Turbocharge Support Responses with an AI Copilot for Freshdesk

AI in customer support is used primarily for operational process optimization.

Two dimensions of AI capabilities for improvements can be considered, depending on the agents’ role.

The first dimension includes tools created for agents’ performance enhancement. In this domain, optimization becomes possible through routine task automation or semi-automation.

In the case of up to 100% automation, responses in chats, e-mails, social media, and other channels of communication are generated automatically. The human role here is outcomes analysis and improvement implementation.

In the case of semi-automation, AI provides a human with information or suggestions, and a human agent responds to the request by hand. Outcome supervision is a human’s task as well.

Below is the list of features examples related to agents’ assistance:

• automated responses;

• response suggestions;

• sentiment analysis.

The key capability that AI provides in the realm of response automation is personalization. Each response, the ready-to-send or a suggested one, is tailored to a particular customer and a particular case. It’s achievable due to AI model training based on conversations and the connection between AI models and databases containing real-time information.

The second dimension includes mechanisms and technologies for optimization in which human agents aren’t involved directly. These operations are performed under the hood.

Examples of such operations include the following:

• personalized product and service recommendations;

• requests routing;

• personalized content delivery for self-service;

• dynamic customer profile.

On the operational level, the tasks of AI injection in customer support are the reduction of time, effort and costs required for the delivery of personalized and accurate information to customers.

On the strategic level, the ultimate goal of AI implementation in customer support is improved customer experience that, in turn, increases customer loyalty.

CoSupport Agent and CoSupport Customer are AI assistants. Compared to chatbots, they are capable of handling more complex and creative tasks. 

The difference between three popular AI-based tools for customer support is as follows:

• Chatbots are relatively simple tools based on solid algorithms. They follow certain If-Else rules and provide pre-programmed answers. The level of personalization is zero.

• AI chatbots are AI-powered tools for conversations with language models. GPT is a popular example of such a solution. Today’s AI chatbots like CoSupport Customer have additional features that help them connect to other tools or perform certain additional tasks, such as browsing or text summarization. 

• AI assistants are AI-based solutions that enrich other tools’ functionality with new capabilities. For example, CoSupport Agent integration with Freshdesk extends the CRM’s ability to manage communication with customers and sales using AI skills. The primary task of the integration is to create personalized response suggestions. 

The key AI assistants’ difference is that they are embedded naturally into existing tools to automate processes and enhance users’ performance. 

In the case of CoSupport Agent and CoSupport Customer as AI assistants, they are the examples of Freshdesk integrations. The solutions integrate naturally to Freshdesk, and customer support agents can use their functionality within the same interface while leveraging additional features.

Both solutions integrate with Freshdesk directly, meaning that agents use the same Freshdesk interface enriched with AI-driven features.

CoSupport Agent integration into Freshdesk is quick and easy and includes the following steps:

• install the CoSupport Agent browser extension;

• log in to Freshdesk;

• navigate to Freshdesk to process requests;

• AI creates the response suggestions;

• edit the best suggestion if needed and send a response.

Suggestions are based on conversations with a particular customer; hence, they are personalized and reflect all “knowledge” of a particular company’s products and services that an AI assistant comprehended during the model training.

CoSupport Customer integrates with a chatbot on your web site. You provide CoSupport AI team with an API key for your chatbot, and a chatbot starts generating responses autonomously.

CoSupport AI uses exceptional measures to guarantee data security. Below is the list of the key approaches:

• Data that is used for CoSupport Agent  training is anonymised, meaning that all personal information is removed from databases. Thus, only impersonal information is used for training.

• In both solutions, operations with data are performed on CoSupport AI’s side, with data storage on dedicated servers.  Moreover, all data are used for one particular project needs; project A’s data is never used for project B.

CoSupport Agent works as a browser extension. Customer support agents download the extension while an administrator tunes the system for flawless integration. CoSupport AI solution is one of the easiest Freshdesk integrations to implement.

When agents work in Freshdesk, they use the following features (in the Freshdesk interface):

• Response suggestions with placeholders. Suggestions are based on conversations and information on the company’s products and services on which an AI assistant was trained.

• Semantic analysis of the text with highlighted intents and key details. It makes request understanding easier and reduces response time.

Both key features aim to enhance customer experience and increase customers' loyalty.

CoSupport Agent also functions as a browser extension. Once you provide the CoSupport AI team with an API key for your chatbot, CoSupport Customer begins generating responses autonomously, without the involvement of human agents.

To access the AI assistant’s functionality, you should follow these steps:

1. Data extraction for training models: Data is collected from the sources provided by a company. The primary source is Freshdesk, which includes all customer correspondence. Additional sources include the knowledge base, documents describing products and services, policies, instructions, and guidelines for customer support agents. CoSupport Agent works with all types of textual information. We can also use call transcripts.

2. Data anonymization: All data goes through a separate model that anonymizes it. Personal characteristics such as names, credit card info, and location (up to 16 types of data) will never be used for training our models.

3. Model training: All materials are fed to the models.

4. Solution launch: Access is provided to admins and end users (human agents).

The time required for Freshdesk AI assistant integration depends heavily on the amount of data for model training and handling organizational issues. The process can take up to a month.

Most of the data is sourced from Freshdesk. We utilize the following types of data derived from the CRM:

• Tickets (we need 1000+ tickets for training)

• Emails

• Product reviews

• Comments

• Call transcripts

Additional data sources include:

• Articles from the knowledge base

• Content from internal and external websites, including FAQ sections

• Policies

• Step-by-step support instructions

Any documentation used by your customer service agents to perform their tasks serves as a valuable data source for feeding the AI models. We handle all types of textual data in various formats, including PDF, Excel, Google Sheets, JSON, and more.

The pricing policy for CoSupport Agent integration with Freshdesk is based on a scheme “fixed fee for one agent for one month.” The number of responses within a standard plan is limited. When the number of responses exceeds this limit, an additional charge is applied. To resume, the monthly charge for CoSupport Agent depends on the customer support team size and agents’ workload (the number of requests and, consequently, responses). Get the quote.