KENDRA CONNECTORS
Senior Experience Designer
the
PROJECT
Amazon Kendra customers want a simple GUI with validation to effectively minimize their workload when connecting data sources to their index.
Prior to the availability of the Amazon Kendra connectors, to crawl 3rd party data sources, customers required the technical prowess (coding JSON and/or YAML) to set up specific custom APIs outside the console, using Amazon Cloud Formation*.
*Non-technical users preferred the intuitive GUI of competitors such as MS Azure or Google Cloud Search...
​
Customers needed a feature that allows Kendra to crawl a selected 3rd party data source and index content to a single point of truth.
Experience Design
Interface Design
Visual Design
User Testing
what is a
CONNECTOR?
A Connector is set of instructions allowing Amazon Kendra to crawl selected 3rd party collaboration systems and index queried content to a designated S3 bucket.
The Amazon Kendra Connector provides the customer with an in-console GUI that supports content indexing from 3rd party collaboration systems, and a wizard that allows the customer to configure an easily customizable automated recurring crawl of designated data sources.
It works in 3 simple steps:
-
designate a home index (like an S3 bucket)
-
connect and configure an 3rd party data source
-
run your Amazon Kendra search
*For definitions select terms from the link menu on the right.
The finished product in the flesh...
my
ROLE
I joined the project as Senior Experience Designer, partnering with the PM, SDM and UX Director to design an in-console GUI with a wizard to enable and authorize Amazon Kendra to connect with a 3rd party collaboration system, and crawl it for content. I was engaged in every facet of the project's development including, process audits, personas development and user stories, workflows, wireframes, and high-fidelity prototypes through user testing.
Microsoft Office 365
Adobe CC
Asana
Figma
Responsibilities: Product design including, experience, interface and visual design, wireframing, prototyping and user testing.
feature
SCOPE
Customer needs to:
-
Select their desired data source (if applicable) from a library of collaboration systems,
-
Specify the entities to be crawled and content to be included/excluded from indexing.
-
Have the option to configure data sources as APIs to use in Cloud Formation.
-
Define the security and access specific to each data source.
-
Set up an IAM role to access repository credentials and designate an universal storage S3 bucket.
-
Configure and automate recurring instructions far Amazon Kendra to crawl connected data sources.
user
EXPERIENCE
As an Amazon Kendra customer, I want a simple GUI to enable me to effortlessly and quickly connect my organization’s data sources to my index, without having to leave the Kendra console or writing custom code.
​
Scenario: Configuring a Kendra connector to crawl and index 3rd party content that is important to me.
-
Given I’m on the Amazon Kendra data management page...
-
And I have configured an Kendra index to house my data sources of interest.
-
When I Click on the button “Add data source”...
-
I can pick a new data source from a library of collaborative systems and link it to my current Kendra index.
-
The data source sync enablement wizard runs directly in the console...
-
And went I use the step-by-step GUI to configure and customize and enable my Kendra crawler to sync with the new data source...
-
I see a notification that I have successfully added an new data source for Amazon Kendra to automatically and incrementally crawl for new and updated content
Administrator/Content Owner
​
I'm a data administrator and I want ingest documents to build my index and establish my search engine for my data analytics team.
​​
I want to spin up an index quickly, with clear validation so I can move on to the critical next steps with assurance. I want to easily connect my organization’s data sources to my index, so that I can ingest documents to build my index and establish my search engine
​
“What needs my immediate attention today?”
-
I want to spin up an index quickly, with clear validation so I can move on to the critical next steps with assurance.
-
I need a simple GUI that allows my team to minimize our workload when connecting data sources
-
I need to correctly configure security, with validation and viability to ensure this is done correctly
“I am am AWS search manager and I want a simple GUI with validation to effectively minimize their workload when connecting data sources to their index.”
Amazon Kendra console aministrator
my
PROCESS
Upon joining the Amazon Kendra design team I became the 4th "owner" of the 18 month-old Connector design process, which consisted of a vast array of Sketch libraries, pages, art-boards and symbols recently translated into Figma. This tool transition, along with the rapid turnaround of junior designers and interns, made for a disarray of conflicting design methodologies regarding the project UI. Instead of simply releasing the next round of 3rg-party specific Connectors, salvaged from bits and pieces of previous versions, we had to take a step back to re-evaluate the design process.
Pouring through extensive user research, product usability testing and clearly defined requirements and workflows, it was time to build out an equally solid master Figma library of components and presets. With the generic shell of a configuration wizard as a foundation, designers can quickly "fit" a custom Connector to unique content picking presets and components from a vetted master library.
​
As the new Design Lead, I readily accepted the task of building, maintaining and updating this "living" repository while also being hands-on in the process of designing new Connectors, 3 weeks at a time, and on-boarding the next rotation of designers/interns.
As a corner-stone to my design process consisted of IDEO's 6 phases: Empathize - Ideate - Design - Test - Iterate - Implement (and Repeat)...
design
SPECIFICS
-
Create a component library/design system master to streamline the connector design cycle.
-
Design an in-console multi-step data source enablement and index configuration wizard.
-
Create a paginated menu library of compatible collaborative systems to which Amazon Kendra can connect. Also, show already connected data sources in title or list arrangement.
-
Leverage radio buttons, multi-select drop-downs and auto-suggest for immediate validation.
-
DON’T leave the wizard when linking data source access - use modals for logins and authentications.
-
Use Polaris approved on-screen messaging patterns to communicate/confirm actions.
visual
IDENTITY
I frequently referenced AWS brand strategy and attributes when developing the visual language and component interaction. From font pairing over color theory to illustration style, building upon Polaris, AWS' well established (and recently updated to version 3.0) design system helped me create a holistic experience that felt trusted, intuitive and familiar at each touch-point.
AMAZON Squid Ink
AMAZON Orange
Blue 600
Blue 300
Blue 100
Red 600
Orange 600
Green 600
user
EXPERIENCE
-
Designate an index to collect content from data sources - commonly, an Amazon S3 bucket.
-
Scan connector library to select the 3rd party data source of interest for Kendra to crawl.
-
Configure custom instructions for content crawling and enable the connector.
-
Edit, automate and monitor recurring crawls for every collaboration system important to your organization.
the
PROTOTYPE
-
I have configured an Kendra index to house my data sources of interest (such as an S3 bucket).
- I click on the button “Add data source”...
-
I pick a new data source from a library of collaborative systems and link it to my Kendra index.
-
The data source sync enablement wizard runs directly in the console...
-
And I use the step-by-step GUI to specify data source details...
-
I then define the access and security to my new data source...
-
I configure my sync settings...
-
I have the option to set field mappings...
-
I review my configuration settings and activate my connector...
-
I see a notification that I have successfully added an new data source for Amazon Kendra to automatically and incrementally crawl for new and updated content.
project
SUCCESS
-
Originally slated as a Tier 3 (internal-facing feature refresh), Increasing popularity and service requests among the non-technical user group, caused AWS to categorize it a Tier 2 (external feature update).
-
Through a well maintained master component library, connectors can now be built from a generic foundation, using case-specific presets and revised copy.
-
With 20 data source connectors already synchronized and another 8 in production, the Kendra AI/ML team has scheduled an additional 12 collaboration systems to be crawled by Q3, 2022.
-
Technologically savvy customers who prefer linking 3rd party data sources to Kendra, using JSON and/or YAML code in Cloud Formation, increasingly turn to the GUI for linkage exploration of specific collaboration systems.
