At a Glance
About IVR, Bots and Conversational AI
Softdial Contact Center™ (SCC) allows users to create automated applications and connect to external automated services to handle a variety of tasks for voice and text sessions, from the simple and repetitive to the more complex.
Automation allows you to do more with the call center resources you have, making the best use of agents’ time by reserving them for higher value tasks and more personalised service.
With SCC, users can:
- create Interactive Voice Response (IVR) process flows for call center voice sessions
- create text bots to deliver basic handling for text sessions
- integrate external conversational AI and bot services.
All automated applications and integrations can be set up within SCC by non-technical personnel, in a fraction of the time of conventional programming methods, by using Softdial Scripter™, SCC’s fully featured workflow design application.
Comprehensive IVR software toolkit
Scripter offers a comprehensive toolkit for creating IVR call center process flows, or ‘scripts’, from the simple to the complex. With these tools, scripts can:
- take user input (DTMF or voice)
- interact with a database or CRM
- play recorded or Text to Speech output
- perform session control, e.g. transfer to a live agent
- record a customer voicemail.
By combining these building blocks, IVR scripting can be designed to:
- block or unblock a customer’s credit card in case of fraud and then send them an SMS to confirm the status change.
- handle an inbound or outbound call in its entirety (where local regulations permit)
- validate and route an incoming call to a queue or a blended live agent
- accept transfers from live agents to handle parts of an inbound or outbound call
- handle sessions of other media types such as email and text messaging using SCC’s Text to Speech/ Speech Recognition features.
Agent transfer to IVR system
Call center agents can transfer calls to the IVR phone system mid-conversation to take some of the strain of repetitive tasks, or handle a particular process, for example:
- to present mandatory terms of business and capture an ‘accept’ response
- to take payment details for PCI compliance where sensitive customer information, e.g. account passwords, must not be revealed to the agent handling the call.
On completion, the IVR application can then pass the call back to the original agent.
IVR scripting is designed using the drag-and-drop tool within Softdial Scripter™. To create a script, the user simply drags appropriate logic blocks, or ‘steps’, onto the design area and connects them together to form a flow.
Below is an IVR call center script example, using a few standard steps from over 50 provided.
This example script can be assembled and tested in 15 to 20 mins by a relatively inexperienced user.
When activated, this script will:
- Play an introductory greeting to the caller, requesting them to enter an account number
- Capture the account number and save it to a script variable field, or exit on fail
- Perform a database lookup based on the account number captured
- If lookup fails, play a prompt requesting the user to re-enter the account number
- If successful, re-queue the call to a live agent, attaching the account data to the session
- End the IVR call transaction for the IVR application.
Complete solution for bots
Softdial Contact Center™ (SCC) provides a complete solution for bots for any type of text media – chat, SMS, email, WhatsApp, etc.
There are two main types of text bot:
- rules-based – can only respond to very specific commands. If the customer doesn’t use recognised commands or words, the bot doesn’t understand and cannot respond
- AI-powered – uses machine learning and artificial intelligence to provide the best response.
Both kinds of bots are supported by SCC, and depending on the complexity involved, can be either:
- entirely composed within Scripter, or
- third party products, integrated using Scripter.
Creating a text bot script
With Scripter, users can design text bots to handle a wide variety of customer service activities, e.g. delivery confirmations, health centre enquiries, product maintenance requests and information, timetable requests.
Scripter enables access to any external system or information source, to collect data to help make decisions and improve the bot dialog.
The image shows an example of a simple Weather Bot script.
When activated, this script:
- asks the customer for their nearest town
- executes a custom step to retrieve current weather from openweathermap.org
- advises the customer of weather status, description and temperature.
To help understand customer intention, Scripter contains a Text Processing Library, including a Natural Language Processing (NLP) engine.
Read more about Natural Language Processing (NLP) for Bots.
Integrating external bot and conversational AI services
Scripter users are not limited to what the tool itself provides. To take the customer conversation as far as it can go without involving an agent, Scripter allows easy integration with external bot and conversational AI engines.
In this case, Scripter acts as an intermediary for all text communication between the customer and the external service, for example:
1. customer sends text
2. Scripter passes customer text to the external service for processing
3. the service may respond in a variety of ways, e.g.
- with a response to pass to the customer
- with an instruction to Scripter to pass the text communication to a live agent, for instance because the customer requested it, or because the service interprets that otherwise the customer may sign off.
When a session is transferred from a bot or live agent to another live agent, all session history is also made available, removing the need to re-establish who the customer is.
See also this example use case.