Often in our conversations we want our agent to have different conversations with our contacts based on how they respond to a question or based on information that was previously collected. There are several ways to achieve this using CloseBot.
​
Below are several examples of how this can be achieved, as well as guidance for what is the best method for your unique use case.

Most of the time, adding instructions to an agent node to do "this" if they say "this" and do "that" if they say "that" will work very well. This option is usually the best way to route conversation flow based on responses from the contact.

In the example above, if the contact only needs paint, the agent will collect year, make and model of the car then book to the Paint Calendar, but if the contact needs body work, the agent will find out if they would like PDR or Traditional Bodywork before collecting year, make and model and book those contacts to the Bodywork Calendar.
​
I typically follow a standard format for these scenarios. I will tell the AI what information needs to be gathered. Then on the line directly below it, I will add "if" statements for how I expect the conversation to go depending on the answer the contact gives. That formatting is for my own benefit and does not effect the ability of the AI to do its job.

Sometimes the service the contact wants leads to very different conversations. For instance, your agent is for a real estate business that helps with Buyers and Sellers.

As you can see in the example above, the agent finds out what the needs of the contact are, then it uses the appropriate exit to send the contact to the correct agent node for their needs.
​

Although this Agent can be built within a single agent node. Many users find it simpler to route their contacts to two separate agent nodes to reduce cost, simplify troubleshooting, and reduce the amount of tools an agent has access to at any certain point of the conversation leading to smarter tool usage.

Although nearly any case can be built using the agent node, there are still some good uses for our legacy actions like True/False and Switch.

If we already know what product or service the contact wants before the conversation with the agent starts, we can route the conversation to different paths directly from the start by adding a switch that checks the value of a contact property and routes accordingly.

Above we show a Switch that is checking the value of a field without using AI before it routes the conversation to a specific path. This works best when the property the switch references is a single dropdown. We build a case for each option that is available in that dropdown.
​
Notice how even though we only have two cases, there are 3 paths that exit the switch. The lowest path from a switch will always mean none of the options apply. We can use that path to show how we can route paths from our Legacy objective action using a Switch that uses AI.

In this Example, we assume that we will know what service the contact wants before they start talking to the agent. If that field is not set we will want to have a way to route the conversation appropriately through conversation.

Above, we create an objective that will find out what service the contact is interested in. When that objective is completed, the contact will exit the objective and enter the switch that will use the previous conversation to route the contact to the appropriate next step.
​

After connecting all of the exits from the switches, we have a job flow that will automatically route people to the correct agent node if the contact's "Service" field is a known value. If it is not known, the agent will find out if the contact is looking for a new install or service to an existing system and route the contact to the correct node so it can have the specific conversation based on their needs.