Patterns for chatbot integration with the contact center
10. 1. 2024
A chatbot using artificial intelligence, especially generative AI, is becoming an essential part of customer care. It is also used for back office support of employees. In this blog, we want to describe the possible approaches to integrate chatbot with contact center.
Today, most large companies with a wide customer base already use a chat channel for customer support, taking orders and handling processes. In this case, the chat widget is mostly integrated into the company's website. If the customer starts using the chat, this communication is connected to the contact center, where a avalaible agent with the appropriate skills is found by routing rules and is subsequently connected to the customer and communicates via the chat channel. Usually, the transcription of the chat communication is created as an interaction in CRM, or linked to the chat archive. Tentpo access requires the availability of agents.
We are an IT consulting firm that develops and adapts IT solutions for customer communication and customer data management. The challenge of this time is also the migration of IT infrastructure to the cloud, the development of cloud-native solutions and the use of cloud services that we offer to our customers. We are partners of the world's largest cloud providers and we are looking for the most suitable solution and cloud architecture for the customer.
We are consulting partners of the largest cloud providers, and therefore when choosing a chatbot for a customer, we always consider in particular: :
- Oracle Digital Assitent in Oracle Cloud Infrastructure
- AWS Lex chatbot with AWS Bedrock genAI models
- Microsoft Azure AI Bot Framework Service
Chatbot can be integrated with the contact center with these possible patterns :
- Chatbot as proxy
- Chatbot as agent
Chatbot as a proxy assumes orchestration for the chatbot with the new Message Router component and integration with the contact center. The customer communicates through a defined channel such as webchat, teams, or whatsapp with the bot, unless there is a situation where the transfer of the conversation to a contact center agent is defined. The router component has logic in it, where it decides whether the dialogue with the customer goes to a bot or a contact center that directs it to the right agent. The bot stays in the loop and can collect logs and information from the conversation, filter messages, or add additional content to the dialog.
In the event that the cahbot does not know how to serve the customer, or the customer directly requests the connection of a live agent, the so-called "hand off" process. When detecting the need for a live agent, all collected data is collected so that the person has the context of previous communication, customer identification. The router ends the communication with the chatbot and calls the contact center to assign an agent to this chat communication. The picture shows an example with the Genesys contact center.
The second variant of the solution, the so-called "bot as an agent" is the integration of the bot into the orchestration of the contact center, where it is considered an agent with special skills. The contact center has the functionality to integrate the bot into the flow. The conversation between the customer and the bot can be transferred to a contact center agent called "handoff" protocol, where the bot moves the dialogue to the contact center, which directs it to the right agent. The main advantage of this solution is simplicity and transparency. There is no need for duplication of channels and all existing communication channels will be used very effectively.
The third method represents the connection of the chatbot directly with the customer/user without the possibility of routing to the contact center. The solutions we offer offer a number of channels via webchat, ms teams, whatsap, facebook messenger, slack, sms, email. The connection between the superstructure and the voicebot is also very interesting, when the framework is able to communicate by voice and recognize speech.
It is important to point out that the offered chatbot frameworks are multilingual, they can recognize the language in which the customer communicates and respond to him in this language. It is in today's global world to serve customers in their language.
Our solution is also offered by corporate chatGPT, which can learn the context of the company and answer customers' questions related to the content of the company's activities. It is also important to mention the recognition of personal data, their removal from the transcription, as well as the recognition of the customer's emotions.
Contact us so that we can show you all the options and deploy the chatbot for you as soon as possible.
Back to Blog