MyBOT allows you to build chatbots of varying complexities & scale with ease.
A best-in-class Conversation Flow builder, robust integration capabilities and ability to deploy the bot on Website, Facebook Messenger or one of 14+ channels, MyBOT has made Chatbot building easier than never before.
The platform and it’s capabilities revolves around a chatbot and includes sections to help build, deploy, train, look at analytics as well as other associated capabilities such as Live Chat, Broadcast, Campaigns, Portal Users, chatbot users information, etc.
In the following sections we will try to understand what is the functionality of each of these sections and how they can help you in building your perfect chatbot.
As the name suggests, this section is what you would use the most while building your chatbot. It is further subdivided into FAQs, Paths, Entities and Data sources.
This section is what primarily drives the automated responses and is powered by MyBOT’s proprietary NLP engine. Adding FAQs is fairly straightforward, you just click on Add FAQ, add the query, add any variations of the primary query and then add the appropriate response. The response can be a simple text message as well as can be used to trigger paths. The FAQs can get more complex by adding and tagging entities in the query. The intent determination from the user query is made by the NLP engine based on the training provided.
There is also an option to upload all the FAQs at once via the CSV upload utility. This is the quickest way to train the chatbot if you have data already available in a FAQ format. FAQs/Intents can be categorized as well to manage them better as well as be tagged to a language while building multilingual bots. For Business plan users, you can also add and manage custom synonyms and stopwords from this workflow.
Paths are complex conversational flows which would involve more interactive elements than what the FAQ module can provide. The entire conversation design/structure of the chatbot can be modeled using the conversation modeler. Each specific aspect of the chatbot or a workflow can be built as its own path and then linked appropriately to put in place the entire chatbot. This modular design can be very helpful for triggering certain chatbot functionality from various entry points which may include FAQs as well.
Each path is composed of interlinked nodes. Each node is meant to perform a particular function. Some of the nodes are – Sending a simple message, carousel nodes for displaying rich menus or item listings, sending a message with quick responses/options, various multimedia nodes like send image, video and audio, faq filters, setting attributes, identity and many more.
There are a number of nodes focused around integration as well. JSON API node to quickly link to any standard REST API which supports various authentication modes as well as quick, easy ways to use the response values. The chatbot can also be linked to Zapier which can be used to connect to various other third-party services. Other than the above, there are a lot of GUI driven native integrations provided by the platform which include – Salesforce, Google Sheets, Google Calendar, SMS providers and more which will be added to the list.
Entities are critical data points or values which are extracted from a user query/input. The system provides inherent capabilities to determine entities and isolate them from the actual user intent. There are a number of system defined entities which are provided like – Date, Time, Date and Time, Amount, Cities, Countries and many others. Also, along with these system defined entities you can also define your own custom entities which could be list of products or services or anything else which can be an important data point for taking a particular action or understanding of the user intent.
2.4 Data Sources
Data sources are provided as a capability to facilitate loading of information to help the chatbot respond to certain queries. A data source is typically a two-dimensional data table which could be a list of entities along with the values of their corresponding attributes. File formats supported are CSV and Excel. Once the file is uploaded, intents and entities are automatically generated from the file and are consumed by the chatbot for training. The chatbot will now automatically start responding to queries around the attributes of any of the entities which are present in the tabular data.
The Intents and Entities can also be independently modified to probably add more variations to them or changing the response language to make it more meaningful.
3. Processing Information
We understand that during the course of your conversation with a user, you will need to process a lot of information. MyBOT provides you the use of Attributes to make it easy to collect, process, persist and even report on information collected. Some of the key capabilities include:
3.1 Collecting Information in Attributes
There are various means of collecting information from your bot user and saving it as an attribute. These include the use of nodes like Request User Data, Identity and JSON API in your conversation flow. You can also initialize the web chatbot with attributes to set the initial context. Once you set an attribute value using one of these Nodes, it is automatically persisted for use in the conversation flow as well as for reporting & viewing later.
3.2 Processing Attributes
Nodes like Script Node and Set Attribute allow you to set & process values stored in an attribute. Script Node in particular is a powerful means to do various kinds of processing on the attribute values including string manipulation to mathematical operations.
Besides the above, all Display Nodes like Send Message, Send Message with Options, Send Carousel etc allow you to use the attributes as a placeholder within the display text.
