Traditional AI
Explainable AI
“This engineer has a 70% chance of resigning”
No reason given.
“This employee has a 70% chance of quitting BECAUSE of low survey scores and no promotions in 3 years.”
An MSME-focused Non-Banking Financial Company (NBFC) was facing a serious employee turnover problem. Most workers stayed only 8 to 12 months. Because the firm handles confidential customer data and private personal details, this constant hiring and departing created major data privacy and security risks during financial audits.
To fix this, the company needed a way to predict which employees were likely to leave and act before they resigned.
Traditional AI models often act like a “black box”—they tell you what will happen (like giving a score saying an employee has a 70% chance of leaving), but they don’t tell you why.
Explainable AI (XAI) opens that black box. Instead of just giving a score, it uses simple language to explain the specific reasons behind the prediction—such as low job satisfaction, burnout, or lack of promotions.
“This engineer has a 70% chance of resigning”
No reason given.
“This employee has a 70% chance of quitting BECAUSE of low survey scores and no promotions in 3 years.”
Data Collected: HR combined employee surveys, attendance records, and exit interviews.
The Prediction: The AI flagged a key group of engineers at high risk of quitting.
The Explanation: XAI showed the exact causes: poor career growth, recent spikes in time off, and low ratings for manager support.
The Action: HR didn’t have to guess. They immediately set up mentoring programs, clearer career tracks, and leadership training for managers.
Clear Explanations: HR sees the exact reasons behind a high-risk score instead of relying on guesswork.
Targeted Action: Teams can fix specific problems—like offering flexible hours or adjusting pay—instead of applying generic solutions.
Builds Manager Trust: Managers are far more likely to act on AI advice when they can see the logic behind it.
Spotting Unfair Bias: XAI reveals how the model makes decisions, ensuring it doesn’t unfairly flag employees based on age, gender, or department.
Testing “What-If” Scenarios: HR can run simulations to see how much turnover drops if they improve specific areas, like training or recognition programs.
Step | Action | Description |
|---|---|---|
1. Gather Data | Collect Information | Combine HR records, attendance, survey results, and exit interviews into one secure system. |
2. Build the Model | Train AI | Use machine learning algorithms alongside XAI tools to track what drives turnover. |
3. Show the Logic | Share Insights | Display clear reasons for risk scores (e.g., “70% risk driven by lack of promotion”) and check for bias. |
4. Take Action | Fix the Cause | Roll out targeted solutions (like mentoring or perks) and simulate how changes will lower turnover. |
By moving from basic AI to Explainable AI, companies shift from reacting to resignations after they happen to preventing them early. XAI turns complex data into clear, human-centred actions that protect business data and keep top talent.
Atanu has demonstrated exceptional analytical capability by identifying the root causes of attrition within the team and recommending effective remedial measures. Leveraging his AI driven model, he has enabled the organisation to predict potential attrition and proactively implement strategies to mitigate risks before they materialise. His insights have established a forward-looking attrition management framework, strengthening workforce stability and organisational resilience.
An investment consultant firm (MSME) with 10+ employees, looking forward to become compliance on POSH regulations. 2025-26 the organisation had few employees and POSH compliance was not taken seriously by its board.
Detailed work on making the firm POSH compliance was taken up and work was delivered within 2 weeks.
Within just two weeks, a comprehensive POSH compliance framework was successfully established for the organization.
Policy Drafting & Approval: A detailed POSH policy was drafted, reviewed by senior leadership, and formally approved by the Board of Directors.
Defining Sexual Harassment & Consequences: Sexual harassment was explicitly defined as a violation of the Code of Conduct. Consequences were incorporated into the policy, and appointment letters were updated to reflect these changes.
Formation of Internal Committee (IC): A 4member IC was constituted, headed by a senior female leader. An external member was appointed in line with POSH regulations.
Standard Operating Procedure (SOP): A detailed SOP was designed, outlining inquiry procedures, daywise activities, and actionable steps for the IC team.
