Next, we will pass the text output to our LUIS action where we will provide this text as our Utterance Text. If we expect our emails to be received in Html format, we can use the Html to text action to remove the Html markup so that when we insert the content into Dynamics 365, we are not including this additional formatting. We can include a Subject Filter if we are controlling how the email is created, such as an email form on a web page. Our flow will be triggered when a new email arrives in our customer support mailbox. To call LUIS from flow, we need to copy the Key String. We can use this endpoint to perform additional testing through a tool such as Postman or from a custom application. With our testing complete, we can publish our application to a Staging or Production Slot.Īfter our application has been published, we will have access to a starter key and an endpoint URL. Once we have created our Intents and Utterances, we can then test and train our application. The more descriptions we provide, the more accurate our LUIS model will be. In our Power Outage Intent, we will describe some sentences that we expect to receive via email that imply there is a power outage. Utterances are related statements that infer an action, in natural language, provided by a user. With our Intents defined, we need to go into each one and provide some utterances. Within LUIS, we are going to define four different Intents: LUIS has been covered in a previous Flow Blog post, so please refer to that post if you are new to creating LUIS applications. Once we have created our case within Dynamics 365 and made the impacted department aware, we will acknowledge the customer’s request by sending a follow-up email that includes their case number.īefore we work on our flow, we need to build our LUIS application. This way, they can prioritize the resolution of the incident. If the customer concern is a high priority, such as a power outage, we want to let our Field Services know right away. We will use LUIS to interpret the intent of the email that our customer has sent us so that we can log it appropriately within our Dynamics 365 case. LUIS stands for Language Understanding Intelligent Service and uses machine learning to allow people to build applications that can receive user input in natural language and extract meaning from it. Next, the body of the email is passed into our LUIS connector. Instead of an employee reading, and reacting, to these emails, Microsoft Flow will apply a filter on specific keywords. To achieve their goals, Contoso Energy is going to use Microsoft Flow to automate the retrievals of emails that are received in the customer service inbox. This allows us to prioritize their case correctly and ensure the right department can address high priority incidents asap! Some of these requests are higher priority than others so it is important we understand the user’s intent when we log the support case. Especially, when you consider the people that can help solve the customer's issue, are not immediately informed of the issue.Īs a power company, there are many customer requests that can be made including reporting a power outage, billing queries, street light replacement and vegetation brushing to name a few. Sifting through emails manually to log support cases inside of their ERP (Microsoft Dynamics 365) is a time consuming and low-value activity. In addition to mobile, social and telephone conversations, they continue to receive customer support requests over email. Fortunately, we can take advantage of LUIS.ai which will understand the customer’s intent so we can route this information correctly.Ī power company, called Contoso Energy, would like to improve their customer service and wants to provide many channels for customers to solicit assistance. But, to achieve this, we need some intelligent technology.
If we can interpret their intent, we can ensure the correct people within our organization are aware of the issue, to provide a more timely outcome. To achieve our new customer service goals, we can leverage Microsoft services like Microsoft Flow, LUIS.ai and Dynamics 365.Ī strategy we want employ is to better understand the types of issues our customers are having as soon as possible. The question becomes how can I scale my customer service department, without adding labor costs? The answer lies in the phrase “work smarter, not harder”. It typically isn’t practical to meet this demand through additional headcount. As organizations focus on customer obsession, they need to scale their organization to exceed elevated customer expectations.
0 kommentar(er)
