A Tale of Two Businesses : Cisco AI & AnytimeTV


The Prompt 


AnytimeTV is an Organization which has a platform to sell their Broadcasting services to Consumers. AnytimeTV has an online portal where Agents/Admins/Business owner’s login to do their respective tasks on a daily basis. You are expected to design views for all three user types to help them accomplish their tasks daily.



  • Agents (market and sell packages to Consumers)

  • Admins (monitor uptime, manage packages, Consumers, Agents, Bots)

  • Business Owners (monitor revenue, subscriptions, new markets


Understanding The User

My approach to deal with ambiguity is to ask a lot of questions, so with the aim of performing primary user research, I set out to contact people in Sales & Marketing roles (Agents), Operations roles (Admins) and Finance roles (Business Owners) on LinkedIn. However, due to slow response, I failed. Then I realized that my biggest challenge was that I couldn’t contact real users. 


So, how did I change my design direction?


I pivoted to perform secondary research. Due to time constraints, I focused my efforts in understanding user needs and behaviors of Admins & Agents


For the user types, I found articles and company docs that have created personas in similar roles at companies like SumoLogic and Salesforce. I read multiple posts on online forums like Quora and Reddit that crowd sourced answers to questions like “What are the biggest pain points of a sys admin?”


Building The Persona

Based on my research findings, I aggregated information into user personas to represent people I will be designing for. I refined the users to represent specialized roles as

Admin → Sys Admin

Agent → Sales Rep

admin persona.png
agent persona.png

Mapping the user flow

While keeping the user goal in mind, I decided to tackle one major use case for every user type. I walked in the user’s shoes and mapped out the user flow.

What technology can help improve the current experience?

Now that I had mapped out user flows, I set out my design goal to improve the current user experience to achieve their goals. My Technologist instincts kicked in, and I decided to understand technologies that similar platforms use to advance experiences of their users

I learnt that current platforms mine data heavily and use AI services to improve user experiences : 


  1. Cognitive Services 

  2. Predictive Modelling 

  3. Recommendation Engines

  4. Dashboard views & Analytics 


How can Cisco AI help AnytimeTV?

With the design goal of using AI to improve the current UX, I explored Cisco AI products to see how it can provide services to AnytimeTV. 

Admin Experience

I started the admin experience by incorporating Cisco AI services.


After mapping the experience, I brainstormed key screens and layouts, first on paper and finally in Sketch


Link to Admin Wireframe


Agent Experience

Next, I incorporated Cisco AI into the Sales rep experience.


After mapping the experience, I brainstormed key screens and layouts, first on paper and finally in Sketch


Link to Agent Wireframe


User Testing, Setting KPIs and More​ 

Due to the time limit, many design decisions in this design challenge were made based on secondary research and assumptions. If time allows, more user research and task-based user testing should be done to get solid insights and feedback.

  • Do first time users feel the flow intuitive and easy to use?

  • Designing for edge cases

  • Designing for accessibility

  • Setting KPIs for business outcome


Then based on the feedback from real users, more iterations should be conducted to polish the design in the future.


Design Implications : Ethics in AI

As a UX designer, I always try to think critically of my designs. While AI can empower users, there is currently various research being conducted in academia and industry around Ethics in AI. As for my design, there can be various questions to be answered :

  • Would the system become so automated that Admins & Agents lose their jobs?

  • Would privacy be a concern if AI assistants are listening into conversations with leads?

  • Will Admins become so dependent on assistants for diagnosis that they lose their critical thinking ability/ make them lazy?

  • What if the algorithm is trained incorrectly? Would this cause the entire system to be at risk?

  • What if the algorithms are biased?

Hey, you reached the end. Thank you for reading!

San Francisco Bay Area