Akshay Khamgaonkar

Scenario: An AI agent is deployed in a telecom company’s customer service system. Its role is to analyze incoming customer support messages and triage them: auto-resolving basic queries, routing technical issues to relevant specialists, and flagging high-risk or potentially angry customers for escalation to a supervisor. One day, the AI mistakenly escalates a minor billing question to the company’s crisis response team due to misclassifying the tone of the message as threatening. This leads to an unnecessary internal incident response, confusion for the customer, wasted staff hours, and reputational damage when the customer shares the interaction online. Who's Responsible? Primary accountability may lie with the product/deployment team for failing to implement human review or setting too low a threshold for escalation. Secondary responsibility could lie with the AI design team, if the model had known deficiencies. Systemic accountability rests with the organization, for not having oversight, training, and escalation policies in place. Design Safeguards for Transparency and Traceability Decision Logging - Every AI decision should be logged with context Explainable Interface - Provide human-readable justifications for escalations Human-in-the-Loop (HITL) - Require a human reviewer to approve the AI’s recommendation Feedback Loops - Feed this outcome back into the training dataset, track false positives and use them for retraining. Clearly Define Roles and responsibilities

A great example of balancing multiple objectives is in Customer support chatbots, where the system needs to minimize response time (for efficiency) while maximizing customer satisfaction (for quality). These two goals often conflict—fast responses might lead to incomplete answers, while highly personalized responses might take longer. To Handle this situation, we can prioritize the query types into Simple queries to prioritize Speed and Complex or emotional queries to prioritize satisfaction. We may introduce some Smart rules like - 1. If customer sentiment is negative or confused, favor satisfaction. 2. If queue length is high and query is low-impact, favor speed. 3. Never exceed a max response time of XX seconds, even when optimizing for satisfaction.

Exception handling in invoice processing, AI can improve exception handling by analyzing patterns in frequent exceptions, such as mismatched invoice amounts or missing POs, and learning from human interventions. By tracking key data like exception types, resolution methods, and time taken to resolve, the system can identify the most effective solutions and apply them to similar cases. Continuous feedback from users ensures that the AI adapts and refines its decision-making process over time, ultimately reducing manual effort, improving accuracy, and enhancing processing efficiency. The AI should prioritize learning from frequent, high-impact exceptions by tracking resolution success and user actions, focusing on patterns that lead to effective outcomes.

I think employee performance evaluation is one process where AI can track the employee performance based on various parameters and assigned targets. However, it might not be able to take decision related to promotions or putting employee on Performance improvement plan or termination as performance might be impacted due to personal problems, family issues or any financial crisis which AI might not understand well as humans do. There could be a situation where a 2nd top performer is in need of promotion more than one who is a top performer, in that case human intervention is required. AI should take into considerations all the personal, medical, financial or mental aspects of an employee while performance evaluation and should pause and consult a human before taking decision. So I feel these parameters should be the criteria for AI to escalate.

What’s one rule where you would fully trust AI to make the call? Answer: Data backup and retention policy, where AI can make data-based decision for saving the files, archiving unused or old files and retaining data from backup as and when required. one rule where you would never trust it? Why? Answer: We would never trust AI to make Hiring and termination decisions as there are many impacting parameters that needs an human view to evaluate the situation and take decisions. It cannot be rule based though it can be done partially using AI like shortlisting resumes or documenting performance issues but not take decision if individual should be hired or terminated.

In your work environment, what’s one task or decision that feels “too human” to hand over to AI? Answer: Preparing a consolidated report that meets leadership reporting expectations, alongside delivering proper presentations, can be challenging when they constantly evolve with each role and needs to be fetched from different systems. As individuals attempt to align numbers and insights with their specific targets, the data often needs to be broken down to its lowest level. This allows each person to see their individual performance data. However, it’s equally important that the overall data remains useful for leadership at the top level, enabling them to make informed decisions. How might you reimagine it to make AI a valuable contributor? Answer: AI tool may integrate multiple systems to enable user level access and can summarize the data across all levels to provide leadership with actionable, high-level insights that are essential for informed decision-making. By using AI-driven analytics, we can continuously adjust and optimize reports based on real-time data, ensuring both granular details for individual contributors and strategic insights for executives.