AUTOMATED CUSTOMER INTERACTION SCORING AND UNFAIR DISMISSAL RISKS

1. INTRODUCTION

Automated customer interaction scoring refers to the use of artificial intelligence, speech analytics, customer relationship management software and algorithmic assessment tools to evaluate employee performance during customer interactions. These systems are commonly used in call centres, banking institutions, insurance companies, retail organisations and customer support departments.

Automated systems may assess employees according to speaking speed, response time, customer satisfaction ratings, conversation duration, politeness, emotional tone, problem-resolution rates and compliance with prescribed communication scripts. Employers may subsequently use these scores to determine performance ratings, bonuses, promotions, disciplinary action or termination.

However, automated customer interaction scoring creates significant employment-law concerns when inaccurate algorithms, discriminatory evaluation standards or unreliable customer feedback influence dismissal decisions. Employees may face termination because of accents, language differences, disability-related communication patterns, customer prejudice or technical failures beyond their control.

In India, dismissal based primarily on automated performance scores must be examined against applicable employment legislation, contractual obligations, natural justice and constitutional protections.

2. LEGAL FRAMEWORK GOVERNING AUTOMATED PERFORMANCE SCORING

A. INDUSTRIAL RELATIONS CODE, 2020

The Industrial Relations Code, 2020, establishes provisions concerning industrial disputes, standing orders, disciplinary matters and employment termination.

Where an employee qualifies as a worker under the applicable statutory framework, dismissal based on automated customer interaction scores may raise questions concerning misconduct, disciplinary procedure, retrenchment and available remedies.

Employers must comply with applicable procedural requirements rather than treating an algorithmic score as automatic proof of misconduct.

B. CODE ON WAGES, 2019

The Code on Wages, 2019, provides protections concerning wage payments and prohibits specified forms of discrimination in wages.

An employer cannot automatically withhold earned remuneration or impose unauthorised financial penalties merely because an employee receives an unsatisfactory customer interaction score.

C. RIGHTS OF PERSONS WITH DISABILITIES ACT, 2016

The Act provides protections against disability discrimination in employment.

Automated communication assessments should not unfairly disadvantage employees with speech impairments, hearing-related disabilities or other conditions affecting communication.

D. DIGITAL PERSONAL DATA PROTECTION ACT, 2023

Customer interaction scoring may involve processing employee voice recordings, transcripts, performance records and other identifiable personal information.

Employers must consider applicable data-processing, notice and security obligations under the Act, subject to its commencement and applicable rules.

E. CONSTITUTIONAL PROTECTIONS

Articles 14, 16 and 21 provide important protections against arbitrariness, discrimination and violations of dignity and privacy, particularly in public employment and other circumstances where constitutional protections apply.

3. MAJOR LEGAL RISKS ASSOCIATED WITH AUTOMATED CUSTOMER SCORING

A. ALGORITHMIC DISCRIMINATION

Automated systems may disadvantage employees because of regional accents, pronunciation, gender, age, disability or cultural communication differences.

For example, an employee speaking with a regional Indian accent may receive a lower communication score despite providing accurate customer assistance.

B. UNRELIABLE CUSTOMER SATISFACTION RATINGS

Customer ratings may reflect personal prejudice, unreasonable expectations or dissatisfaction with organisational policies rather than actual employee performance.

Dismissal based exclusively on such ratings may therefore be unfair.

C. INACCURATE SPEECH RECOGNITION

Speech-recognition software may incorrectly interpret words, misunderstand regional languages or fail to identify the context of a conversation.

Such errors can produce misleading assessments of employee behaviour.

D. DISPROPORTIONATE DISMISSAL

An isolated low customer interaction score should not automatically establish serious misconduct or justify termination without examining the surrounding circumstances.

E. LACK OF TRANSPARENCY

Employees may not understand how the algorithm calculates performance scores or which factors contribute to negative assessments.

F. ABSENCE OF HUMAN REVIEW

An organisation that dismisses employees solely through automated scoring risks ignoring relevant circumstances, technical errors and individual explanations.

4. SIX IMPORTANT JUDICIAL DECISIONS

The following decisions establish relevant employment-law principles. They are primarily direct or analogous authorities concerning dismissal, disciplinary fairness, discrimination and evidentiary reliability rather than judgments specifically concerning AI-based customer scoring.

1. D.K. YADAV V. J.M.A. INDUSTRIES LTD. (1993)

Authority: Direct employment-law authority concerning natural justice.

The Supreme Court emphasised that termination affecting employment rights must satisfy principles of fairness and natural justice.

Relevance: An employer should not terminate an employee merely because an automated customer scoring platform generates a negative rating. The employee must receive an appropriate opportunity to understand and challenge the allegations.

2. MANEKA GANDHI V. UNION OF INDIA (1978)

Authority: Direct constitutional authority on fairness and non-arbitrariness.

The Supreme Court established that procedures affecting individual rights must be fair, reasonable and non-arbitrary.

Relevance: Public employers must ensure that automated performance assessment does not become an arbitrary mechanism for terminating employees. Transparent criteria and fair review procedures are essential.

3. CENTRAL INLAND WATER TRANSPORT CORPORATION LTD. V. BROJO NATH GANGULY (1986)

Authority: Direct authority concerning unfair employment conditions.

The Supreme Court examined oppressive contractual employment provisions and recognised that unconscionable terms may be invalid in appropriate circumstances.

