Customer cases

Use cases

Manufacturing

Waste reduction

Case Any production process generates waste – either because of non-conformity with quality requirements or because of using more inputs than necessary. The golden batch is the optimal way to produce a batch. Applying AI and data technology raises that bar.

Solution Implement centralized data collection of the production process to ensure all relevant data are captured and are easily accessible for analysis. Based on these data, build digital and AI models of the production process: the digital twin. The insights can be used to advise human operators on how to control production or they can be applied to intervene automatically.

Impact AI and more data generate improved insights, allowing for better control of the production process. This better control facilitates less quality issues and less wasted inputs. The golden batch norms are sharpened once again.

‘Creating a digital twin allows for better control of the production process‘

---

Opportunity detection

Case When deciding to pursue a sales opportunity, sales executives typically try to estimate where to focus their efforts based on the size of the deal, the likelihood to win and the effort required to win. Small deals with a lot of effort to win will most likely not be pursued. For some organizations, however, these long tail opportunities could mean significant additional revenue. What if the effort required to win could be reduced by using AI technology?

Solution A detailed client analysis shows that most effort is required to identify the opportunity and to select the right portfolio item to offer. For both efforts, we start a separate advanced analytics track simultaneously. The first track focuses on opportunity identification through using data, data analytics and AI technologies. The other track uses the same approach to select the right portfolio item. To make life easy for the sales executive, the results are automatically added to his sales opportunity system.

Impact This will differ per client and industry. Results achieved at a large B2B manufacturing company demonstrated an increase in revenue of 10%.

---

eCTD creation

Case Generation of eCTD (electronic Common Technical Documents) requires a lot of work on gathering and analyzing internal source documents. The right sections of the source documents have to be highlighted and recognized in order to be transfered to the right section in the requested dossier. Besides analysis, interpretation and (often) rewording are part of this challenge too.

Solution Augment the work of regulatory specialists by applying a text analytics solution, so the creation or update of eCTDs can be automated step-by-step:

1. Conceptual search – Identify all correct documents that are related to a specific section in eCTD
2. Natural Language Processing – Match the right sections in the correct documents
3. Summarization – Create summaries of the right sections for the eCTD dossier
4. Language transformation – Adjustment of wording from/to regulatory text
5. Step-by-step automation of all sections and combined into one system. We apply rapid prototyping to learn fast and configure our framework for eCTD.

Impact Both productivity and regulatory compliance are positively impacted by using AI text analytics to automate a part of the eCTD process. Repetitive work is outsourced to the algorithm and substantive accuracy of regulatory content is improved upon.

Retail

---

Tenant mix

Case Consumer spending habits are changing – more spendings shift from physical stores to e-commerce, while consumers are increasingly looking for experiences rather than products. Malls need to reinvent their business in order to survive and serve their visitors in the best way possible. Making the right decisions in creating the optimal mix of tenants is key. An additional challenge is to assign tenants to the optimal location within a shopping centre to strenghten each others position and boost cross-visits of stores. How and where to start when addressing this complicated challenge?

Solution Advanced analytics has the potential to revolutionize the way tenant decisions are made within shopping centres. Often, shopping centres already have access to significant amounts of data regarding customer needs and shopping behaviour. The right analytical skills and tools can turn this information into valuable insights. Additionally, synergies among stores and tenant locations within the centre provide information about cross-conversion between stores and categories. Both customer- and tenant insights combined help mall operators make informed business decisions and work towards offering customers the optimal tenant mix.

Impact Using advanced analytics in selecting the right tenants offers mall operators the tools and knowledge to master the optimal tenant mix through fact-based decision-making. By adding optimal shop location, cross-visits and cross-conversions to the equation, shopping centres possess an unique competitive advantage on both customer and business side, allowing revenues to increase by 20%.

---

Customer service improvement

Case In most customer service departments, interactions with customers are recorded to use them for quality management, training and customer experience analysis purposes. Manually executing customer experience analysis by listening to the recorded conversations is quite labour intensive. As valuable as getting a glimpse of real life customer interaction may be, it does not allow for large volume customer experience analysis.

