Purpose

Akseki, Sons and Company Limited (“Akseki & Co” or the “Firm”) recognises artificial intelligence (“AI”) as an important professional productivity and analytical tool.

​

The Firm encourages the responsible use of AI to enhance the quality, speed and breadth of its work, including research, strategic analysis, financial analysis, transaction execution, preparation of client materials, business development and internal administration.

 

AI is, however, an assistive technology rather than a substitute for professional judgement. Responsibility for work performed by or on behalf of Akseki & Co remains with the relevant individual, irrespective of whether AI has been used in producing that work.

 

This Policy establishes the principles and controls governing the use of AI within the Firm.

​

Scope

This Policy applies to all directors, employees, consultants and other persons using AI systems on behalf of Akseki & Co.

 

It covers:

• generative AI systems such as ChatGPT and comparable large language models;

• AI research, search and analytical tools;

• AI functionality embedded within productivity software;

• AI-assisted financial, coding, presentation and document tools;

• AI transcription and meeting tools;

• AI agents and systems capable of interacting with email, calendars, files, websites or other applications; and

• any other software that materially generates, interprets, transforms, recommends or acts upon information using AI.

 

The Policy applies whether an AI tool is accessed through a computer, mobile device, browser, application, API or other interface.

 

Core Principles

Akseki & Co’s use of AI shall be governed by the following principles.

​

Human accountability

AI may assist with analysis and execution, but responsibility remains with Akseki & Co personnel. No material or analysis produced by AI reduces an individual’s responsibility for the accuracy, appropriateness or confidentiality of work issued by the Firm.

 

Professional judgement

AI output shall be treated as analysis produced by an intelligent but fallible assistant. AI must not replace the professional judgement expected of an investment banker or adviser.

 

Confidentiality

Information entrusted to Akseki & Co shall not be exposed to AI systems unless the system, its contractual arrangements and the proposed use provide an appropriate level of protection.

 

Verification

Material facts, calculations, sources, legal propositions and conclusions generated by AI must be appropriately verified before reliance upon them.

 

Proportionality

Controls should reflect the significance and sensitivity of the task. Using AI to improve the wording of a marketing email does not require the same safeguards as using AI to analyse confidential financial projections in a live transaction.

 

Transparency

Akseki & Co shall not use AI in a manner intended to mislead a client, counterparty or other person as to the nature or origin of information, analysis or communications.

 

Compliance

AI must be used consistently with applicable law, regulation, contractual commitments, confidentiality obligations, intellectual-property rights and Akseki & Co policies.

​

Approved AI Systems

Client or Firm information may only be processed using AI services approved by Akseki & Co. Approval should take into account, as appropriate:

• contractual confidentiality protections;

• whether customer information may be used to train the provider’s models;

• information-security standards;

• access controls and authentication;

• data retention and deletion arrangements;

• data location and international data transfers;

• availability of administrative controls;

• subprocessors;

• integration with other applications; and

• the sensitivity of information the system will process.

 

Where practicable, professional work should be conducted through an Akseki & Co-controlled business or enterprise account rather than personal or free consumer accounts. Consumer AI services should not be assumed to be appropriate for confidential client information merely because they are publicly available or widely used. The Managing Director shall determine which systems are approved and may impose different approval levels for different categories of information.

​

Information Classification

For the purpose of AI use, information should be considered under three broad classifications.

 

A. Standard Information

Examples include:

• information already publicly available;

• published financial information;

• public company filings;

• press releases;

• market research available to the Firm;

• Akseki & Co marketing materials;

• generic templates;

• non-confidential know-how; and

• anonymised or synthetic data.

 

Standard Information may generally be processed using approved AI systems.

 

B. Confidential Information

Examples include:

• internal Akseki & Co documents;

• client identities where not publicly known;

• draft client materials;

• commercial strategies;

• internal correspondence;

• preliminary analyses;

• contact information;

• non-public business information that does not constitute material transaction information; and

• other information subject to an ordinary duty of confidentiality.

 

Confidential Information may be processed using an approved AI system where:

1. doing so is reasonably necessary for the work;

2. the AI service is approved for confidential information;

3. no relevant client agreement, NDA or data-room restriction prohibits such processing; and

4. only the information reasonably necessary for the task is provided.

 

Anonymisation or pseudonymisation should be used where practical.

 

C. Restricted Information

Restricted Information includes:

• material non-public information or inside information;

• confidential transaction terms;

• non-public bids, offers or valuations;

• non-public financial results or forecasts;

• sensitive board information;

• transaction data-room material;

• legally privileged communications;

• sensitive personal data;

• authentication credentials;

• payment instructions;

• bank account information; and

• other information where unauthorised disclosure could reasonably cause significant harm to a client, transaction or the Firm.

