AI Scenario Analysis: The Three Values of Pre-Sales Customer Service Robots in E-commerce
-
What makes a good AI scenario? How does a pre-sales customer service robot perform in e-commerce? The author explores these two questions by analyzing three values in a B2B context and shares insights.
Product managers often talk about scenarios, so before diving into this AI scenario, let’s briefly define what a scenario is: who does what, when, where, and what they need.
So, what constitutes a good scenario? After some research, I found many standards, with experts offering diverse opinions. Drawing from three years of B2B experience, I’ve summarized that a good B2B scenario should embody three key values:
PS: This article doesn’t address whether a business should pursue a particular initiative. To answer that, we’d need to analyze competitors, internal resources, market competition, etc. Here, we focus on the value of a scenario for product managers in implementation, leaving other factors aside.
Over the past few years, AI has surged in popularity, but recently, the hype has cooled, and even previously enthusiastic investors have grown cautious. From AI’s perspective, there are two reasons:
To put it simply, if AI is a hammer, there aren’t many nails that truly need it. Many problems that appear to be nails are actually screws, solvable with a simple screwdriver.
Today, I’ll discuss the AI pre-sales customer service scenario, focusing on integrating customer service robots into e-commerce workflows. Unlike generic customer service robots that aim to "cut costs"—a hard sell to B2B clients—this scenario is less challenging in e-commerce. Let’s examine the customer profile:
Based on the profile and scenario definition, the AI pre-sales robot scenario can be understood as: an e-commerce pre-sales customer service robot responds to buyer inquiries, guides them to make purchases, and assists in improving conversion rates.
Now, let’s explore its performance through three lenses: commercial value, customer value, and product value.
For AI to succeed, it must deliver value to users. In B2B products, this boils down to two goals: efficiency and revenue growth.
In the e-commerce pre-sales scenario, we can outline the pain points of different roles in a business and assess whether the robot can address them.
The AI pre-sales robot can help stores in three ways:
Commercial value, broadly speaking, is about driving business growth and revenue for B2B companies. Narrowly, it’s whether clients recognize and pay for that value.
First, the big picture: if the market size for a service provider is small, it limits the company’s growth potential. Additionally, the scenario must be replicable and scalable to deliver commercial value. If there are only a handful of similar clients or each requires heavy customization, it’s a tough business.
Consider this data: according to iiMedia Research’s "2019 China E-Commerce Semi-Annual Development Report," China’s mobile e-commerce user base is nearing 700 million, indicating massive demand.
On the supply side, take Taobao as an example. Its 2018 report shows 437,000 sellers with annual revenues exceeding ¥1 million, equivalent to 437,000 CEOs, with 2,252 surpassing ¥100 million.
To achieve such results, pre-sales teams handle enormous traffic, requiring robust staffing—and inevitably facing the aforementioned pain points. Tracing this, the AI pre-sales market holds significant potential.
Moreover, e-commerce sellers often follow standardized workflows, especially within industries like appliances, footwear, or skincare. This makes the scenario replicable. Standardization will be discussed later under product value—patience!
B2B requires clear ROI calculations for clients, whether in revenue growth or cost savings. If unclear or unacknowledged, deals fall through. Non-pre-sales customer service robots often struggle here.
Current NLP capabilities can only partially address user queries, leaving many to human agents, making immediate cost savings hard to demonstrate. Even with formulas like Total Value = Reduced Customer Service Labor, clients aren’t compelled to buy.
Some argue large clients invest for long-term ROI, but scaling efficiency requires internal optimizations and system integrations, leading to custom projects—hard to replicate. With few large clients, competition is fierce. Small clients, meanwhile, prioritize revenue over savings in tough times.
The AI pre-sales robot, however, offers real-time responses and smart replies to save costs while boosting conversions and revenue—a beacon for clients in a downturn.
This transforms the equation into one even Gauss would applaud:
Total Value = Increased Conversion per Query + Additional Revenue + Reduced Labor Costs
Here, even if labor savings aren’t immediate (due to robot maintenance), marketing impact is instant. Quick revenue gains + long-term efficiency prove clients’ foresight—so: folks, buy it!
With customer and commercial value established, the third factor is product value: can it be a great product?
What makes a great product? It’s complex, but key criteria include: reusability, standardization, ROI, and scalability.
Data is to AI what fuel is to cars, electricity to devices, and boba tea to me! The pre-sales scenario not only accumulates data but ensures high-quality feedback.
How? Good cases are typically measured by user reviews (rare) or manual labeling (costly). Here, purchases serve as the metric, letting robots learn and improve in a closed loop:
User Query → Robot Judges Purchase → Business Iterates → Smarter Robot
This builds industry-specific data maps, reusable across similar sellers (standardization).
Earlier, I hinted at product standardization. As noted, e-commerce workflows are similar within industries, especially for specific product categories. Thus, we can:
- Build industry/vertical knowledge graphs for smart replies.
- Develop AI marketing packages based on category strategies.
Once standardized, the AI solution can be deployed across similar sellers—smooth sailing!
The input-output ratio here refers to whether the service provider's solution can truly attract customers after implementation. On a deeper level, it's about whether it can help users solve problems and close deals, realizing the promised customer value rather than just making empty promises.
In AI pre-sales scenarios, customer inquiries are limited to specific areas such as products, logistics, promotions, and negotiations. Given the current level of NLP technology, satisfactory response effectiveness can be achieved. Moreover, buyers usually consult with purchase intent, so an AI bot only needs to provide answers and skillfully guide them (intelligent marketing), making the conversion rate much higher than randomly asking passersby if they need fitness services.
Scalability means that a company can derive a series of new products and services from its core competitive products to meet customer needs, thereby driving greater business growth.
A small and refined product is a good path, but if the boss wants to explore another route with more potential, a mere customer service tool won't suffice.
Fortunately, if a product manager starts working on AI pre-sales scenarios and the boss expects a grand vision, it's not impossible. Here, we expand based on actual e-commerce scenarios and briefly list some new services:
After elaborating on the three aspects of value in AI pre-sales scenarios, it's clear that this is indeed a promising opportunity worth pursuing.
Of course, this isn't the main point. The lengthy discussion aims to help us reflect on what makes a good AI scenario and understand its value to ensure successful implementation.