Ethics, Governance and the Future of Customer Acquisition

Introduction

Customer acquisition is no longer only a question of persuading a prospective buyer to exchange money for value. It is also a question of whether the invitation to buy was lawful, informed, fair, secure, and socially legitimate. Digital platforms allow firms to collect extensive behavioral data, personalize offers in real time, automate communication, and use artificial intelligence (AI) to rank, target, and convert prospects. These capabilities create commercial opportunities, but they also create risks: unauthorized surveillance, manipulation, discrimination, misleading claims, bribery, insecure data practices, and the erosion of consumer autonomy.

For entrepreneurial firms, ethics and governance are not peripheral compliance activities. They are mechanisms for protecting customers, managing strategic risk, preserving brand legitimacy, and building durable trust. A business may achieve short-term conversion through aggressive tactics, but long-term growth depends on repeat purchasing, referrals, employee commitment, investor confidence, and permission to operate in markets.

This chapter examines the principles and systems required for responsible customer acquisition. It addresses consent, privacy, consumer protection, anti-bribery, discrimination, dark patterns, responsible data use, AI governance, global selling, autonomous agents, predictive personalization, and trust-based growth.

Learning Objectives

By the end of this chapter, students should be able to:

1. Distinguish ethical, legal, and governance dimensions of customer acquisition.
2. Evaluate consent, privacy, consumer-protection, anti-bribery, and discrimination risks.
3. Identify dark patterns and assess their effects on consumer autonomy.
4. Design principles for responsible collection, use, retention, and sharing of customer data.
5. Establish governance controls for AI-assisted selling and automation.
6. Analyze global challenges involving jurisdiction, culture, language, and enforcement.
7. Assess autonomous agents, predictive personalization, and trust-based growth.
8. Apply an ethical decision framework to a customer-acquisition case.

1. From Conversion Efficiency to Legitimate Exchange

A conventional sales funnel treats acquisition as a sequence from awareness to consideration, purchase, and retention. An ethical perspective adds a legitimacy test at every stage. A prospect should understand who is communicating, why the communication is occurring, what is being offered, what data is being used, and what choices are available.

A useful distinction is:

• Lawfulness: conformity with applicable legislation, regulation, contractual duties, and enforceable standards.
• Ethics: judgment about what is fair, respectful, honest, and responsible, including conduct not yet specifically prohibited by law.
• Governance: the structures, roles, processes, controls, records, and incentives through which an organization directs and monitors conduct.
• Trust: a stakeholder’s justified expectation that the firm will act competently, honestly, and with regard for the stakeholder’s interests.

Legal compliance is a minimum threshold, not a complete ethical standard. A tactic may be technically permissible in one jurisdiction but still be deceptive, exploitative, or inconsistent with a firm’s stated values. Conversely, ethical governance should not rely only on individual goodwill; it should make responsible conduct easier and misconduct more difficult.

The central principle is informed and voluntary value exchange. Customers should receive genuine value, have meaningful choice, and not be subjected to deception, coercion, or exploitation of vulnerability.

2. Consent, Privacy, and Consumer Protection

2.1 Consent as a quality of permission

Consent is not merely a checkbox. High-quality consent is generally:

• Freely given: refusal does not create disproportionate disadvantage.
• Specific: permission relates to identified purposes rather than an unlimited future use.
• Informed: the person receives understandable information about the processing or communication.
• Unambiguous: an affirmative action indicates the person’s intention.
• Revocable: withdrawal is practical and does not require unreasonable effort.
• Documented: the organization can demonstrate what was agreed, when, and through which interface.

Different legal regimes use different categories and conditions for processing personal data. Organizations must therefore identify the jurisdictions and roles involved rather than assuming that one global consent design is sufficient. In some circumstances, processing may rely on a lawful basis other than consent, but that does not remove duties of transparency, minimization, security, fairness, and respect for individual rights.

Pre-ticked boxes, bundled consent, confusing language, forced account creation, and consent obtained through repeated interruption may produce a record of acceptance without producing meaningful choice. A responsible design separates essential service terms from optional marketing permissions and makes withdrawal no harder than granting permission.

