Recently, while preparing for the AI Application Forum at the CFC conference at the end of the month, I've been engaging in intensive discussions with brand owners about AI applications. The feedback I've received is quite interesting, and I've roughly divided it into three stages. Initial stage: meeting minutes, customer service replies, contract review, bid document generation, etc., which are more administrative tasks. Intermediate stage: internal knowledge bases, image recognition for store inspections, consumer demand insights, sales report analysis, etc., where some small business scenarios start to be applied. From my interactions, most companies are still in the above two stages. Very few companies have embedded AI into frontline sales processes to try to change how salespeople visit stores, promote products, or manage distributors—almost none. This is what I consider the advanced stage of AI application. Last month, BCG released a report titled "How AI is Reshaping Distributive Sales Channels in Emerging Markets." This report doesn't just talk about AI trends in general; it specifically breaks down what AI can do in the FMCG distribution system, how to do it, and the data performance after implementation. I spent some time translating and reading it in depth, and I think there are several insights worth sharing. Especially for sales managers at brand companies, the scenarios described in this report might be the reality we face in the next two to three years, because AI's evolution and capabilities are advancing very rapidly! AI Assistants Change Salesperson Execution Actions Let me start with a common phenomenon that I believe many will relate to. Many brand companies' sales systems appear digitized on the surface. Salespeople have SFA, visit systems, and data dashboards. But if you look closely at frontline execution, you'll find that many actions still rely on human judgment: which stores to visit first today, what products to promote upon arrival, what promotional policies to discuss, and whether to increase SKUs at stores. In the BCG report, this stage is called GTM 2.0: having basic digital tools but inconsistent execution and unstable results.
Note:
GTM 1.0 is manual sales planning and execution;
GTM 2.0 has basic digital systems but unstable execution;
GTM 3.0 is large-scale deployment of digital tools and data analytics;
GTM 4.0 is AI empowering frontline sales, sales management, and back-office operations. What AI can do is give companies the opportunity to leap to GTM 4.0. What does that mean? Let me describe a scenario: a frontline salesperson visits 20-30 stores daily. In the past, they relied on experience to decide visit order and product recommendations. Now, they have an AI sales assistant on their phone. In the morning, it tells them: "You have limited time today; prioritize these 8 stores, and this route is the most efficient." At the first store, the AI reminds: "Last time, this owner complained about two out-of-stock SKUs, which have now been replenished. You should focus on promoting product X because it has a 30% gross margin per case. Here's a suggested script." This isn't a concept; it's a scenario that's already been implemented. In the BCG report, a mid-sized home cleaning brand in an emerging market equipped its salespeople with an AI assistant for order recommendations and lost-order recovery reminders. Within a month of launch, sales increased by 11%. Another consumer goods company used AI for dynamic route optimization and personalized reminders. After deployment, face-to-face time with customers increased by 25%, and the number of product lines sold per visit increased by 8%. Regarding frontline sales visits, I've had discussions with several city managers. Their biggest headache is the varying skill levels of salespeople. Good salespeople rely on experience; poor ones rely on habit. On the same route, some can sell three more products, while others just go through the motions. What AI does is essentially turn your company's best sales experience into a real-time coach on every salesperson's phone. It's not something you forget after training; instead, at each store, there are specific, data-driven suggestions. A couple of days ago, I spoke with a CIO, and we agreed: ideally, we could "distill" the experience of top salespeople and turn it into everyone's capability. This aligns perfectly with the direction of the BCG report. AI Changes Not Just One Action But the Entire Logic of Sales Management We've just talked about frontline salespeople, but if we look from the perspective of sales managers, AI might change much more. The BCG report outlines ten AI application scenarios covering planning, execution, and operations. For sales managers, I think two directions are particularly noteworthy. The first is "store-specific" sales planning. In the past, brand companies made sales plans by region or channel, offering a standard solution. But in reality, each outlet varies significantly: different sizes, customer bases, competitive environments, and historical