zapp2photo - stock.adobe.com
A central tenet of Salesforce's philosophy is providing customers the means to create codeless applications. Its chatbot technology takes another step in this direction.
The goal in building most chatbots is the automation of tasks -- creating an interface that will respond to select user commands and then proceeding to execute an action or kick off a workflow. The Salesforce platform simplifies this complex interface and process by embedding most of the moving parts, which are prebuilt, into the platform itself.
At its simplest, a chatbot is the front end to an API, preconfigured for a chatbot-style UI, commonly called an agent. How does the agent work with Salesforce to execute tasks and workflows? It is connected to the Salesforce Lightning rules engine.
Here are the steps to create a chatbot:
- Select a platform for generating an agent. There are many to choose from -- DialogFlow, which is hosted by Google, is a good example.
- Generate the bot, per the platform's instructions. You will need to configure it, setting up user expressions it will respond to (intent) and identifying objects in the expressions (entities).
- Train the bot. Practice using the expressions to see if the objects are working as planned, and add intents and entities as needed. You will also have to allow for errors.
Most platforms can help you do all of this codelessly.
After you create a chatbot, it must be connected to Salesforce Lightning. That's where the API comes in. Sometimes, the chatbot platform will direct integration tools -- software development kits with the appropriate classes built in -- but sometimes, it won't. If not, there's middleware out there that can invoke the API. From there, set up the app in the Salesforce Lightning app manager, add it to the utility bar, select chatbot component and you're done.
However, it's easier to create a chatbot on the Salesforce platform, without having to connect to it externally. For that, Einstein Bot is available.
In Salesforce Bot Builder, the intent model is built through a manager, even adding chatbot response pauses so the interactions resemble humans'. The bot is then activated, making it ready for training, as described above. Within the Einstein Bot process, the bot can train against an extensive ready-made body of historical agent training data.
From this point, the bot has access to the full power of Einstein. It can be tied to a broad range of Salesforce objects and processes, accessing customer profiles and historical data, triggering workflows and storing data collected in the user exchange.
For highly complex Einstein Bots, Salesforce offers Einstein Bots for Developers, where complex custom coding can be done.
Finally, there's a dashboard available to monitor the bot's performance, once it's deployed.
Dig Deeper on Salesforce CRM
Related Q&A from Scott Robinson
Actionable insights can be gleaned from brand surveys and reviews, competitor reviews and social media. However, not all insights are insightful or ... Continue Reading
CX teams don't traditionally include the data scientist role, but adding one can help derive more value from customer data than a CX analyst or ... Continue Reading
Content management systems in the healthcare and legal industries use AI to automate tasks, saving businesses the time and money spent on employees ... Continue Reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.