3.3 Reporting on Attributes
MyBOT provides two key mechanisms to let you view & analyze the information being collected. This includes:
– User Report: You can configure the available columns on this report to include specific attributes
– User Details: For an individual user, MyBOT automatically displays the entire list of available attributes and information collected against those for that user.
4.1 Test the Chatbot
Once you have started building the chatbot, you would want to test out the conversation flow structure as well as the NLP responses. This can be easily done from the home page of your account, clicking on the three dot menu icon for a bot and selecting the option “Test Bot”.
Once the chatbot is built and trained to satisfaction, it needs to be deployed on any of the multiple messaging channels which are supported. The same chatbot can be associated with multiple channels without any additional efforts. The platform does the part of normalizing the conversational experience for the channels configured and providing a consistent experience tied to what was built as part of the FAQ training and the conversational paths.
There are 14+ channels on which the chatbot can be set up which include Whatsapp, Messenger, website widget as well as standalone web link, mobile SDKs for Android and iOS, Telegram, LINE, Slack, Skype and many more..
The steps for configuring the chatbot for each of these channels are different and are amply documented in the user assistance guides for the channel setups.
Once the chatbot is built and deployed on the channels, there are various statistical data points which are collected for the chatbot usage and are made available via the multiple Dashboards as well as the Users section which particularly focuses on the chatbot users. The Dashboards cater to a specific aspect of the chatbot and highlights such metrics –
The Engage dashboard primarily provides stats around the chatbot engagement. Some of the key elements here are the number of user interactions, new user growth and usage trends, and some metrics around the conversation type which include average messages per conversation and conversation duration. It also common button actions and frequent messages data.
The Responses dashboard primarily focuses on the NLP aspects for the automated responses to user queries. You get to see trends of total questions asked to the bot and the split between answered and unanswered questions. You can also drill down to see the individual questions also with details of the chatbot response, NLP engine’s score determination for the match and you can also train the chatbot from there.
The Retain dashboard helps you with data and insights around user retention. You get a clear depiction of the repeat usage patterns of the chatbot users with time and can understand when they drop off post-joining.
The Support dashboard primarily contains statistics around the live chat aspects and agent performance. Some key metrics are average time to respond, the average time for a user to wait for an agent, inbound and resolved query trends, etc. Also, these data points can be grouped by days, agents or agent category to provide more meaningful insights.
6. Other capabilities
There are a number of other capabilities which can enrich your chatbot outside what is provided in the build section. Below is a brief description of these –
6.1 Live Chat
Live Chat capability for the chatbot also comes as a pre-packaged solution without dependency on a third-party service. This capability augments the automated responses capability of the chatbot to handle cases which the chatbot is not currently trained for or for cases where a human intervention is necessary to perform a certain task. Chats can be seamlessly handed over to an agent, live chat agents can be assigned to categories/groups, chats can also be transferred between agents and chat context is maintained and can be exported/downloaded.
Broadcast capability helps you to proactively reach out to your users for certain announcements as well as send any promotional activities you might have for your chatbot users. Broadcasts can be sent to all users or users can be segmented as well to allow for more targeted and relevant marketing messages. Various stats can also be viewed around the broadcasts which are sent out. You can also set up periodic broadcasts or broadcasts which trigger a certain path/flow. It is to be noted that broadcasts are channel dependent and there could be restrictions on promotional content as well as the timing of the broadcasts which are channel-specific.
Automated campaigns can be designed to re-engage the users or to nudge them to complete a particular action if they have dropped out of it. Each campaign can have a sequence of steps which can send out a message to a user or trigger a path. Each of these steps can have a particular delay between that step and the one prior to it.
Users’ subscription to campaigns can be managed via the nodes for subscribe as well as unsubscribe.
6.4 Portal Users
For a complex chatbot, more than one users might be required to manage the bot. The platform provides the ability to invite multiple users to the chatbot in various capacities. You can add chatbot administrators, live chat support agents, or bot builders. The access to the various workflows would be controlled depending on what role the particular user is provided. It is to be noted that a user who already has another account on the platform cannot be invited to a different account’s chatbot.
The Account section helps you manage your MyBOT account, it shows the total number of interactions you have consumed, the number of bots you have as well as plan and billing information, including current plan details, any associated trials, payment history, etc. You can upgrade your plan based on your requirements from this section.