Query your inbox with natural semantics
“Insulation quotes we’ve given”
“Cylinders used for food or drinks”
“What is the difference between duplex and super duplex?”
Horton Miller is an industrial fluids company that specializes in metal protection. They have just hired a new sales employee that received a question from his prospect regarding non-ferrous protection. Fishing through past emails while on the phone would take too long. His only option is to ask the client if he could get back to them, hang up, and search his email or ask a more tenured employee.
Simple, yet powerful features
Natural Language Understanding
NLU is an ever growing technology and is quickly becoming a must-have. This is true machine learning and it has one goal: to understand the true meaning of your query. Far beyond simple key-word search.
Multiple results
Within the machine learning model, NLU Search provides multiple results for your natural query. Number of results can be customized per customer request.
Highlighted findings
Vector identification in certain indexes allow the model to pinpoint an answer, then it leaves it to artificial intelligence to highlight. Highlighting potential answers saves you time, making you faster.
Clickable results
Many times our customers want to open the entire email chain for more context. NLU Search offers clickable links to open such chains.
User results rating
Notice the thumbs up/thumbs down. This allows you to provide feedback to the model so it can provide more accurate results in the future.
Attachments Included
Most emails that contain valuable information have attachments. NLU Search pulls the attachments that are correlated with the email chain in order to give you all the information at your finger tips.
How it works
This solution is built entirely in AWS. Don't have an AWS account yet? We will get you all set up and secure.
You own your data
Our solution is deployed in AWS. If you don't have an AWS account yet, we will create one in accordance to the "Well Architected Framework" to ensure security. Your emails stay with you.
Secure
Data transfer from your email server to an Amazon S3 bucket is encrypted. All transformations take place in a private subnet where theres no outside access.
Data Transformation
Emails are put in a raw S3 bucket, then transformed to JSON for the machine learning model. The model learns the data and stores in an index.
Seemless Retrieval
Queries are submitted through interface into a machine learning endpoint for understanding and retrieval.