sales shares. AI can generate a customized sales plan for each store based on store profiles and sales history: what to promote, what scripts to use, and what promotions to offer. In the past, only top sales managers could do this based on experience, but now AI can do it at scale. The second is personalized performance management. Traditional sales management sets uniform goals, fixed incentives, and one-size-fits-all training. But the gap between salespeople is often large. The report describes a scenario where AI can analyze each salesperson's conversion rate, visit efficiency, category coverage, and other metrics, then provide personalized goals and training suggestions. For example, it might tell a salesperson: "Your conversion rate is 46%, while the city average is 70%. The issue lies in the product recommendation stage after entering the store. Here are two actions you can practice." What does this mean for sales management? In the past, a regional manager overseeing three cities could barely manage by holding weekly meetings, reviewing reports, and spot-checking visits. If AI can identify each salesperson's problems in real-time and provide targeted improvement suggestions, the regional manager's management radius can expand. Of course, AI applications also include other areas, such as sales support tasks to improve efficiency. What Are the Barriers to AI Sales Applications in the FMCG Industry? After discussing what AI can do, the reality is that we often know it's important but don't know where to start. It feels like AI is "omnipotent," but when it comes to specific business scenarios, something always seems off. From my exchanges with brand friends, AI has been fully applied to internal contract review, bid document generation, image recognition for store inspections, meeting minutes, etc., but they rate it as "below average." Most scenarios are about improving administrative efficiency, and truly high-value applications on the sales side haven't been fully found yet. The BCG report also mentions key obstacles to AI adoption, with three core points: 1. Lack of a clear AI strategy. Many companies are running pilots everywhere—today an AI customer service, tomorrow an AI video—but without a coherent approach. What business problems should AI solve, and what commercial value should it create? There's no clear AI value planning. Of course, this is understandable, as AI has only become widespread in the past year or so, and exploration takes time. But I believe that by the second half of the year, there should be some clarity on AI strategy. 2. Weak data foundation. The ceiling of AI is never determined by the model but by the company's internal data assets. For most companies, it's not that they lack data, but that data is scattered across ERP, DMS, SFA, e-commerce backends, Excel spreadsheets, and WeChat groups. Without integrated data, AI can only do superficial work. 3. Difficulty in frontline adoption. The BCG report mentions a phenomenon: even when AI tools are available, teams still revert to traditional workflows. The technology is ready, but the organization isn't. Incentive mechanisms aren't aligned with AI, and salespeople feel that AI recommendations differ from their habits, so they don't use them. After a round of discussions, I have a feeling: the real challenge with AI isn't the technology itself, but the restructuring of sales processes. Every organization has inertia. If you don't break the workflow, you're just adding an AI tool. Final Thoughts Undoubtedly, AI is iterating rapidly, and the next stop for AI implementation in the FMCG industry will definitely be the sales system. In the past, FMCG companies relied on people to run, monitor, and manage sales. In the future, AI won't make these experiences disappear, but it will reorganize them. It will let salespeople know which stores to visit today; let city managers know where market problems lie; and let managers know where resources should be invested. So at this CFC conference's AI Application Forum, what we most want to discuss is not "whether FMCG companies should use AI"—that question no longer has much discussion value. What's truly worth discussing is: Which scenario should AI enter first? Who takes the lead? How do sales, digital, and IT departments coordinate? When data is incomplete, can we start with a small scenario and get it running? For many years, the FMCG industry has relied on people to push products to terminals case by case. In the future, AI may not replace these people, but it will redistribute their experience, actions, and management methods. For brand companies, when AI leaves the office and enters the market, that's when it truly starts to impact business. Finally, regarding AI implementation practices in the FMCG industry, "New Distribution" will host the "AI Application Forum" in Hangzhou on May 27-28, featuring case sharing and peer exchanges on specific AI application scenarios for brand owners and distributors. For detailed agenda, please contact staff via the conference QR code.