Compliance Checklist: A structured compliance checklist was created, with clear timelines and designated responsible/accountable persons for each task.
Awareness & Training Program: A customized awareness program was rolled out, including posters, email campaigns, and training modules tailored to the organization’s culture. All notices were updated and displayed within the office premises at visible places for easy access to all employees.
Audit Scope and Checklist: An audit scope document and a checklist was created to review the effectiveness, of the POSH compliance implementation. Based on the Adit scope an audit plan was created and various controls systems were enhanced for continuous monitoring and identify gaps in the process.
Atanu possesses comprehensive knowledge and hands on experience in POSH compliance, which he effectively applied to create and implement a robust compliance framework for the organisation. His ability to translate regulatory requirements into practical policies and actionable processes reflects both depth of expertise and strong execution skills. By drafting the POSH policy, establishing the Internal Committee, designing SOPs, and rolling out awareness programs, Atanu ensured that the organisation not only met statutory obligations but also fostered a culture of integrity and respect. His proactive approach and structured planning have positioned the firm as fully POSH compliant within a remarkably short timeframe.
An upcoming NBFC and startup firm, have established their presence in Indian Financial Industry map with its innovative products and AI driven technologies. The firm had received global funding opportunities, however, due to lack of proper governance mechanism, the funds were not credited to the organisation. The firm was looking forward to setting up of Governance mechanism for its employees.
Detailed work on setting up of Governance mechanism was taken up and work was delivered within 3 weeks.
Existing Governance Mechanism: The firm did not have any Governance mechanism. Which had resulted in an internal fraud, with an exposure of close to INR 1 crore. As a result of which the investors were looking forward to a governance mechanism with in the firm.
Weak documentation: The firm followed a process which is mostly verbal and people dependent. There was no documentation, process note, governance policy and letters were issued to the employees at whims and fancies of the Management.
Creation of Policies: Two major policies were created:
1) Corporate Governance Policy – which defined the various types of letters to be issues by the organisation and its impacts based on the philosophy and principal of Natural Justice and prevailing laws.
2) Consequence Matrix – which defined various malpractices and various letters to be issued to an employee based on the malpractices defined in the policy document.
Example: The action of an employee that will result in breach of Code of Conduct from Organisation’s point of view and the disciplinary action that the organisation will initiate as a result of the same is defined in the Consequence Matrix.
Governance policy defines; there should be an investigation conducted by the relevant function within the organisation to establish the violation in Code of Conduct and the concerned employee should be issued a show cause notice to provide him / her an opportunity to explain his / her side of the story.
Formation of a Management Committee: amongst the senior leadership team for review of the disciplinary actions taken. The roles and responsibilities of the Management Committee were defined. The committee to function not only to review any disciplinary letter issued by the organisation, but also to act an appellate body to reconsider any disciplinary action as an exception. The committee to meet every quarter and decide and approve the efficacy of the consequence matrix in building the right culture within the organisation.
Awareness Campaign: all employees were made aware of the newly implemented policies 1) Corporate Governance policy and the consequence matrix. The employees were clearly communicated on various Do’s and Don’t’s and adherences under these policies through various email campaign, notices, training programs. Changes were made in the appointment letters and termination as a result of a malpractice was defined. these were done with a view to create a deterrent and build a right culture within the organisation.
Manual process Vs Automation: The policy designed created a robust environment and deterrent within the organisation. However, the process and implementation continued to remain manual with human intervention, thus increasing the possibility of errors and omission due to advent of human intervention. A detailed automation plan of the process was submitted and waiting to begin soon.
Letter template; Audit Scope and Checklist: with a view to minimise the risk of manual error various letter template were created to have uniformity and consistency in disciplinary letter issuance process. An Audit scope document and an Audit checklist was also created in order to have uniformity in the disciplinary action decision and following every process steps.
Atanu has successfully designed a detailed corporate governance implementation process, aligning it precisely with business requirements. His structured and actionable approach ensured seamless delivery, and we look forward to the next phase of automating the process under his guidance.