Pratish Deshpande

Pick metrics that reflect the customer experience and process goals (e.g., total wait time, not just time at reception). Use sampling, data validation rules, and cross-checking with other sources to catch incorrect data early.

Business Excellence leverages Voice of Employee (VoE) to identify process pain points, improve engagement, and drive innovation. It ensures that employee insights shape continuous improvement initiatives. Challenges include lack of trust, poor feedback mechanisms, and data not being translated into action. Strategies: Use anonymous surveys, regular open forums, and transparent follow-ups to build trust and show responsiveness. Example: In a shared services center, VoE feedback highlighted frustration over repetitive manual approvals. This led to a successful automation project that reduced approval time by 60%.

Yes, KRIs can be used to manage a process by identifying potential risks early and enabling proactive mitigation. However, companies often prioritize KPIs because they focus on performance and results, which are more directly linked to business goals. Benefits of KRIs: Early warning of potential issues (e.g., rising defect rates in manufacturing) Supports risk-based decision-making Enhances compliance and resilience Limitations of KRIs: Harder to quantify and interpret than KPIs May lead to false alarms or overlooked risks if poorly defined Example: A bank may use a KRI like the number of failed login attempts to detect potential fraud risk, complementing KPIs like customer satisfaction.

Organizations should use BPR when radical, transformative change is needed, often starting from scratch, and Lean Six Sigma when aiming for continuous, data-driven improvement of existing processes. The two approaches can complement each other—BPR can redesign core processes, while Lean Six Sigma can refine and sustain them. They are not fundamentally incompatible.

An automated weekly performance dashboard in my previous organization appeared efficient because it saved time and looked professional. However, it was ineffective as it focused on vanity metrics, lacked context, and often misled decision-makers. It gave a false sense of control without driving real improvements, failing to align with strategic goals or support informed actions.

In financial services, AI hallucinations, such as incorrect investment advice or misinterpretation of regulations, can cause confusion, loss of client trust, or regulatory issues. To reduce this risk, I would use constrained prompts, implement validation checks against trusted data, and include a human review layer before sharing outputs with clients. Adding disclaimers or confidence indicators can also help users identify and question uncertain responses, ensuring greater reliability and accountability.

The AI could help prevent compliance risks like misrepresentation of financial products, disclosure of confidential information, or making unauthorized commitments. I envision the AI offering feedback through subtle, real-time suggestions—highlighting risky phrases with tooltips or offering compliant alternatives—ensuring minimal disruption while enhancing accuracy and adherence to guidelines.

There are four main ways to build AI solutions, each with its pros and cons. 1. Conventional AI (rule-based, ML): Good for clear, structured problems like fraud detection. Easy to explain but limited with unstructured data or complex tasks. 2. Fine-tuning an LLM: You take a big pre-trained model and adapt it to your specific domain. Gives great results for focused tasks but needs decent data, skills, and compute. 3. Training from scratch: Full control, but very expensive and complex. Only makes sense for cutting-edge research or unique problems nobody else is solving. 4. Prompt/flow engineering: Fastest and cheapest. No retraining—just smart prompts. Great for quick tasks like summarizing or drafting, but not super reliable or customizable. In short: Use conventional AI for structured, classic tasks. Fine-tune if you need domain-specific smarts. Train from scratch only if you must. Prompt engineering for speed and ease without heavy lifting.

One ethical dilemma an AI in a BPO might face is when a client demands a service refund, but granting it could negatively impact an employee’s performance metrics. The AI could lean towards denying the refund to protect the employee, even if it’s not the fairest option for the customer. My approach would be to have the AI prioritize honesty and fairness for the client while considering the employee’s well-being. It should always favor long-term trust over short-term metrics.

Suppose AI wrongly flags a normal transaction as fraud and locks the customer’s account. The customer’s upset. The truth is, the AI didn’t mess up on its own, people built it, trained it, and let it run without enough checks. So the blame falls on the designers, the testers, and the company that rolled it out. You can’t just blame the technology. To avoid this, there should be humans reviewing key decisions, clear records of what the AI does, and regular system checks.

When AI goals clashes, it should admit 🙂 . No pretending to know what's “right.” Just flag the message and hand it off to a human, since most of the time, the clash is human fault. If we feed conflicting priorities then AI will clash. So modify instruction after getting notified for clash.