The parent of CEWELL (中金智汇) was the big data business division set up in 2014 by “Centrin Data,” a big data and cloud computing company. Back then, “Centrin Data” wanted this business division to vigorously develop cloud services for corporations. In 2017, “CEWELL” began to operate independently after being officially registered as a company. Recently, it received 50 million yuan (approximately US$7.87 million) in financing from Centrin Data.
“Generally speaking, we are defined as a big data application company providing smart services for corporations. The company currently focuses on smart customer service operations, such as smart marketing, smart customer service, and client-side risk control,” said LIAO Changru, CEO of CEWELL.
With regards to products, CEWELL first set about its work encompassing three scenarios of call centers—quality testing, marketing, and operation:
Medium and large enterprises that provide batch services for consumers have relatively big call centers. Service quality monitored and managed by man, requires excessive amounts of work. Additionally, spot checks conducted can easily give rise to speculative behaviour.
CEWELL’s solution is Smart QC, which, featuring speech recognition, semantic analysis, and distributed storage, converting multitudinous customer service recordings into text files and then identifying and analyzing them in the early stages. The system can regulate batch monitoring with a rules engine: For example, by setting up quality checkpoints such as business terminology and polite expressions, the machine automatically scores the business proficiency of all the service personnel at work and later filters out low-score calls, which will be subjected to quality inspection.
Different quality testing methods are applied respectively to different industries. For example, customer sentiments can be monitored through routine consultations, while conformity to regulations is monitored through insurance sales to ensure custom services do not get out of line and mislead customers, etc. Smart QC can make rules in light of industry differences and corporate needs.
In fact, many users make calls to inquire about products. When this happens, whether or not customer service can seize the opportunity to sell seems particularly important. CEWELL’s smart marketing system is aimed at this scenario. It raises relevant products and services and consequently makes the sale materialize through a process in which user questions and preferences are compared with the features of products and services.
The last area is with respect to corporate operations. The performance and service quality of customer service and the work of employees all require a quantifiable indicator. CEWELL prepares statistics on these dimensions, analyzes the data, and then gives them to companies as an overall basis for management.
LIAO emphasized that these application scenarios are mainly supported by two CEWELL platforms: One is a big data platform. CEWELL has optimized its open source system and as a result it can sustain the entire lifecycle of data which includes early-stage acquisition, processing and analysis, and upper-level integration of algorithm models. The other is a text analysis platform which mainly applies self-developed technology to the semantic processing of a huge amount of data in conversations.
Target customers generally come in two categories: companies that provide customer services use it for internal control, and third-party companies providing business process outsourcing (BPO) use it to serve their customers. At present, most of these services are provided to the credit cards, insurance, online consumption, and express delivery sectors.
Currently, revenue is mostly from project-based transactions. The amount depends on data volume and complexity, and millions of yuan are usually involved.
CEWELL has a team of about 60 staff members, most of whom are engaged in research and development; fewer than 10 do marketing. The company shares resources with the group when it comes to sales, technology, and more.