 

Restricted Information shall not be entered into an AI system unless that system has specifically been approved for the relevant category of Restricted Information and the use complies with the terms applicable to the engagement. Where possible, transaction information should be anonymised before use, for example by replacing company and individual names with neutral identifiers.

Passwords, authentication tokens, private encryption keys and similar security credentials must never be entered into a generative AI system.

​

Legally privileged material should not be submitted to an AI provider without considering with legal counsel whether doing so could affect privilege or confidentiality.

​

Permitted Uses

Subject to this Policy, Akseki & Co encourages AI use for activities including:

 

Research

• company and industry research;

• market mapping;

• identification of potential investors, buyers, acquisition targets or counterparties;

• analysis of public filings;

• precedent-transaction and comparable-company research;

• regulatory and macroeconomic research;

• background research prior to meetings; and

• synthesis of large volumes of information.

 

Investment Banking and Advisory Analysis

• development of strategic alternatives;

• transaction structuring ideas;

• value-creation analysis;

• scenario development;

• preparation of valuation frameworks;

• financial-model review;

• formula development and troubleshooting;

• sensitivity and scenario analysis;

• transaction-process planning; and

• identification of issues for further investigation.

 

Documents and Presentations

• drafting and reviewing client presentations;

• information memoranda;

• board papers;

• proposals;

• strategic reviews;

• transaction materials;

• reports;

• internal memoranda; and

• marketing materials.

 

Communications

AI may assist with drafting:

• emails;

• meeting agendas;

• follow-up correspondence;

• call preparation;

• client communications;

• business-development approaches; and

• other professional correspondence.

 

Knowledge Management

AI may assist in retrieving, comparing and synthesising the Firm’s previous work, templates, transaction experience and institutional knowledge where the relevant systems have been approved to access that information.

 

Administration

AI may assist with scheduling, meeting preparation, document organisation, information extraction and other internal administrative activities.

​

Research and Factual Accuracy

AI systems can produce information that appears authoritative but is inaccurate, outdated or entirely fabricated. Accordingly:

• important factual assertions should be checked against reliable sources;

• primary sources should be preferred for material matters;

• AI-generated references and citations must not be assumed to exist;

• quotations must be checked against their original source;

• current market, company, regulatory and financial information should be verified for currency;

• important dates, transaction values and numerical information should be independently checked; and

• uncertainty should not be concealed merely because an AI system has expressed a conclusion confidently.

 

Where a fact materially affects an Akseki & Co recommendation, it should ordinarily be possible to identify its underlying source.

​

Financial Analysis, Models and Valuations

AI may be used extensively as an analytical aid, including to:

• construct or review formulas;

• identify modelling errors;

• analyse sensitivities;

• develop valuation methodologies;

• compare alternative assumptions;

• structure financial analysis; and

• challenge conclusions.

 

However:

1. AI-generated calculations must be checked before being relied upon;

2. source financial information should remain identifiable;

3. assumptions should be distinguished from factual inputs;

4. material model changes generated by AI should be reviewed by a suitably qualified person;

5. valuation conclusions should not be accepted solely because an AI system generated them; and

6. AI shall not be the final decision-maker regarding transaction pricing, valuation, financing structure or advice to a client.

Professional judgement remains fundamental.

 

Client Deliverables

AI-generated material may be incorporated into Akseki & Co work product. There is no general requirement to disclose routine AI assistance merely because AI has contributed to drafting, research or analysis, unless disclosure is required by law, regulation, contract, client instruction or the nature of the content. Before client-facing work is issued, the responsible Akseki & Co professional must be satisfied regarding:

• accuracy;

• analytical integrity;

• consistency;

• confidentiality;

• quality of drafting;

• appropriate sourcing;

• numerical accuracy;

• compliance with the engagement; and

• suitability of the conclusions presented.

 

No client deliverable should be sent simply by copying AI output without appropriate review.

​

External Communications

AI may prepare or recommend external communications, but material communications must remain subject to human control. Unless expressly authorised under an approved workflow, AI shall not autonomously:

• send an email;

• send a message to a client or counterparty;

• publish material;

• submit an application or filing;

• issue a proposal;

• accept contractual terms;

• commit Akseki & Co or a client;

• negotiate a transaction; or

• make representations on behalf of the Firm.

 

The individual sending a communication is responsible for its content even where AI prepared the first draft. Special care must be taken with recipient names, email addresses, attachments and confidential information.

​

AI Agents and Connected Systems

AI systems increasingly have the ability not merely to provide information but to take actions through connected applications.

