# Innovate in 48 - Kumo Kaisers

## Who are the Kumo Kaisers?
The Kumo Kaisers are comprised of members from the public cloud services engineering and project management functions. IE Script Kiddies.

| Member | Role | Location |
| -------- | -------- | -------- |
| Matt Betts | Lord and Master | Jacksonville, FL
| Philip Balcar | C# Hactivist | Seattle, WA
| Toni Goessman | SNOW Queen | Tucson, AZ
| Blair Paine | A Recovering Consultant | London, England
| Vladimir Shvets | Just Vladimir | Jacksonville, FL

## Solution Name
Visual Research Assistant

## Describe your solution; What problem you are solving
It is very time consuming to research an error message, whether it comes from the system, middleware, or within an application. When people are experiencing problems, they expect a solution in the shortest amount of time possible. The Visual Research Assistant does the heavy lifting of taking the important parts of a screenshot and parsing it into usable text so ServiceNow and the internet can be searched rapidly for related resolutions. The results can then be attached to a SNOW ticket, which enhances the application for future similar errors.

## Solution Features
- Captures a screen shot within the app, so no additional tools needed
- allows for automatically uploading the screen shot to a ServiceNow ticket
- Prevents PII data for accidently being attached to a ticket
- performs rapid searches of multiple sources based on keywords
- continues to expand knowledge reuse, by attaching results to the ticket
## Technologies and architecture used
- [Azure Logic Apps](https://docs.microsoft.com/en-us/azure/logic-apps/)
- [ServiceNow API](https://docs.microsoft.com/en-us/connectors/service-now/)
- Azure Cognitive Services
- [Optical Character Recognition (OCR)](https://docs.microsoft.com/en-us/azure/search/cognitive-search-skill-ocr)
- [Text Analytics PII Recognition](https://docs.microsoft.com/en-us/azure/search/cognitive-search-skill-pii-detection)
- [Key Phrase Extraction](https://docs.microsoft.com/en-us/azure/search/cognitive-search-skill-keyphrases#:~:text=The%20Key%20Phrase%20Extraction%20skill,talking%20points%20in%20the%20record.)
- Bing Web search
- C# .NET framework UWP application platform
- Windows 10 Forms

## What your code is designed for

The Kumo Kaiser app is designed to enable operations teams to leverage the power of using screenshots to search existing knowledge base articles and related incidents within ServiceNow, as well as performing web searches on the internet. Besides retrieving the data, it allows the user to seamlessly upload the image to their open SNOW ticket. Once the text is converted and the ticket updated, the keywords become searchable in the future.

FIS is a highly regulated organization, often dealing with PII and other sensitive information. The app will first scan the screenshot, and in the event, PII data is detected, the image is rejected, and the process exits with no further action.

If the image does not contain PII data, and if the user specified there is an open SNOW ticket, the image is uploaded into the work notes of the ticket.

The text is then passed through a Keyword extractor. Based on the keywords, ServiceNow Incidents and Knowledge articles are reviewed and an internet search is performed to find related information. If the user requested their SNOW ticket updated, the results are attached to the ticket.

## What your code was written in
- C#
- Azure Logic App JSON

## Open source or proprietary software used (if applicable)
- Universal Windows Platform (UWP)
- JSON.NET

## Why it's cool

The Kumo Kaisers did not need traditional development skills and instead made use of readily available range of APIs exposed as connectors.

Comprised of operations engineers, the team set out to deliver a solution leveraging as much no-code/lo-code technology as possible.

This allowed for rapid deployment of a complex solution, leveraging cloud native artificial intelligence technologies. (Plus, it sharpened our Googling skills).

From an engineer's perspective, this solution allows for significantly more robust documentation on SNOW Incidents with very little overhead. When producing Root Cause Analysis (RCA), there is a real-time stream of visual events as opposed to manually finding and attaching screenshots and notes through-out the incident triage process. This allows knowledge to be readily reused again and again.

## WOW Factor
> According to the Global Knowledge Team, reuse of knowledge allows cases to be solved faster, with 50 - 60% improved time to resolution and a 50% increase in first call resolution. Resources are used more effectively, with a 70% improved time to proficiency, a 20 - 35% improved employee retention and, a 20 - 40% improvement in employee satisfaction.