Work

About

Bloxy AI Copilot

Adopt Generative AI to assist network administrators and security analysts with configuration, product learning, and decision making through quick answers and data enrichment.

Adopt Generative AI to assist network administrators and security analysts with configuration, product learning, and decision making through quick answers and data enrichment.

Adopt Generative AI to assist network administrators and security analysts with configuration, product learning, and decision making through quick answers and data enrichment.

Adopt Generative AI to assist network administrators and security analysts with configuration, product learning, and decision making through quick answers and data enrichment.

Role
Role

Lead UX Designer

Lead UX Designer

Timeline
Timeline

October 2023 - March 2024

October 2023 - March 2024

Team
Team

1 UX Designer, 1 UX Researcher, 2 Product Managers, 8 Data Scientists & 3 UI Engineers

1 UX Designer, 1 UX Researcher, 2 Product Managers, 8 Data Scientists & 3 UI Engineers

Challenge

We were tasked to explore the potential of integrating an Generative AI feature into our product. The Bloxy AI assistant would help users understand granular security insights and provide contextual, personalized help. With this new feature, we hope to improve feature adoption, enhance operational efficiency, and keep pace with the fast-evolving industry.

Impact

I spearheaded the UX research and design of the product feature from 0 to 1, shaping the entire user experience. By presenting UX innovations to executive leadership, I successfully secured buy-in from key stakeholders to integrate GenAI into our product.

Challenge

We were tasked to explore the potential of integrating an Generative AI feature into our product. The Bloxy AI assistant would help users understand granular security insights and provide contextual, personalized help. With this new feature, we hope to improve feature adoption, enhance operational efficiency, and keep pace with the fast-evolving industry.

Impact

I spearheaded the UX research and design of the product feature from 0 to 1, shaping the entire user experience. By presenting UX innovations to executive leadership, I successfully secured buy-in from key stakeholders to integrate GenAI into our product.

0-1

First GenAI Feature in Infoblox

80

Users adopted Bloxy

65%

Increase in User Satisfaction

OVERVIEW

Pioneering Generative AI feature in Infoblox

Our top customers have inquired about whether we are introducing AI features into our product. In response, we clarified the requirements and developed an MVP feature through multiple iterations and usability testing with Infoblox customers. This approach allowed us to evaluate the usability, precision, accuracy, and potential of our GenAI feature.


Business Goals

Summarize Help Documents within our platform

Help users resolve configuration-related issues without navigating to other platforms or search engines. Provide quick summaries of their configuration questions to reduce customer support tickets and free up our team to focus on more pressing issues that need human intervention.

Utilizing AI to empower data-based decision-making

Our team was tasked with exploring the possibility of incorporating AI and addressing pressing challenges related to onboarding, understanding security events and misconfiguration issues.


Solution Summary


My Contribution

As the lead designer for an AIOps design feature, I played a key role in shaping the project from conception to execution. I spearheaded the creation of the experience vision and strategy, ensuring our approach was aligned with organizational goals and customer needs.

Presenting to executive leadership

I developed detailed prototypes that brought our vision to life. By presenting comprehensive customer insights and these prototypes to the executive leadership team, I conveyed the value and potential impact of our design, sparking excitement and stakeholder buy-in.


Skip to the final design deliverables

OVERVIEW

Pioneering Generative AI feature in Infoblox

Our top customers have inquired about whether we are introducing AI features into our product. In response, we clarified the requirements and developed an MVP feature through multiple iterations and usability testing with Infoblox customers. This approach allowed us to evaluate the usability, precision, accuracy, and potential of our GenAI feature.


Business Goals

Summarize Help Documents within our platform

Help users resolve configuration-related issues without navigating to other platforms or search engines. Provide quick summaries of their configuration questions to reduce customer support tickets and free up our team to focus on more pressing issues that need human intervention.

Utilizing AI to empower data-based decision-making

Our team was tasked with exploring the possibility of incorporating AI and addressing pressing challenges related to onboarding, understanding security events and misconfiguration issues.


Solution Summary


My Contribution

As the lead designer for an AIOps design feature, I played a key role in shaping the project from conception to execution. I spearheaded the creation of the experience vision and strategy, ensuring our approach was aligned with organizational goals and customer needs.

Presenting to executive leadership

I developed detailed prototypes that brought our vision to life. By presenting comprehensive customer insights and these prototypes to the executive leadership team, I conveyed the value and potential impact of our design, sparking excitement and stakeholder buy-in.


Skip to the final design deliverables

Jump to final design

Jump to final design

Jump to final design

Jump to final design

RAPID ITERATION CYCLES


Our Approach

I facilitated collaborative design and UX canvas sessions to work with stakeholders to identify the business goals and problems we’re solving.

