Natural Language
Processing(NLP) Salesforce Integration

This article will provide an overview of how Natural Language Processing (NLP) Artificial Intelligence (AI) algorithms can be used and Integrated into your Salesforce Org.

NLP Overview

NLP used machine learning algorithms based on provided text to determine the context and sentiment.

Context Analysis will review the major words used in a text string and determine the subject, verb, adj etc. There are many APIs that provide this service, such as IBM Watson, Microsoft Cognitive Service, Wit, Metamind, and Google Cloud. It just depends, how much context you need. Some services like Google, can be fed and unlimited stream of text, like a Contract Document, and provide a summary, or ambiguities. Others are better at simple phrases.

Sentiment Analysis will review the overall tone of of the text provided. It will return a score of -1 to 1, -1 being negative, and 1 being positive.

Application of NLP

Lead/Contact/User Personality Profiles
With the mix of Context and Sentiment Analysis we can create personality profiles for any Lead, Contact, or User. We can use any type of text input: Email, Chatter, Voice (Voice to Text Twillio integration), Social Media (Twitter, Facebook, LinkedIn, etc.).

Here is an example of using the Pope’s Twitter handle (@pontifex_es) to create a personality profile:



screen-shot-2016-10-21-at-11-08-19-am  null

Uses for this type of analysis are endless:
Determine best day and times to follow up with leads/contacts.
Determine best methods to communicating with leads/contacts/users.
Employee background checks
Employee performance reviews.

Tone Check
Tone check would add another layer of professionalism to all outgoing communications in your company. Similar to Spell Check, that looks for misspelled words, Tone Check would tell you the overall Tone of the email, text, phone calls or chatter. We can also include validation rules to warn or prevent certain phrases that have a negative tone from being sent to a client. This way you can monitor all communications with clients and ensure sales and customer service upholding your company’s branded image.

We can also include predictive algorithms to suggest alternative words for the selected phrases that does match the tone you are trying to achieve.



Escalating Service Request
Service requests, phone calls, emails chains, and web chats, can all be analyzed with sentiment analysis in real time, thereby allowing issues to be escalated to senior support reps to handle unhappy customers. This can also be used to remove or warn a support rep who is using a negative tone with a customer.

Get in contact with us today so we can help strategize the best ways to implement natural language processing.