Part B

Responsible Artificial Intelligence (AI) Policy

This Part B is our deeper AI governance framework, serving both as public disclosure and internal accountability structure. It is not a marketing document.

1. Introduction

This Responsible Artificial Intelligence (AI) Policy (“Part B”) sets out PROCEKA TechSol Private Limited’s governance framework for the use of AI, both in client-facing contexts and internal operations. Unlike Part A above, which explains our AI approach to website visitors and clients, this Part B is also intended to serve as our internal accountability framework for AI-assisted business operations. It is not a marketing document.

2. Purpose

This Part B establishes the governance principles, oversight mechanisms, and accountability structures that apply whenever PROCEKA uses or considers adopting AI, in order to build justified trust with clients, professionals, regulators, and business partners.

3. Scope

This Part B applies to all current and future use of AI at PROCEKA, including AI used in client engagements, internal operations, the Website, and any future SaaS, marketplace, portal, or mobile application.

4. Definitions

Terms used in this Part B have the meanings given in Part A, Section 4, unless otherwise specified. “AI Governance” means the policies, processes, and oversight PROCEKA applies to the adoption and use of AI. “Human-in-the-loop” means a process design where a human professional reviews AI output before it takes effect.

5. Vision for Responsible AI

PROCEKA’s vision is for AI to serve as a decision-support technology that improves the speed and consistency of routine work, while professional judgement, accountability, and client relationships remain firmly in human hands. We do not aim to automate professional judgement itself.

6. AI Governance Principles

Our AI governance is built on the principles of human-centred design, human oversight, transparency, explainability, accountability, fairness, non-discrimination, privacy by design, security by design, confidentiality, accuracy, continuous improvement, responsible innovation, regulatory compliance, and proactive risk management, each addressed further in this Part B.

7. Ethical Commitments

PROCEKA commits to using AI responsibly, protecting confidential information processed through AI tools, respecting intellectual property rights, complying with applicable law, maintaining transparency about our AI practices, continuously improving our AI governance, reviewing emerging risks, and ensuring human support remains available to clients and users.

8. Human Oversight

Human-in-the-loop review is a standing requirement for any AI-assisted output that could affect a client’s business, financial, tax, legal, or regulatory position, before that output is relied upon or communicated.

9. Human Decision Making

Decisions with material consequences for a client — including professional advice, filings, and recommendations — are made by appropriately authorised human professionals at PROCEKA, informed by AI-assisted analysis where used, not determined by AI systems independently.

10. Human Review Requirements

Human review of AI-assisted output must, at minimum, involve a qualified professional assessing the output for accuracy, completeness, and applicability to the specific client situation before it is finalised or communicated.

11. Transparency

We aim to be transparent with clients about when AI has been used to support a piece of work, upon reasonable request, and we do not misrepresent AI-assisted output as purely human-generated work or vice versa.

12. Explainability

Where AI tools support a professional conclusion, the reviewing professional aims to be able to explain the reasoning behind that conclusion in terms independent of the AI tool itself, so that clients receive an explanation grounded in professional reasoning, not merely a restated AI output.

13. Accountability

PROCEKA, through its qualified professionals, remains fully accountable for advice and deliverables provided to clients, regardless of whether AI tools were used to support that work. Use of AI does not diminish or transfer professional accountability.

14. Fairness

We aim to apply AI tools consistently across similarly situated clients and matters, avoiding selective or inconsistent use that could produce unfair outcomes between clients.

15. Bias Prevention

Before adopting an AI tool for use in client-facing work, we consider, where reasonably practicable, whether the tool carries known risks of biased or skewed output, and apply human review specifically to identify and correct such issues where they arise.

16. Privacy Protection

AI-related data processing is subject to our Privacy Policy in full. We do not use AI tools to process personal information beyond what is necessary for the stated purpose, and we apply data minimisation where practicable when submitting information to AI tools.

17. Confidentiality

Confidential client information is handled with particular care when AI tools are involved. Where reasonably practicable, we avoid submitting highly sensitive client information to general-purpose AI tools that may retain or use such information beyond the immediate task, and we evaluate AI providers’ confidentiality commitments before adoption for client-related work.

18. Cyber Security

AI tools and integrations are assessed for security risk before adoption, consistent with our general security practices described in our Privacy Policy, recognising that AI systems can introduce novel risks such as prompt injection or data leakage, addressed further in Section 34.

19. Data Quality

Where AI tools rely on data we provide (such as internal knowledge base content), we take reasonable steps to keep that underlying data accurate and current, recognising that AI output quality depends significantly on input data quality.

20. Responsible Data Usage

We use data in connection with AI tools only for purposes consistent with our Privacy Policy and the specific purpose for which the data was originally collected, and do not repurpose client data for unrelated AI training or development without appropriate basis.

21. AI Training Data

PROCEKA does not currently train its own proprietary AI models. Where we use third-party AI tools, training data practices are governed by that provider’s own policies, which we consider when evaluating a provider for adoption.

22. AI Output Verification

AI-generated output intended for client-facing use is verified by a qualified professional against source material, applicable law, and the specific facts of the matter, rather than accepted at face value.

23. AI Limitations

We recognise that AI systems can produce inaccurate or fabricated output (“hallucinations”), may reflect biases present in training data, may not reflect the most recent regulatory changes, and lack genuine understanding or accountability. These limitations inform every governance mechanism in this Part B.

