The Full-Scale War of Conversational Open Platforms: Technology or Business Competition?
-
The competition in conversational open platforms—is it determined by technology or business?
After the WeChat AI team reintroduced its latest product, the 'WeChat Conversational Open Platform,' at the 2019 World Artificial Intelligence Conference (WAIC), the field of intelligent conversation has become even more lively. WeChat stated that this platform, launched in May, will be open to individuals, businesses, or organizations with conversational or customer service needs, allowing users to create and customize their own conversational AI bots based on their requirements.
Tencent's bold moves in the conversational open platform space are not unique. Looking at current developments both domestically and internationally, from major internet tech brands like Microsoft and BAT to startups specializing in AI conversation, everyone is developing their own proprietary conversational open products.
Large and small companies have already entered the fray, with a plethora of conversational open platforms or AI bots, each showcasing unique AI features, creating a fiercely competitive landscape.
However, this may not just be a battle of technology.
The explosive growth of conversational open platforms at this moment is no coincidence.
The surge in users brought by mobile internet, coupled with the rising demand for new consumption and services, has intensified users' pursuit of personalization, ultimately leading to contradictions. The segmentation and personalization of needs in a massive user base are ushering in an era where 'everyone will have their own personalized conversational AI bot,' and this has spurred the emergence of numerous conversational open platforms serving this goal.
Laziness has always been a key driver of technological advancement. Users, spoiled by the thoughtful designs of mobile internet products, have entered a state of being pampered—'clothes come when stretched out, food comes when opened up.' So-called precise recommendations, one-click completions, and voice-only interactions are all results of laziness.
The consequence is that major products must not only cater to a vast user base but also address each user's 'lazy demands,' creating a natural conflict between standardization and personalization in terms of labor costs and resource investments.
User 'laziness' manifests in two ways, which are also the main focuses of intelligent conversational platforms:
1) Horizontal choices are too broad; users are too lazy to choose and need AI recommendations.
For example, when a user tells the system they want to eat Sichuan cuisine, the conversational system automatically recommends the most likely dishes based on the user's preferences and completes mechanical actions like placing an order.
2) Vertical experiences require multi-step operations; AI conversation is needed to simplify the process.
For example, when a user wants to pay their phone bill, older systems required navigating through multiple service levels. Now, the user can simply say, 'I want to pay my phone bill,' and the system automatically brings up the payment interface, eliminating the hassle of multiple steps.
These are just two simple examples, but so far, any valuable intelligent conversational application revolves around these two directions.
It is precisely because of their maturity in addressing these pain points that conversational open platforms are gradually becoming a mainstream trend.
At the same time, almost all major internet companies are talking about AI and To B (business-to-business) services. Conversational platforms, which combine AI and To B, have become the best way to stake a claim in this space.
Baidu's UNIT, Alibaba's Xiaomi, and Tencent's WeChat Conversational Open Platform have all taken the lead in the intelligent conversation field, laying the groundwork for future AI scenario empowerment.
This is not hard to understand. The smooth transition from consumer internet to industrial internet highlights the advantages of conversational platforms. The shift from To C (consumer) to To B inherently relies on bidirectional connections between users and enterprises, and conversation is one of the best choices.
This is especially true for Tencent. While pivoting to To B, the open platform itself amplifies Tencent's strength in To C connectivity.
Users have lazy demands; tech giants have growth demands—it's a perfect match.
Stimulated by user needs and new technologies, the development of conversational open platforms has led to a competitive trend. Each company's strengths, influenced by business strategies and corporate philosophies, present unique highlights.
Although the WeChat Conversational Open Platform follows a standard open-platform model, WeChat remains its biggest asset.
As we know, behind WeChat lies a vast ecosystem comprising subscription accounts, service accounts, mini-programs, and more, connecting most users, businesses, and third-party service providers in the mobile internet—almost synonymous with mobile internet itself.
Within this massive system, Tencent's advantages are evident. With proximity comes opportunity—simply deploying service accounts and mini-programs allows Tencent to quickly capture a significant share of the AI conversation market, securing a strong starting position.
For Alibaba, commercial scenarios based on e-commerce operations are its foundation and strength.
Recently, Alibaba launched an AI customer service bot—the new 'Ali Xiaomi'—for Taobao and Tmall users. Data from 2017 shows that, thanks to widespread use in e-commerce scenarios, Ali Xiaomi handles an average of 2 million conversational turns per day.
In the vertical business domain, Alibaba's conversational open platform can be directly applied to commercial clients, providing ample business scenarios for AI training and data collection. By first deepening its presence in business scenarios and then expanding horizontally to other fields, Alibaba's conversational open platform is carving a clear development path.
Baidu's conversational open platform advances along two parallel tracks.
On the technology front, Baidu's reputation as the 'military academy of internet technology' is no exaggeration. UNIT's core technologies—demand understanding, conversational control, natural language processing, and knowledge mining—objectively hold a certain technical lead.
