AI-powered voicebots not only facilitate large-scale operations but also provide instant data feedback. They can quickly get customers’ responses during phone calls, ascertain which parcels are ready for delivery, which are not, and the reasons behind.
In the pre-delivery stage, AI voicebots undertake the crucial role of confirming order statuses with every customer. Considering the millions of COD parcels queued for delivery daily and the prevalent poor call connection rates within the ASEAN region, the scale of this task is substantial.
Each interactive call made by the voicebots verifies several pivotal details. They confirm the customer’s desire to receive the parcel, ensure address accuracy, check customer availability and obtain payment confirmation. Parcels are dispatched for delivery only after these key details are verified. This systematic approach significantly reduces unnecessary delivery costs and minimizes the occurrence of failed deliveries.
For every failed COD delivery, voicebots would engage with the customer again to comprehend the failure reasons. Based on the provided reasons, they proceed to either cancel orders or schedule redelivery, as per the customer’s preference.
Based on our data, deploying WIZ Talkbots for COD parcel delivery typically results in a 10% rise in delivery success and a substantial 60% reduction in costs, especially when Talkbots work as copilots with human agents.
Delighting customers through generative AI-powered service
Long wait times sabotage customer experiences
To address millions of e-commerce customers’ inquiries, it’s common for platforms to rely on IVR and BPO agents to manage the overwhelming volume of calls. However, this often leads to customer experience that is far from satisfactory. Customers find themselves in a maze of button presses, sorting through various question categories, listening to impersonal responses, or waiting for a long time to be connected to a human agent. Sometimes, the wait time can extend to staggering 72 hours, a delay that leaves customers deeply frustrated.
Enhancing customer satisfaction through immediate, multilingual support
24/7 available generative AI-powered chatbots eliminate the need for pressing multiple buttons and waiting in long phone queues. They offer immediate responses, significantly improving customer satisfaction. Their comprehensive conversational capabilities allow them to handle multi-round dialogues in texts, covering a wide range of questions. Furthermore, these chatbots can support multiple languages, ensuring globally diverse customer bases are catered to.
Real-world impact: Reducing cost by 90%
Picture an e-commerce platform with a BPO team spreading across over 100 countries. They spend a minimum of USD 300K per month on BPO services. When calling BPO agents for support, customers typically face waiting times ranging from 4 to 24 hours.
By implementing a cutting-edge solution like WIZ.AI’s generative AI-powered chatbot, the company can have a unified customer service platform that supports 50+ languages in their target markets. The chatbot can address FAQs, carry out open-ended Q&A in multiple rounds of dialogue, and provide omnichannel support. The result? A significant increase in customer satisfaction scores (Net Promoter Score, NPS) and a whopping 90% reduction in costs.
The future of e-commerce and AI
At present, AI technologies necessitate human orchestration for designing workflow decision trees. Every step of automation is cultivated by humans behind the scenes.
As we look forward, we envision AI technology to automate tasks in the e-commerce spectrum with minimum human intervention. That’s the unprecedented potential LLM agents* can unlock. LLM agents can transcend mere conversational capabilities and apply knowledge to solve tangible problems. Consider an LLM agent assisting with inventory management strategies. It can not only inform sellers of trend forecasts and popular SKUs based on machine learning algorithms but also aid sellers in procuring and restocking relevant SKUs. Furthermore, LLM agents with different roles can collaborate with each other. For instance, a seller operations agent can liaise with a customer service agent, informing customers about hot items that have cashback or free shipping, or trendy items that are newly available in their store. The prospects are limitless!
*LLM agent or LLM powered autonomous agent system consists of an LLM functioning as the brain, and three other crucial components for planning, memory and tool use. In other words, LLM agents possess autonomous decision-making and task execution capabilities.
Initial demo indicates autonomous planning capabilities of LLM agents
While LLM agents are still emerging, we already have early evidence demonstrating their capacity for autonomous planning. One of our experts at WIZ.AI has created an initial demo video for this.
With our meta prompt system, users no longer need to manually create prompts for LLMs to perform tasks. They simply provide an example of an input-output pair, and the system can swiftly generate a prompt capable of managing similar tasks within minutes, as demonstrated with a new math calculation in our video.
Wondering how generative AI can supercharge your e-commerce business ? Speak to one of our specialists today.
Reference
Shopify. (2023). AI in Ecommerce: Applications, Benefits, and Challenges. Retrieved from:
https://www.shopify.com/blog/ai-ecommerce
Carl Bleich. (2023). How Generative AI is Transforming E-Commerce. Retrieved from:
https://learn.g2.com/ai-in-e-commerce