Solution The good news is that insights from customer conversations with your contact center become readily available with speech-to-text and text analytics AI services. The Speech-to-text transforms voice into written language. Text analytics analyzes and categorizes at a granular level the written down conversations. These categories provide valuable insights into the day-to-day issues your customers are experiencing.

Impact By automating the analysis of customer conversations with their contact center, companies gain a wealth of insight into the customer experience they provide. These insights can be highly automated with modern day cloud AI solutions, assuring near real time insights at low cost. Now all you need to do is quickly take and execute the right actions based on these insights.

‘Automate your customer experience insights with modern day cloud AI solutions‘

Banking

Voicebots (Speech-to-Text)

Case In the banking sector, AI technologies can help boost revenues through increased personalisation of services to customers. Rapidly providing accurate answers to a variety of questions is one of the key aspects for a strong customer relationship. A relatively new application of this helping tool is the voicebot, which translates speech/voice requests to textual input and eventually meaningful answers.

Solution Implement a chatbot with speech-to-text capabilities to process and answer questions and requests 24 hours a day, enabling you to offer your customers constant and continuous service and support. Voice recognition technology enlarges the impact of customer service tremendously, given the fact that 8 times more customers interact by voice than by chat. Not all cases are suitable for being handled by speech-to-text technology – personal topics and complicated items are preferably discussed with a human agent. Scoping the total amount of cases for those fitting to the voicebot is therefore an important starting point. The next step is to train the chatbot with the right range of knowledge and conversations to create smooth interactions with humans. Eventually, the speech-to-text application has the greatest impact when being fully integrated with the bank’s IT landschape for instant execution of requests.

Impact Dependent on the type of organization, the voicebots can significantly reduce the amount of effort needed to:
– Process questions and requests;
– Deliver answers faster, especially when demand is high;
– Free agents to handle more complex questions.
When done right, voicebots will lower costs, improve customer satisfaction and enriche the work of agents.

---

Advice quality

Case Many organizations work hard ensuring the quality of advice their advisors give. Especially in the financial industry, the consequences of non-compliant advice can be significant. Traditionally, quality managers would examine a random selection of all given advices to determine compliancy of those advices. In many cases AI models are able to predict the probablity of an advice being compliant. This opens up opportunities to make advice quality management more effective or more efficient.

Solution Advice is typically documented in the form of text. The text either contains a compliant advice or a non-compliant advice. Quality management classifies a subset of these advices using those labels. This data can be used to create and train a text analytics AI model. With this model we can now predict whether an advice is probably compliant or probably non-compliant.

Impact There are several ways this AI model can be applied to improve the business process. A) Instead of randomly selecting advices for quality management, the AI model makes a probalistic selection containing higher amounts of non-compliant advices. B) Instead of checking a subset, all of the advices can easily be checked. C) If the AI model reaches a high accuracy, non-complaint advices can be filtered out for re-work by senior staff. D) If the AI model reaches even higher accuracy, realtime direct feedback can be given to advisors.

---

Next best client

Case In many organizations, relationship managers need to decide on which people to contact to maintain the relation and/or upsell services. As by definition, a relationship managers’ time is limited, so most organizations have more clients to contact than they can possible reach out to. The better the selection of clients, the more effective and efficient the sales organization can be.

Solution Collect ever more granular internal and external data on client preferences, transactions and e.g events. Ensure these data are both well protected and easy to access. Develop analytical models to predict which client best to contact next, and with what content. With the help of AI, data relationship managers can now make an informed decision on the Next best client.

Impact The efficiency of the sales force as well as client satisfaction are increased. Clients are much more often contacted when they are about to make a decision on continuining services or when adding new ones. With the advice of the AI model on the next best product, not only the timing of the contact but also its content becomes more spot on.

‘Develop analytical models to predict which client best to contact next‘