Given the novelty of Generative AI (though it's become a buzzword), capturing customer reactions to a prototype is more valuable than conducting extensive generative research, as customers can't imagine what they haven't experienced.

We started of writing down the benefit hypothesis of the product features and validated our hypothesis through multiple rounds of iterations and usability tests.


Learning about the different dimensions of AI

To become a more effective AI designer, I took the time to take up AI learning outside of work and received a certification for a course " Product management for AI/ML" from ELVTR. This helped me gain a deeper understanding of the strengths and considerations to develop responsible AI.

Understanding user perception of AI

We conducted 17 customer interviews and market research to drive our planning phase. Our customer interviews included a blend of discovery-type questions and usability tasks, aimed at understanding users' perception of GenAI and iterating on the initial proposed designs.


Key insights that defined the MVP

RAPID ITERATION CYCLES


Our Approach

I facilitated collaborative design and UX canvas sessions to work with stakeholders to identify the business goals and problems we’re solving.

Given the novelty of Generative AI (though it's become a buzzword), capturing customer reactions to a prototype is more valuable than conducting extensive generative research, as customers can't imagine what they haven't experienced.

We started of writing down the benefit hypothesis of the product features and validated our hypothesis through multiple rounds of iterations and usability tests.


Learning about the different dimensions of AI

To become a more effective AI designer, I took the time to take up AI learning outside of work and received a certification for a course " Product management for AI/ML" from ELVTR. This helped me gain a deeper understanding of the strengths and considerations to develop responsible AI.

Understanding user perception of AI

We conducted 17 customer interviews and market research to drive our planning phase. Our customer interviews included a blend of discovery-type questions and usability tasks, aimed at understanding users' perception of GenAI and iterating on the initial proposed designs.


Key insights that defined the MVP

DESIGN & EVALUATION


We validated our proposed designs with 16 customers.

Through these tests, we:

  • Evaluated the usability of the proposed PoC

  • Assessed the validity of answers for different categories and observed the users' reactions to the answers.

  • Gained a deeper understanding of the potential use cases.

  • Evaluated the usefulness of the prompts and answers for each category.


Key Insights from usability studies

Bloxy is an easy way for quick answers

Users expect the AI assistant to provide an easy way to seek quick answers and do not want to spend time deciding what to do. We adjusted our design based on user feedback and created style guides for developers.


Users want to form their own conclusions

Instead of immediately updating a configuration for users, we should allow them to review the alerts and make their own decisions based on the evidence.


Flexibility and control over config settings

Users want the flexibility to adjust recommended settings to align with their corporate best practices when making configuration changes.

DESIGN & EVALUATION


We validated our proposed designs with 16 customers.

Through these tests, we:

  • Evaluated the usability of the proposed PoC

  • Assessed the validity of answers for different categories and observed the users' reactions to the answers.

  • Gained a deeper understanding of the potential use cases.

  • Evaluated the usefulness of the prompts and answers for each category.


Key Insights from usability studies

Bloxy is an easy way for quick answers

Users expect the AI assistant to provide an easy way to seek quick answers and do not want to spend time deciding what to do. We adjusted our design based on user feedback and created style guides for developers.


Users want to form their own conclusions

Instead of immediately updating a configuration for users, we should allow them to review the alerts and make their own decisions based on the evidence.


Flexibility and control over config settings

Users want the flexibility to adjust recommended settings to align with their corporate best practices when making configuration changes.

Bloxy AI Copilot 1.0

Adopt Generative AI to assist network administrators and security analysts with configuration, product learning, and decision making through quick answers and data enrichment.

KEY FEATURES

Immediate Contextual Help

Users can immediately navigate to the context and view Bloxy's responses while performing actions simultaneously.

Immediate Contextual Help


Empower decision making without taking over

Bloxy offers personalized insights and seamless issue resolution based on user configurations. We provide the data to help users make informed decisions and assist them with configuration updates, offering users flexibility and control.

Empower decision making user flow


Easy access to conversation history

Users can search, view, and bookmark previous chats, as they may encounter the same issues. Instead of redoing the chat, they can quickly find the answer by going back. Easy access to conversation history is also essential for auditing and change management purposes.

Conversation History


Improve experience with feedback

Customer feedback allows us to incorporate user input into GPT, enhancing the accuracy and precision of its output. After iterating on multiple designs, we chose to display a thumbs-down icon below each response, as it proved to be the most intuitive option for users.

We will also include a link that directs users to contact customer support if they require further assistance after submitting the feedback form.

