Cisco AI Assistant
Design Strategy |
Trust and Security Management with AI
The problem
Trust: Organizations want to ensure that their data and assets are securely protected to prevent theft, financial losses, and reputational damage resulting from ransomware and other cyber-attacks.
This involves not just the implementation of robust security configurations but also the assurance that these implementations align precisely with their intended security policies.
Eliminating human error is crucial here, as it can lead to vulnerabilities or downtime that jeopardize security.
A major problem arises when IU network engineers are hesitant to place their trust in AI to streamline their workflow and assist in security tasks.
Cisco Mentors
Aanjan Ravi (Product Management)
Laura Tudorache (Product Design)
Molly Bloom (User Experience Research)
Akshay Parakh (Product Design)
Animesh Singh (Product Design)
Feixing Tuang (Product Design
Mentors
Professor Shuying Belivs
Associate instructors
Divya Pandey
Siyona Michale
Team Members
Apurva Patil
Neha Goswami
Nishant Mahendra Prasad
IU Network Engineers
Matt Malott
Jeff Ambern
Project Overview
Time
3 Month
Tools Used
Figma
Figjam
Google Docs
Process followed
For our project to develop a design strategy, we tackled the issue that IU network engineers had with trusting Cisco's AI assistant. To address this, we began with secondary research to deepen our understanding of the AI assistant.
We engaged with stakeholders for a broader perspective and directly spoke with the IU network engineers, the primary users of this software. By adopting an iterative approach, we continuously gathered and integrated feedback from all stakeholders throughout the project. This helped us refine our strategy effectively.
Secondary Research Insights
Our research on using AI in network security has shown that working together, always making improvements, and sharing knowledge are very important.
Enhanced Problem-Solving: Collaboration features enable network engineers to leverage collective expertise when addressing complex security challenges, leading to more comprehensive and effective problem-solving.
Iterative Improvement: Feedback mechanisms allow for the continuous refinement of AI models based on the performance metrics observed during simulations, leading to increasingly effective security measures over time.
Cultural Shift: Fostering a culture of knowledge sharing can drive innovation and employee engagement. Recognizing contributions can encourage network engineers to take an active role in improving AI-driven security.
Gaps in Existing Solution
01
Transparency and Explainability of AI-Driven Insights
While some solutions incorporate AI-driven analytics for threat detection and response, there may be gaps in transparency and explainability regarding how AI algorithms reach their conclusions.
Opportunity:
Prioritize transparency and explainability in its AI-driven features, providing users with clear explanations of AI-generated insights and recommendations. This can help build trust and confidence in the system's capabilities.
02
User-Friendly Interface and Workflow Customization
Some existing solutions may have complex interfaces and workflows, leading to usability challenges for users, particularly in large enterprise environments.
Opportunity:
Focus on user-friendly interface design and customizable workflows. By prioritizing intuitive navigation, contextual help, and flexible configuration options, Cisco can enhance user experience and adoption rates.
Why should you care?
Risk Mitigation
By understanding AI decision-making and refining models iteratively, businesses can effectively mitigate risks, reducing the likelihood of security breaches and associated damages.
Efficiency Gains
Implementing these insights leads to more efficient security operations, as collaborative problem-solving and iterative model refinement streamline processes and reduce manual effort.
Innovation and Adaptability
Embracing knowledge sharing and continuous improvement fosters innovation, enabling organizations to swiftly adapt to evolving security challenges and maintain a proactive stance against emerging threats.
Our Mission
To promote trust in AI-driven solutions by enhancing transparency and explainability in AI decision-making processes for network engineers.
Meet Katherine, a Network Engineer at IU.
Our strategy is tailored to empower professionals like Katherine, network engineers seeking to enhance security trust and streamline operations with advanced AI-driven solutions.
Katherine
Network Engineer
Age: 28
Background
Katherine holds a Bachelor's degree in Computer Science with a focus on cybersecurity. She has been working as a network engineer for the past 5 years, specializing in designing, implementing, and maintaining secure network infrastructures for large organizations.
Tech Proficiency
Proficient in networking technologies such as firewalls, intrusion detection systems, and VPNs.
Experienced in implementing security measures to safeguard against potential risks.
Utilizes AI-driven solutions to augment security operations, leveraging advanced technologies for threat detection, anomaly identification, and proactive risk mitigation.
Challenges
Reliability Concerns: Doubts about AI's reliability and accuracy may hinder trust in its recommendations for critical security tasks..
Interpretability Challenge: AI's opaque decision-making process can create uncertainty, making it difficult to understand and validate recommendations.
Goals
Katherine strives to enhance her organization's security using AI-driven solutions, streamlining operations and mitigating emerging threats while promoting collaboration and continuous improvement within her team.
Our Vision
To revolutionize network security by empowering network engineers with advanced AI-driven solutions that ensure transparency, collaboration, and trust in decision-making processes.
Testing our MVP with Matt from IU UITS
We had the opportunity to test out our initial ideas with Matt from IU UITS and gather his feedback on the concepts for further iterations.
Feedback
Decision Tree Feature: Users (Matt from IU UITS) appreciated the clarity provided by the decision tree feature, allowing them to understand how decisions are made.
Collaboration Aspect: The collaborative features, particularly in-line comments, were well-received, enabling effective teamwork and communication.
Citation Sharing: Users liked the option to share citations, facilitating broader participation in discussions and knowledge sharing.
How does this strategy increase the trust and confidence of network engineers in AI?
Informed Decision-making:
AI Insights Panel provides detailed explanations, meter/confidence scores, and decision tree visualizations, enabling network engineers to understand AI-driven security decisions thoroughly.
Real-Time Guidance:
AI Guidance Panel offers real-time suggestions and best practices, supported by customer feedback ratings, facilitating efficient rule creation and modification for network engineers.
Collaborative Workspace:
Collaborative Incident Analysis features enable real-time collaboration and knowledge sharing among network engineers, enhancing teamwork and problem-solving.
Final Solution