Product Design and User Experience
Product Design and User Experience
In my 5+ years at Gridspace, we focused on improving interactions between customers and companies. Our focus was to elevate contact center experiences using Agentic AI’s and improve internal processes for the world’s largest consumer brands using the latest ML and proprietary LLM technologies.
This browser-based IDE allowed users to edit pre-made call handling flows from popular templates developed in-house by our team of conversation designers or create new ones based on specific customer needs.
Each node in the flow, when clicked, opens a panel that allows for configuration. In this view, users are able to set parameters for a new Grace agent.
A playbook is a set of instructions for handling customer requests and answering questions specific to the business.
One approach is to configure a Grace agent to be broader, covering main business functions, while handing off to more specific Grace agents finely tuned for nuanced inquiries.
After a Playbook is created you can interact with a Grace agent and immediately start coaching it. This view shows how to correct for speech-to-text errors and correcting LLM outputs.
Many Grace agents can be set up to loop between themselves, ensuring the caller is handed off to the Grace agent best suited for their needs.
Setting up a call center in 5 minutes was our mission. This concept showed how Gridspace Flow worked with a preset Grace agent in a walkthrough experience. The goal was for the user to customize a few parameters and easily tailor a template for their needs before creating an account.
What is going on in my company today? This was the single question every executive wanted to know. Pulse was designed to identify live calls from negative trend sets in order to stay aware of developing issues. The live dashboard also displays daily metrics for contact center activity and other identified trends for continued monitoring.
Clicking into a trend card would serve up all the conversations that a user can choose to review as well as export graphs for reports.
I was able to spend time with managers who oversaw teams of agents and map their process for coaching and ensuring quality and compliance.
This view would monitor the progress of individual managers' work queues and provide metrics for reporting.
In addition to viewing the performance of the department as a whole, managers were able to view individual call specialists and gauge progress.
Using the latest ML and AI technologies, each call was processed, summarized for intent and emotional valence, and labeled in real-time.
This view’s intent was to provide a historical snapshot of a customer’s interaction with the company, tracking key events as well as a summary of all conversations.
This view allowed users to pose questions and then interact with the models output. Very similar to Chat GPT but trained specifically on customer data with a proprietary LLM.
This feature allowed the creation of outbound caller campaigns targeting current or prospective customers by enabling Grace agents to help automate business development departments.
Each campaign was tracked and configurable - such as the number of allowable call attempts and the definition of call time windows to deter abuse of the feature.