Public:SC Office

Vision

File:Customer and Employee Experience together.png

Customer and Employee Experience - both target markets targeted by Kore.ai

Trying to build a Holistic Solution which is a mix of :

* Knowledge Management & Enterprise Search like getguru.com, rewind.ai

* Team Messaging like Microsoft Teams / Slack

* Collaboration Wiki like Notion.so

* Time Tracking like HubStaff.com

User Experience

Conventional Team Messengers have Navigation on the left, and Chatroom on the right like this

	Desktop	Mobile
Slack	File:Slack Desktop.png Slack Desktop. Left Navigation. Right chatroom	File:Slack Mobile.png Slack Mobile. Left Navigation. Right Chatroom
MS Teams	Anand to fill screenshots here from Google Images	Anand to fill screenshots here Google Images

A problem with Team Messengers is that everytime you have a chat, you later have to do extra work in filing your tasks, decisions, files from the chat into a Knowledge Base (like Confluence).

GetGuru.com tries to solve this. It integrates into the messengers (like Slack, Teams) and creates a Knowledge base for you (it offers a Wiki, or Knowledge base).

Our user experience offers

• A Messenger + (Self Organizing) Knowledge Base inside a single place.
• Not just a Knowledge Base, but also your "Workplace".

Down the line we will delve more into "Self Organizing" and "Workplace". But first let's see what the anatomy of our User experience will look like

Web	Desktop	Mobile
File:SC Office User Experience Anatomy.png SC Office Web User Experience: Left Navigation. Center Workplace. Right Chatroom	Navigation could be docked & pinned on the left of desktop User can continue to open whatever apps they want on their desktop. That becomes their workspace On every app, there would appear a "Chat Anywhere" button. Or this chat anywhere pane would appear docked with Navigation Panel	Multiple Hamburger to choose Navigation Once chosen Navigation appears on top bar Workspace or Chatroom appears next. Whatever was accessed by user last time Multiple hamburger to choose between Workspace and Chat

Navigation

N level Nested

Slack and Teams have a limit of 2 or max 3 level in Navigation. We would let N level. We would use Mediawiki category structure to let users create categories

Yet (nearly) Flat

You can see chat conversation and activity (in workplace) when you click on any node. You can see

1. Only the activity / chat in selected node
2. All the activity / chat nested inside the selected node

Navigation with Dashboard

see Navigation Panel in TalkingDB

for every "View" created inside article page, we would create an item in navigation panel too

Workplace

Wiki pages

Mediawiki Article pages will open as user clicks on a node in Navigation

Spreadsheet / Kanban / Tasks

While Mediawiki powered articles pages would be the default option in Workplace, but like in Clickup you would have option to see

Mindmap

See Mindmap and Graph Software

⚠️ Your favorite Desktop or Web App

See

• Desktop.com
• https://www.producthunt.com/products/rambox

Chatroom

Chatroom for every node

A chatroom will be created for every node selected inside navigation

Configure Channel & Filter against nodes

connected will create child nodes against certain category inside navigation

Labeling Chat messages

#upskilling should help HR team to know what needs their attention

While customers respond, enterprise search would be used to source answer from documents

Enterprise search will use Document Chatbot. For every incoming customer message, a question or query would be formed and passed to the Enterprise search engine. Where a response will be suggested

Queries would be be context sensitive

Enterprise search should decide the filters that must be applied on queries such as, only the documents that belong to

• Certain category be be responded from (lets say we had a category / article named "North India" then queries must be returned only from inside that space
• Only data from inside customer's search history would be responded to

These contexts would be applied by NL on certain intents such as

• Last order
• other intents that are easy to train

User doesn't has to waste time deciding the location where message goes

People start typing a message, and a list of channels are searched, and suggested where the message must go

People start typing the message based on a window they had open

Notifications

Using "near natural language" the notification would be split in following types

Where I am

a Driver / responsible

as of RACI

Timeboxing

See here

Organizing Knowledge

Chrome plugin to organize knowledge

Static URL

Time would be tracked when a navigation node is selected, and some internet history is used. I a URL is opened more than once while working while keeping the navigation open, then it's link would be added with "open count" posted against the URL. And stored inside Knowledge

Query String URL based on a customer parameter

experimental. don't code it yet until the idea matures

Prompt Guideline when Colleague is Working

Managers (CXO, PMs Or Leads) will be able to train CV to identify when Colleague is about to do a key action on their screen, like:

• commiting some code
• Posting a job on upwork
• Contacting a candidate over linkedin / email

A guideline article from wiki would be pulled up on the screen. Colleague would be able to either just read the guideline or mark the checklist in guideline wiki as complete.

The guideline would prompted on colleague's screen using an on-screen notification appearing just next to their caret/mouse/button where they were performaning this action

Talk to your Screenshots

1. Developers can ask the time taken by them on technologies, modules, and projects
   1. It can be used by developers to improve themselves
   2. it can be used to revisit some past work
   3. it can be used by sales teams for better time estimates
2. Such as we can ask "how much time was taken to implement SSO", "how much time was taken add dgraph"
3. It can also be used to audit timesheets for a filed claim
   1. Recently I (mayank) have to manually audit screenshots when a client claimed that we used time to learn basis such as git and redis. But the manual search to prove that the calim was invalid too large effort and could be saved by talking to screenshots

Enterprise Search

Inside the constantly increasing organized knowledge, an enterprise search would be performed. Any questions asked to the Document chatbot would be answered by

• crawling all the documents crawled inside the selected navigation node and answered
• learning from all the past conversations that happened inside selected navigation node and answered

Capabilities

Navigation to Workplace

For every navigation note, we would memorize the Workplace that was left last open. Maybe for a particular node the user prefers to see their Favorite App first, than the Wiki page. It would remember and open only that

Chatroom to Navigation

Chat messages to link to navigation node

While you are chatting you can do # and mention to a node in Navigation

⚠️ Automatic switching of Navigation Node

While chatting if you start to talk you went "off topic" and started to talk about a topic which goes in another Navigation node, then AI would automatically detect and nudge you to either MOVE or COPY your communication over there. And even JUMP you to the navigation node

Time Tracking

Track time based on what node is open

Notice what node is currently selected by user in the navigation. Track user's time inside that node.

Idle Time detection

Like they have in Hubstaff

Screen to Navigation-node

Uses CV (with NL and Document AI) to understand whats open on user's screen. And based on that automatically navigate the user to another node in navigation. See this video between Tushar and David

Chat to Task/OKR/Objectives CRUD

uses NL

Chat to Wiki

via Document Similarity

Success Metrics

The metrics with which we measure that SC Office is more helpful than MS Office

Users are reminded when KPIs aren't met

ICs

Engineers

Weekly: Release URL

Weekly: QA Report

Managers

All Managers

Weekly (start): Issue Board

Bi-weekly: QA Report

Bi-weekly: client updates

Project Managers

Weekly (end) : Money released

CXOs

Team has enough work

Team has received feedback