Relevance: Employers should not rely on one-sided contractual provisions that permit dismissal solely on the basis of undisclosed or unexplained automated performance scores.

4. GLAXO LABORATORIES (I) LTD. V. PRESIDING OFFICER, LABOUR COURT, MEERUT (1984)

Authority: Direct employment-law authority concerning disciplinary misconduct and service rules.

The Supreme Court considered the relationship between employee misconduct and the applicable standing orders governing disciplinary action.

Relevance: An employer must establish that the employee's alleged failure satisfies the applicable definition of misconduct or another legally valid ground for disciplinary action. A low algorithmic score is not automatically proof of misconduct.

5. ROOP SINGH NEGI V. PUNJAB NATIONAL BANK (2009)

Authority: Direct employment-law authority concerning evidentiary reliability.

The Supreme Court emphasised that disciplinary findings must be based on relevant evidence and that suspicion cannot replace proof.

Relevance: Automated customer interaction reports should be supported by reliable underlying records, including actual conversations, relevant transcripts and contextual evidence. Unsupported algorithmic conclusions should not be treated as conclusive proof of employee misconduct.

6. VIKASH KUMAR V. UNION PUBLIC SERVICE COMMISSION (2021)

Authority: Direct authority on reasonable accommodation and disability equality.

The Supreme Court recognised reasonable accommodation as an essential component of substantive equality under disability law.

Relevance: Customer interaction scoring must accommodate employees whose disabilities affect speech, communication speed or other assessed characteristics. Employers should not penalise disability-related communication differences without considering legally required adjustments.

5. EMPLOYER RESPONSIBILITIES AND COMPLIANCE REQUIREMENTS

Employers implementing automated customer interaction scoring should adopt the following safeguards:

TRANSPARENT EVALUATION CRITERIA: Clearly explain the scoring methodology, performance indicators and consequences of poor ratings.

ALGORITHM AUDITING: Periodically examine automated systems for linguistic, gender, disability and other discriminatory biases.

HUMAN SUPERVISION: Require meaningful human assessment before imposing serious disciplinary penalties or termination.

EVIDENCE VERIFICATION: Preserve original customer interaction recordings, transcripts and relevant performance records.

CONTEXTUAL ASSESSMENT: Consider customer behaviour, system failures, organisational policies and external factors affecting performance.

REASONABLE ACCOMMODATION: Provide appropriate adjustments for employees with disabilities.

EMPLOYEE APPEAL RIGHTS: Establish an internal procedure for challenging inaccurate scores and unfair evaluations.

PROPORTIONALITY: Distinguish between minor performance deficiencies, repeated underperformance and genuine misconduct.

DATA PROTECTION: Maintain appropriate controls over employee recordings and performance information.

DOCUMENTED DISCIPLINARY PROCEDURES: Record the reasons for adverse employment decisions and the evidence supporting them.

6. EMPLOYEE RIGHTS AND AVAILABLE LEGAL REMEDIES

Employees who believe that automated customer scoring has resulted in unfair dismissal may:

Request an explanation of the performance assessment.

Challenge inaccurate transcripts and customer interaction records.

Seek access to relevant evidence in accordance with applicable employment rules.

Raise an internal grievance against discriminatory scoring.

Challenge disability-related discrimination under the Rights of Persons with Disabilities Act, 2016.

Contest dismissal through the appropriate labour authority or industrial tribunal, where legally applicable.

Challenge arbitrary public-employment decisions through appropriate constitutional proceedings.

Seek reinstatement, back wages or other appropriate relief where legally justified.

The available remedies depend on the employee's legal classification, applicable service rules, employment contract and statutory framework.

7. PRACTICAL COMPLIANCE AND RISK MANAGEMENT

Organisations should establish a formal policy governing the use of AI-based customer interaction assessment systems.

First, performance indicators should measure genuine service quality rather than relying exclusively on speaking speed, conversation length or emotional tone.

Second, customer feedback should be evaluated alongside objective indicators such as issue resolution, accuracy, customer complaints, procedural compliance and service outcomes.

Third, employers should conduct regular audits to identify whether particular linguistic or demographic groups consistently receive lower scores without objective justification.

Fourth, organisations should maintain a review mechanism allowing employees to challenge incorrect assessments before disciplinary consequences are imposed.

Finally, termination decisions should be based on verified evidence, applicable service rules and fair disciplinary procedures. Automated scoring should support managerial evaluation rather than replace independent judgment.

8. CONCLUSION

Automated customer interaction scoring can improve workplace performance management, but excessive dependence on algorithmic assessments creates substantial risks of unfair dismissal, discrimination, privacy violations and procedural injustice.

The decisions in D.K. Yadav v. J.M.A. Industries Ltd., Maneka Gandhi v. Union of India, Central Inland Water Transport Corporation v. Brojo Nath Ganguly, Glaxo Laboratories v. Presiding Officer, Roop Singh Negi v. Punjab National Bank and Vikash Kumar v. Union Public Service Commission provide important judicial foundations for evaluating the legality and fairness of employment decisions based on automated assessments.

Ultimately, an employee should not lose employment merely because an algorithm assigns an unfavourable customer interaction score. Employers must establish reliable evidence, apply objective performance standards, provide reasonable accommodation and follow fair disciplinary procedures. Human oversight, transparency and accountability are essential to ensuring that automated workplace evaluation remains consistent with Indian employment law and fundamental principles of justice.