Public:Sales AI: Difference between revisions

From Docs | Smarter.Codes
(Created page with "Sales AI is a product we are building for internal use. Category:Sales AI")
 
No edit summary
Line 1: Line 1:
Sales AI is a product we are building for [[internal use]].
Sales AI is a product we are building for [[internal use]]. It would be a virtual assistant that augments our Sales colleagues to be better in sales.
 
= Roadmap =
 
== R1: search across all sales communication on upwork and google docs. Month1 ==
This should reduce 5hr per week of a full time sales rep
 
== R2: identify "near duplicate" snippets with 75% accuracy. Month2 ==
This should reduce another 5hr per week of a full time sales rep
 
== R3: Auto label snippets with 75% accuracy. Month3 ==
 
== R4: Prompt snippets to sales reps with 75% accuracy. Month4 ==
 
== R5: Chatbot that guides visitors on website with 75% accuracy. Month5 ==
 
 
= Capabilities =
 
== Real time push data into Customer Data Platform ==
Put all the customer communication that happened during sales process in a [[:en:Customer_data_platform|Customer Data Platform]]. See how we [https://docs.google.com/document/d/1Za92EiH-bP7HF_-Z7ruqIZeSqfTgZAhxUEzTsH4d6Ag/ proposed putting a CDP to one of our clients]
 
While Segment.com let you do it pretty nice, you can consider [https://www.google.com/search?channel=fs&client=ubuntu-sn&q=customer+data+platform+using+apache+stack open source solutions for CDP] (connect with Ricky - he was trying to do same for BestOfBharat.com)
A sales rep would be able to use keywords to Search across all..
 
* cover letters submitted on upwork
* replies received from upwork
* on whatsapp
* on email
* on telegram
 
== Identify the repeat strings from historical sales communication ==
Sales reps would be able to see "near duplicate" strings like as shown in this [[Snippets List]]
 
There are some [https://github.com/google-research/deduplicate-text-datasets open source that identify near duplicates in Natural Language].
 
== Auto Label those snippets ==
Click on any snippet from [[Snippets List]]. The labeling in there would be done automatically by machine. Labels have to be first Identified, then applied. To identify labels we can..
 
* Look for keywords highlighted in google docs in those cover letters
* do a keyword density extraction on our sales communication
* Use WikiDigger to find out wikipedia terms for are there for [[Guidelines on Making First Contact Onboarding of Sales People#Identify the Keywords|Tools, Problems, Goals, etc]]
 
== Be present in everything Sales Reps do ==
Monitor their real time calls
 
Stay on their desktop as they are drafting replies to prospects
 
== Prompt those snippets to Sales Reps when similar communication is detected ==
 
== Chatbot that guides prospects to become hot leads ==
See [[Chatbot for HC Sales]]
 
[[Category:Sales AI]]
[[Category:Sales AI]]

Revision as of 07:26, 14 January 2024

Sales AI is a product we are building for internal use. It would be a virtual assistant that augments our Sales colleagues to be better in sales.

Roadmap

R1: search across all sales communication on upwork and google docs. Month1

This should reduce 5hr per week of a full time sales rep

R2: identify "near duplicate" snippets with 75% accuracy. Month2

This should reduce another 5hr per week of a full time sales rep

R3: Auto label snippets with 75% accuracy. Month3

R4: Prompt snippets to sales reps with 75% accuracy. Month4

R5: Chatbot that guides visitors on website with 75% accuracy. Month5

Capabilities

Real time push data into Customer Data Platform

Put all the customer communication that happened during sales process in a Customer Data Platform. See how we proposed putting a CDP to one of our clients

While Segment.com let you do it pretty nice, you can consider open source solutions for CDP (connect with Ricky - he was trying to do same for BestOfBharat.com) A sales rep would be able to use keywords to Search across all..

  • cover letters submitted on upwork
  • replies received from upwork
  • on whatsapp
  • on email
  • on telegram

Identify the repeat strings from historical sales communication

Sales reps would be able to see "near duplicate" strings like as shown in this Snippets List

There are some open source that identify near duplicates in Natural Language.

Auto Label those snippets

Click on any snippet from Snippets List. The labeling in there would be done automatically by machine. Labels have to be first Identified, then applied. To identify labels we can..

  • Look for keywords highlighted in google docs in those cover letters
  • do a keyword density extraction on our sales communication
  • Use WikiDigger to find out wikipedia terms for are there for Tools, Problems, Goals, etc

Be present in everything Sales Reps do

Monitor their real time calls

Stay on their desktop as they are drafting replies to prospects

Prompt those snippets to Sales Reps when similar communication is detected

Chatbot that guides prospects to become hot leads

See Chatbot for HC Sales