On the scenario front, as an underlying technology, conversational open platforms often pair with voice assistant platforms for easier implementation and monetization. Baidu's DuerOS has been in the market for a while, with partners like Skyworth TVs already using 'Xiaodu Xiaodu' as a wake word.
Microsoft's Xiaoice, one of its three global AI product lines, has iterated to its seventh generation, with increasingly refined technical capabilities.
In its core conversational engine, Xiaoice has demonstrated distinct leadership across all seven iterations. For example, the first generation focused on making conversations interesting, the fourth on emotional resonance, and the seventh on breakthrough engagement—each upgrade enhancing its conversational prowess.
Through this process of technical accumulation and iteration, Xiaoice has transitioned from passive to主导对话 (leading conversations), increasingly embodying human-like attributes and showcasing Microsoft's technical strengths in AI conversation.
While it's hard to compare whose technology is superior, Microsoft's conversational platform may offer a better user experience in interactions.
The market expansion playbook for AI conversational platforms is no different from other AI scenarios: rely on a few major clients for branding and the long-tail market for scale.
Major clients often develop their own solutions, with collaboration models leaning toward deep customization. While these cases carry weight, they may not be the key to market evolution.
The 'thick long-tail market'—comprising numerous small and medium-sized enterprises (SMEs)—is the main battleground for determining the scale (and success) of AI conversational open platforms, as well as a primary source of confidence for舆论 (public opinion) and资本市场 (capital markets).
Just as B2B enterprises face both scale and个性化需求 (personalized demands) from B2C users, the 'thick long-tail market' for AI conversational open platforms also grapples with矛盾 (contradictions): a large number of enterprises, each with unique (and sometimes entirely different) needs.
What these SMEs want is the focal point of competition among conversational open platforms:
Few are like the酷炫程序员 (cool programmers) who can type out lines of code while wearing sunglasses. In the long-tail market for AI conversation, most SMEs are小白客户 (novice clients) with limited coding skills, let alone AI expertise. Lowering the deployment barrier for these clients (e.g., chain restaurants, small online stores) is the first hurdle.
A WeChat Conversational Open Platform representative mentioned efforts to simplify deployment—like drag-and-drop interfaces—making it as easy as installing software on a PC.
This is a shared goal for all conversational open platforms.
AI离不开数据 (cannot exist without data). Even after technical empowerment, AI conversational bots often require B2B clients' user data for training.
Small and medium-sized enterprises (SMEs) often possess limited data, or if they do, it is not systematically stored. Consequently, AI providers are competing to develop solutions that can operate with high accuracy using minimal user-provided data.
Just as Newton's three laws of motion explain macroscopic movements, scientists continue to seek a single universal law that can explain the entire cosmos. Similarly, in the realm of AI dialogue, there remains an unmet need for algorithmic models that require less data and possess deeper adaptive capabilities.
"I need it now" reflects the urgency many SMEs feel in today's competitive market. The speed of cold-start for dialogue platforms is indeed related to data requirements, but the focus here is on the platform's built-in capabilities.
Although SMEs have diverse needs, they can generally be categorized. Businesses of similar types share common models, intents, entities, functions, templates, and dialogues even before data training begins. The technology provided by platforms requires training but should not start from scratch—it must understand logic and context, adapting to customer needs with minimal data-driven adjustments.
These built-in capabilities often stem from a platform's accumulated experience with enterprises and products. When opened to external users, they become "hidden academic potential" that comes as a bonus.
Additionally, a platform's ability to self-evolve is a critical competitive factor. After all, everyone wants their AI dialogue system to grow smarter over time, not remain stagnant.
The exposure of Google Duplex's造假 by The New York Times reignited discussions about the applicability of AI dialogue systems.
AI dialogue faces two gaps:
Currently, the latter is the primary battleground for platforms, with intelligent customer service dialogues being the most developed area.
However, it must be acknowledged that among today's AI dialogue platforms, technical differences are incremental—no single platform holds an absolute advantage in achieving human-like fluency, richness, and controllability. The qualitative leap remains elusive.
Everyone has progressed from 0 to 1, but none have completed the jump from 1 to 10—This results in B2B applications showing limited sensitivity to technologies that haven't achieved qualitative breakthroughs. Consequently, competition among major AI dialogue platforms ultimately boils down to the strength of their commercial resources.
In essence, despite the hype around technology and user demands for deployment, data volume, and cold-start capabilities, the short-term competition among dialogue platforms remains centered on commercial resources rather than technological superiority.
Thus, the near-term market landscape will continue to be shaped by the existing commercial domains of tech giants—Alibaba's e-commerce ecosystem, Tencent's WeChat environment—with no strict correlation to technical prowess.
In this context, products like WeChat, with their vast long-tail B2B client resources, immediately demonstrate the commercial value of dialogue platforms. WeChat's extensive user base and cross-industry coverage give its open dialogue platform a significant competitive edge from the outset, making its push into this space unsurprising.
In the short term, WeChat may lead the race, but technology will undoubtedly remain the core differentiator in the long run. However, a purely technology-driven victory is not yet visible on the horizon.