24. Professional Review

Professional review is mandatory, not optional, for AI-assisted output before client-facing use, regardless of how confident or well-formatted the AI output appears.

25. AI-Assisted Drafting

Where used, AI-assisted drafting produces a starting point for a document, which a qualified professional then reviews, corrects, and finalises, applying the same standard of care as if drafting from scratch.

26. AI-Assisted Research

Where used, AI-assisted research accelerates the identification of relevant sources and information, which are then independently verified by a qualified professional before being relied upon.

27. AI-Assisted Compliance

Where AI-assisted compliance tracking or flagging is introduced in future, it is intended to help identify potential deadlines or requirements for professional attention, not to make final compliance determinations.

28. AI-Assisted Business Analysis

Where AI-assisted business or financial analysis is used, resulting insights are reviewed by a qualified professional for relevance and accuracy before being presented to a client as a basis for decision-making.

29. AI-Assisted Knowledge Management

Where AI is used to organise or surface PROCEKA’s own knowledge resources internally or for clients, it supports faster access to information and does not itself constitute new professional advice.

30. AI-Assisted Customer Support

If PROCEKA introduces AI-assisted customer support features in future (such as a chat assistant), users will be informed they are interacting with an AI system, and a pathway to human support will remain available for matters requiring professional judgement.

31. AI-Assisted Internal Operations

PROCEKA may use AI to support internal operations, such as document organisation, internal research, or workflow support. Internal use is subject to the same confidentiality, security, and review principles as client-facing use, proportionate to the risk involved.

32. AI-Assisted Future SaaS

If PROCEKA introduces AI-enabled SaaS products in future, they will be governed by an updated version of this Part B specific to that product, published before general availability, in addition to applicable product-specific terms.

33. AI-Assisted Marketplace Services

If PROCEKA introduces a marketplace connecting clients and professionals in future, any AI-assisted matching or recommendation features will be subject to the fairness and non-discrimination principles in Sections 14 and 15, with detail provided before that feature launches.

34. AI Risk Management

Before adopting an AI tool, we consider risks including inaccurate output, bias, cybersecurity exposure, privacy impact, prompt injection (attempts to manipulate an AI system through crafted input), data leakage, and the tool’s track record and provider reputation, and weigh these against the intended benefit.

35. AI Incident Reporting

Suspected AI-related incidents, including a materially inaccurate output that was communicated to a client, a security concern, or misuse, should be reported internally and, where relevant, to the client, promptly upon discovery, and addressed as described in Section 32 of Part A.

36. AI Monitoring

We periodically review how AI tools are actually being used within the organisation, to confirm they remain within the scope and safeguards described in this Policy, and to identify any need for additional training or controls.

37. AI Auditing

Where AI tools are used in a manner that could materially affect clients, we periodically assess whether the tool continues to perform as expected and whether human review processes are being followed consistently.

38. Third-Party AI Providers

Before adopting a third-party AI provider for use in client-related work, we consider the provider’s data handling practices, security posture, and terms of service, in addition to functional suitability, and prefer providers with clear, reviewable data practices.

39. Intellectual Property

Ownership of AI-assisted work product prepared for clients is governed by our Service Engagement Terms & Conditions in the same manner as any other Deliverable, regardless of whether AI tools contributed to its preparation.

40. Client Responsibilities

Clients should not treat AI-assisted preliminary output shared with them (where explicitly labelled as such) as final professional advice, and should raise questions with their PROCEKA professional before acting on any output they are uncertain about.

41. Employee Responsibilities

PROCEKA personnel using AI tools in the course of their work are expected to follow this Policy, apply professional review to AI-assisted output before it reaches a client, protect confidential information when using AI tools, and raise concerns about AI tool suitability or performance through internal channels.

42. Responsible Use by Professionals

Professionals at PROCEKA remain bound by their applicable professional and ethical obligations when using AI tools, and AI assistance does not lower the standard of care expected in their professional work.

43. Prohibited Uses of AI

PROCEKA does not intentionally use AI to make final legal decisions, final tax determinations, final accounting opinions, or regulatory decisions; to impersonate humans without disclosure; to deceive or manipulate users; to discriminate unlawfully; to create fraudulent documents; to generate misleading professional advice; or to bypass required human review.

44. Continuous Improvement

We treat AI governance as an ongoing process, not a one-time policy exercise, and expect this Part B to evolve as our use of AI, applicable regulation, and industry best practice develop.

45. Policy Updates

We will update this Policy as our AI practices evolve, and will provide specific additional detail before any currently “planned” AI feature (as described on our Technology & AI Solutions page) becomes generally available.

46. Governing Law

This Policy is governed by the laws of India, including the Information Technology Act, 2000 and the Digital Personal Data Protection Act, 2023, without regard to conflict of law principles. We have drafted this Policy with reference to internationally recognised responsible AI principles, without claiming compliance with any specific international AI regulatory framework not currently applicable to PROCEKA.

47. Contact Information

PROCEKA TechSol Private Limited
Rajabandha, Ghatgaon, Keonjhar, Odisha, India
Email: info@proceka.com
Phone: +91 91785 66135

48. Effective Date

This AI Usage & Responsible AI Policy (Parts A and B) is effective as of 24 July 2026.