Lead Responses

What will the Assistant do when a response is received?
Emails responses are read by the Assistant and would try and decide the email intent.

The machine learning model/s tries to identify the response to one of the below.

• OutOfOffice - out of office responses
• Dectivated - asks to stops sending any further emails
• AgreedToMeet - agrees to have a meeting at a particular day
• BrochureSent - requests for brochure
• ConnectLater - request to check back later
• Unsubscribed - unsubscribe me
• InterestedToProceed - shows interest in the product and services.
• ReferredSomeone - refers someone else in the organization to take the conversation ahead
• MovedOut - sends a response saying that lead has left the organization
• Retired - sends a response saying that lead has retired
• SeekingMoreInfo - asks for some more information about the products and services
• SupportAutoResponse - when support tickets are created

This helps the assistant take right action and update the lead state.
Some examples are mentioned below:

Response	Inference	Action
I am out of office for 2 weeks. If urgent …	Lead is Out of office.	Connect when lead is back after 2 weeks
I am not the right person. Please don’t email me.	Wrong Lead	Stop further Communication
Please send me your brochure.	Warm Lead	Send the email with brochure immediately. Then stop further Communication
Send me your company profile.	Warm Lead	Send the email with brochure immediately. Then stop further Communication
Setup a meeting for tomorrow 5 PM	Hot Lead	Forward email thread to Sales. Notify Sales on Phone. Stop further Communication with the Lead. Remind sales after 24hrs to avoid this lead being dropped
We are busy in some other priorities right now. Lets talk sometime in Jan.	Warm Lead	Set future communication in Jan.
Unsubscribe me from your mailing list	Unsubscribed	Lead would be marked as unsubscribed.

What is Engaged, Warm and HOT leads?

Assistant will categorize the leads in various states based on the followups and responses. Various possible state are:
New conveys that the lead added cannot be processed as you have used up your leads limit as per the plan. You can Restart these leads in the next period or upgrade your plan and then restart these leads now.
Cold the followup is under progress.
Hot if the lead has accepted to meet.
Warm lead is someone who has opened 3 or more emails or has clicked on a link from the email. A good list for your calling team to call.
Engaged lead is someone who has opened 1 or more emails.
Processed followup is completed as per the sequence and there was no response from the lead.
Out of office received an out of office from the lead. Assistant will restart the followup after lead returns back to office. What if the next email for the lead is going beyond the out of office date for the lead ? Deactivated lead responded back asking to stop sending emails. Lead email id is invalid as verified by the assistant. Email sent to the lead bounced.
Unsubscribed lead used the unsubscribed link in the email or responded asking to unsubscribe. You wont be able to send any more emails to this lead.
Responded state for the lead conveys that there is some response from the lead and Assistant was not able to categorize it to one of the above.
Pending means the email id was NOT certain to be valid (Un-verifiable), so Assistant will process these leads at a slower rate to avoid getting into Bounce issues. But if Stop Followup action was taken on these leads then Assistant will not pick them for processing. You can ForceSend or restart if required.

Teach your Assistant

Some of the responses may over fit or under fit the out-of-the-box ML model and the response received from the lead may get identified in a State which may not be most correct.
Some responses could be very specific to your industry and customer profile and so this may happen when out of the box ML model is used.

Assisted Learning is the possible solution for above problem. You can provide specific terms from the response email and mention which sub_state and state it should be mapped to. This will be specific to your organisation, any response received will be passed through a rudimentary model created from this small amount of data.

• If there is a match with more than 90% confidence then that assisted learning mapping would be used.
• If more than one row matches with more than 90% confidence then the highest confidence mapping would be used.
• If no row with 90% or more confidence matches then anyways the out of box model in 7Targets would be used.

You would also see a note added mentioning which mapping was used from the Assisted Learning, if Assisted learning mapping is used.

Questions answered

• Why some leads are showing as Pending ?
• Why some leads are showing as New ?
• Why some leads are showing Deactivated ?
• What will the Assistant do when a response is received ?