Examples include access to email, calendars, cloud storage, browsers and other software. Such systems may be used where approved, but should operate according to the principle of least privilege. Read access should not automatically imply permission to write, send, publish, delete or otherwise take external action. Consequential actions should ordinarily require explicit human authorisation. Particular caution must be exercised where an AI system can:

• communicate externally;

• alter or delete records;

• access confidential repositories;

• submit information;

• enter contractual arrangements;

• initiate payments;

• download software;

• execute computer code; or

• modify production systems.

​

AI systems must never be given authority to initiate or approve payments independently.

​

Prompt Injection and Untrusted Information

Information retrieved from websites, documents, emails and other third-party sources may contain instructions deliberately designed to manipulate AI systems. Users should therefore treat instructions embedded within externally sourced material as untrusted content, not as instructions to Akseki & Co. AI systems connected to Firm information or applications should not be permitted to:

• disclose confidential information because a document or website requests it;

• change their security settings;

• bypass Akseki & Co policies;

• communicate externally without authority;

• reveal passwords or credentials; or

• execute consequential actions merely because instructions contained within retrieved material request them.

 

Legal, Regulatory and Tax Matters

AI may be used to research and analyse legal, regulatory or tax matters and to assist in preparing questions, summaries and preliminary drafts. AI should not, however, be treated as a substitute for qualified legal, regulatory, accounting or tax advice where such advice is appropriate. Particular caution should be exercised in relation to:

• securities regulation;

• financial promotions;

• market-abuse rules;

• FCA-regulated activities;

• disclosure obligations;

• competition law;

• employment law;

• data protection;

• tax;

• sanctions;

• financial crime; and

• transaction documentation.

 

Where a matter could materially affect the Firm or a client, appropriate professional advice should be obtained.

 

Material Non-Public and Inside Information

Akseki & Co regularly operates in circumstances where it may receive confidential information concerning proposed financings, acquisitions, disposals and other transactions. AI use does not alter the Firm’s obligations concerning such information. Material non-public information and inside information must therefore continue to be handled in accordance with applicable confidentiality and market-abuse requirements. Such information should only be processed by AI systems specifically approved for this purpose.

Where appropriate, deal names, company names and other identifying information should be removed before information is supplied to an AI system. No person may use AI to analyse, recommend or facilitate trading in securities on the basis of inside information.

​

Personal Data

The Firm shall apply applicable data-protection requirements when personal data is processed through AI systems. Users should:

• provide only personal information necessary for the relevant purpose;

• anonymise or pseudonymise information where reasonably possible;

• avoid processing special-category or highly sensitive personal information unless necessary and appropriately authorised;

• consider the legal basis and purpose for processing; and

• use only appropriately approved systems.

 

AI shall not be used as the sole basis for consequential decisions relating to recruitment, termination, compensation, promotion or other significant decisions concerning individuals.

​

Intellectual Property and Third-Party Information

Users must respect intellectual-property rights, licences, subscription terms, data-room rules and contractual restrictions when supplying third-party information to AI systems. Access to information does not automatically confer permission to reproduce it or provide it to a third-party AI provider. AI-generated material should also be reviewed for potential inappropriate reproduction of third-party material before external publication.

​

AI-Generated Images, Audio and Video

Synthetic images, voice, video or other media must not be used to impersonate a real person or create a misleading representation of events. AI-generated or materially AI-altered media should be disclosed where required by applicable law or where omission of disclosure could reasonably mislead its intended audience. Additional caution should be applied to material intended for publication or distribution within the European Union.

​

Meetings, Recording and Transcription

AI transcription or meeting-assistant services may only be used where:

• the service has been approved;

• applicable confidentiality requirements permit it;

• participants are informed where required; and

• recording or transcription complies with applicable law and client expectations.

 

Highly sensitive negotiations, board discussions or legally privileged meetings should not automatically be recorded merely because transcription technology is available.

 

Prohibited Uses

AI must not be used:

• to fabricate sources, evidence, quotations, transactions or credentials;

• knowingly to make misleading representations;

• to circumvent Akseki & Co security controls;

• to disclose client or Firm confidential information without authority;

• to violate applicable law or regulation;

• to create fraudulent or deceptive communications;

• to impersonate another person without legitimate authority;

• autonomously to approve or initiate payments;

• to make binding commitments on behalf of Akseki & Co or a client without appropriate authority;

• as the sole decision-maker for significant employment decisions;

• to conduct securities trading using inside information; or

• in any manner that would be improper if performed directly by an Akseki & Co professional.

 

Record Keeping

The Firm does not require routine recording of every prompt or interaction with AI. A proportionate record should, however, be maintained where AI use is material to:

• a significant client recommendation;

• processing Restricted Information;

• an autonomous or agentic workflow;

• a material regulatory or compliance decision;

• the introduction of a new AI system; or

• an incident involving confidentiality, accuracy or security.