Feedback workflow

KEY FEATURES

Immediate Contextual Help

Users can immediately navigate to the context and view Bloxy's responses while performing actions simultaneously.

Immediate Contextual Help


Empower decision making without taking over

Bloxy offers personalized insights and seamless issue resolution based on user configurations. We provide the data to help users make informed decisions and assist them with configuration updates, offering users flexibility and control.

Empower decision making user flow


Easy access to conversation history

Users can search, view, and bookmark previous chats, as they may encounter the same issues. Instead of redoing the chat, they can quickly find the answer by going back. Easy access to conversation history is also essential for auditing and change management purposes.

Conversation History


Improve experience with feedback

Customer feedback allows us to incorporate user input into GPT, enhancing the accuracy and precision of its output. After iterating on multiple designs, we chose to display a thumbs-down icon below each response, as it proved to be the most intuitive option for users.

We will also include a link that directs users to contact customer support if they require further assistance after submitting the feedback form.

Feedback workflow

IMPACT

Unraveling the possibilities of Bloxy AI

I collaborated with PMs, UI engineers, and the Data Science team to build the MVP. As we improve answer precision and accuracy, we’ll explore new ways to leverage the Bloxy AI Assistant to address customer pain points.

We met the initial business and user goals and will continue iterating based on customer feedback. Bloxy AI is now a key initiative for FY25! 🎇 The next generation of Bloxy will be unleashed in FY 25.4.

IMPACT

Unraveling the possibilities of Bloxy AI

I collaborated with PMs, UI engineers, and the Data Science team to build the MVP. As we improve answer precision and accuracy, we’ll explore new ways to leverage the Bloxy AI Assistant to address customer pain points.

We met the initial business and user goals and will continue iterating based on customer feedback. Bloxy AI is now a key initiative for FY25! 🎇 The next generation of Bloxy will be unleashed in FY 25.4.

REFLECTION

1) Mastering the domain makes me a more effective designer

As a designer new to working on Generative AI product features, I was eager to explore all the possibilities AI offers. Initially, I felt at a disadvantage because we were given requirements based on what developers and data scientists wanted to build.

I wanted to be more involved in shaping the roadmap, and to do that, I needed to understand AI concepts deeply and learn to speak their language. So after weeks of AI training that I've put myself through, I could finally understand the PM's concerns and the technical limitations we had at hand. Most importantly, I was able to understand how we can leverage the strength of GenAI in product development.

2) Think beyond the dataset and focus on customer benefits

When the project first landed on my desk, our PM and engineering teams were focused on executing based on the available data, with initial use cases revolving around summarizing configuration documentations and generating the number of security events.

Through interviews with customers and stakeholders, I used customer insights to help the team understand users' perceptions, expectations, and concerns regarding Generative AI.

Eventually, we shifted our mindset from "How can we make the answers more precise?" to "How can we bring value to our customers by providing actionable insights and make this a profitable feature?".

REFLECTION

1) Mastering the domain makes me a more effective designer

As a designer new to working on Generative AI product features, I was eager to explore all the possibilities AI offers. Initially, I felt at a disadvantage because we were given requirements based on what developers and data scientists wanted to build.

I wanted to be more involved in shaping the roadmap, and to do that, I needed to understand AI concepts deeply and learn to speak their language. So after weeks of AI training that I've put myself through, I could finally understand the PM's concerns and the technical limitations we had at hand. Most importantly, I was able to understand how we can leverage the strength of GenAI in product development.

2) Think beyond the dataset and focus on customer benefits

When the project first landed on my desk, our PM and engineering teams were focused on executing based on the available data, with initial use cases revolving around summarizing configuration documentations and generating the number of security events.

Through interviews with customers and stakeholders, I used customer insights to help the team understand users' perceptions, expectations, and concerns regarding Generative AI.

Eventually, we shifted our mindset from "How can we make the answers more precise?" to "How can we bring value to our customers by providing actionable insights and make this a profitable feature?".

Next →
Boost Productivity with Infoblox's Search
Next →
Boost Productivity with Infoblox's Search
Next →
Boost Productivity with Infoblox's Search
Next →

Boost Productivity with Infoblox's Search

Let's connect!

kadenceyy@gmail.com

LinkedIn

Dribbble

Instagram

© 2024 Designed by Kadence Tang

Let's connect!

kadenceyy@gmail.com

LinkedIn

Dribbble

Instagram

© 2024 Designed by Kadence Tang

Let's connect!

kadenceyy@gmail.com

LinkedIn

Dribbble

Instagram

© 2024 Designed by Kadence Tang

Let's connect!

LinkedIn

Dribbble

Instagram

© 2024 Designed by Kadence Tang

kadenceyy@gmail.com