top of page
hero image.png

UX LEADERSHIP , STRATEGY AND DESIGN

Leading the enterprise AI transformation for customer support across 5 amazon business verticals

  • Spearheaded the UX of 6+ cross-vertical and verticalized AI solutions aimed at improving business productivity.​

​

  • AI solutions resulted in annualized savings of over $8.624 Million.​

​

  • Improved support ticket handling time by 3.32%

​

  • Expanded across multiple markets UK, CA, IN, AU, SG, ZA, AE, IE, FR, IT, DE, ES

​

​

verticals.png

Created and operationalized
AI frameworks through cross functional workshops

I led BXT (Business, Experience, Technology) workshops to align stakeholders on AI opportunities.

These sessions translated business goals, user needs, and technical constraints into the right AI solutions

Group 1073717043 (1).png

Problem:

The current reactive customer service approach and lack of proactive account monitoring make it difficult for both customers and Amazon to optimize business account performance.

Solution:

H360 addresses this by continuously monitoring account health using AI and 80 data points, automatically identifying issues and triggering targeted interventions before they impact business operations.

Group 1410093882.png
Group 1410093883 (2).png

Collaborated with Data Science team to identify more than 80 data points to identify risks


Partnered closely with data science to define and validate high signal data points that enable accurate, AI driven risk prediction
 

Group 1410093379.png

H360 account health dashboard for account managers

The account health dashboard gives account managers a real-time, AI informed view of customer health such as account health insights, trends and metrics.

Group 1410093886.png

Problem:

Customer support agents handle tickets with fragmented context - long conversation threads, multiple handoffs, attachments and prior tickets which leads to increased resolution times.


Solution:

AI-powered summarization gathers all the relevant information from multiple sources and interactions into a clear structured problem statement.

image 9347.png
image 9348.png

Scaled across 12 + markets and 4 business verticals

Scaled across across multiple markets such as UK, CA, IN, AU, SG, ZA, AE, IE, FR, IT, DE, ES.
Scaled across across different verticals such devices and subscriptions.

Group 1410093887.png
Group 1410093888.png

Collaborated with Data science to define LLM output

Partnered with data science team to design the LLM output format based on user research with explicit safegaurds for model hallucinations and accounting for edge cases.

Group 1410093889 (4).png

Problem:

After a customer call, agents spend significant time on repetitive after-call work - documenting tickets, creating follow=ups and emails to the customer etc. This leads to increased resolution times.

​

Solution:

Agentic AI system automates after-call work by reasoning over call context to decide and exexute follow-up actions - emails, follow-ups, ticket annotations while keeping agents in control.

This Was Only a Glimpse.

I would love to share the full version with you. Connect with me on Linkedin, email me at

ganeshkarthik913@gmail.com

bottom of page