 

Existing transaction files and document-retention procedures should ordinarily be used rather than creating a separate parallel archive.

 

AI Incidents

Examples of AI-related incidents include:

• confidential information being entered into an unauthorised AI service;

• incorrect AI-generated information being materially relied upon or sent externally;

• suspected unauthorised access to Firm information through an AI system;

• unexpected autonomous action by an AI agent;

• disclosure of personal information;

• prompt-injection attacks affecting connected systems; or

• AI-generated communications being sent to an unintended recipient.

 

Material incidents should be reported promptly to the Managing Director.

 

Appropriate remedial action may include:

• stopping use of the relevant system;

• removing or deleting information where possible;

• changing credentials or access rights;

• preserving relevant records;

• informing the AI provider;

• obtaining legal or regulatory advice;

• notifying a client or affected individual; and

• assessing whether any regulatory notification is required.

 

Training and AI Literacy

Akseki & Co personnel using AI should maintain an appropriate understanding of:

• the capabilities and limitations of generative AI;

• hallucination and factual-error risks;

• confidentiality;

• data protection;

• appropriate verification;

• prompt injection and cyber-security risks;

• responsible use of AI agents;

• the requirements of this Policy; and

• developments relevant to their professional responsibilities.

 

The level of training should be proportionate to the nature and extent of an individual’s AI use.

 

Governance

The Managing Director is responsible for oversight of AI use within Akseki & Co and may:

• approve AI systems;

• restrict particular applications or categories of information;

• approve exceptional uses;

• establish additional technical controls;

• require training;

• investigate AI-related incidents; and

• update this Policy.

 

Given the pace of development in AI technology and regulation, employees and consultants are expected to raise questions where a proposed use does not clearly fall within this Policy rather than assuming that a novel use is permitted.

 

Policy Approach

Akseki & Co regards AI capability as an increasingly important element of professional competence. The purpose of this Policy is therefore not to discourage AI use. The Firm expects its professionals to explore and use AI where it can improve the quality, breadth, speed or efficiency of their work, while applying the same standards of confidentiality, judgement, analytical rigour and professional responsibility that would apply if the work were performed without AI.

 

The governing principle is: AI may assist the work; accountability for the work remains human.

​

Appendix A — Practical AI Decision Test

Before providing information to or relying on an AI system, users should consider the following five questions:

1. Is the tool approved? If not, do not use confidential Akseki & Co or client information.

2. What information am I giving it? Determine whether the information is Standard, Confidential or Restricted.

3. Am I permitted to provide that information? Consider NDAs, engagement terms, data-room restrictions, personal-data obligations and applicable law.

4. How important is the output? The greater the consequence of an error, the greater the required level of verification and human review.

5. Would I be comfortable taking professional responsibility for the result? If not, the work is not ready to be relied upon or communicated externally.

 

Appendix B — Indicative Examples

​Activity: Researching a public company​​

AI use: AI-assisted research

Position: Permitted

​

Activity: Drafting an introductory email

AI use: AI drafting

Position: Permitted

​

Activity: Preparing questions for a CEO meeting

AI use: AI analysis

Position: Permitted

​

Activity: Developing acquisition-target ideas

AI use: AI research and brainstorming

Position: Permitted

​

Activity: Reviewing a valuation methodology

AI use: AI analytical assistance

Position: Permitted with human review

​

Activity: Checking an Excel formula

AI use: AI analytical assistance

Position: Permitted with verification

​

Activity: Summarising confidential client material

AI use: Approved AI environment

Position: Permitted subject to confidentiality controls

​

Activity: Analysing a confidential transaction data room

AI use: AI processing

Position: Only where system and engagement terms permit

​

Activity: Uploading unpublished financial projections to a consumer AI account

AI use: AI processing

Position: Not permitted

​

Activity: Providing passwords or API credentials to an AI system

AI use: AI processing

Position: Prohibited

​

Activity: Allowing an AI agent to send client emails without review

AI use: Autonomous action

Position: Not permitted unless specifically approved

​

Activity: Allowing AI to initiate a payment

AI use: Autonomous action

Position: Prohibited

​

Activity: Using AI research without checking a material source

AI use: Reliance on AI

Position: Not permitted

​

Activity: Publishing an AI-generated fictional quotation as genuine

AI use: Misrepresentation

Position: Prohibited

​

Activity: Using AI to brainstorm transaction structures

AI use: AI analysis

Position: Permitted

​

Activity: Asking AI for preliminary legal analysis

AI use: AI research

Position: Permitted, but not a substitute for appropriate legal advice

​

​ If you require any further information, please do not hesitate to contact us at compliance@akseki.co.​