2.2 Privacy as contextual integrity

Privacy is more than secrecy. It concerns whether information flows are appropriate to the context in which information was provided. A person who gives an email address to receive an invoice may not reasonably expect that address to be used for unrelated profiling or sold to third parties.

Important privacy questions include:

1. What information is collected?
2. Is each item necessary for a defined purpose?
3. Was it obtained directly, inferred, purchased, or generated by a device?
4. Could the use surprise or disadvantage the person?
5. Who can access it, and for how long?
6. How can an individual inspect, correct, delete, or restrict use where applicable?
7. What happens if a breach occurs?

Data minimization means collecting and retaining only information reasonably necessary for an identified purpose. Purpose limitation means not repurposing data in ways incompatible with the original context without an appropriate justification and communication. Privacy by design embeds these principles into systems before launch instead of treating privacy as a notice added afterward.

2.3 Consumer protection and truthful selling

Consumer-protection principles typically address misleading representations, hidden fees, unfair contract terms, false scarcity, bait-and-switch offers, undisclosed sponsorship, unsafe products, and obstacles to cancellation. Exact rules vary by jurisdiction, but responsible selling should satisfy four tests:

• Is the claim accurate and supported by evidence?
• Is material information presented before the customer commits?
• Could a reasonable customer form a false impression?
• Is the remedy, cancellation, or complaint route accessible?

Claims involving health, finance, environmental performance, security, or comparative superiority require especially strong substantiation. Testimonials and influencer communications should not imply independent experience when the speaker has been paid or otherwise incentivized without appropriate disclosure.

3. Anti-Bribery and Conflicts of Interest

Customer acquisition can create bribery risk through gifts, commissions, facilitation payments, charitable donations linked to a sale, third-party agents, public-sector procurement, and channel-partner incentives. Anti-bribery obligations may apply even when misconduct is performed by an intermediary.

A sound program includes:

• risk-based due diligence on agents, distributors, consultants, and resellers;
• written contracts defining services, compensation, and compliance duties;
• payments proportionate to legitimate services and supported by records;
• approval thresholds for gifts, hospitality, travel, and sponsorship;
• segregation of sales approval, payment, and audit responsibilities;
• training tailored to high-risk roles and markets;
• confidential reporting and protection against retaliation;
• investigation, remediation, and disciplinary processes.

A useful warning sign is a request for an unusual payment, an offshore account, vague consulting work, urgency designed to bypass approval, or compensation unrelated to documented commercial value. Ethical selling also requires disclosure and management of conflicts of interest. A salesperson who benefits personally from steering a customer toward a particular supplier may not be acting impartially, even if no criminal law is violated.

4. Discrimination and Fairness in Acquisition

Discrimination may be direct, indirect, or produced by automated systems. Direct discrimination occurs when a person is treated differently because of a protected characteristic. Indirect discrimination can arise when a seemingly neutral practice disproportionately disadvantages a group without sufficient justification. Algorithmic discrimination may occur when data, labels, objectives, or feedback loops encode historical inequality.

Examples include:

• showing premium offers disproportionately to affluent neighborhoods when location acts as a proxy for race or ethnicity;
• excluding older users from product education because a model predicts lower digital engagement;
• offering different prices or credit terms based on sensitive or proxy variables;
• using language or images that stereotype a group;
• optimizing for conversion in ways that systematically deny certain customers access to beneficial information.

Fairness is not always achieved by simply removing explicit protected attributes. Proxy variables, such as postcode, device type, language, school, or browsing pattern, may reproduce similar effects. Firms should conduct disparate-impact analysis where legally and operationally appropriate, test outcomes across relevant groups, document legitimate business reasons, and provide human review and appeal routes for consequential decisions.

Fairness also involves accessibility. Acquisition journeys should accommodate disability, language difference, low bandwidth, age-related needs, and varying levels of digital literacy. Inclusion expands the addressable market while reducing exclusionary harm.

5. Dark Patterns and Manipulation

A dark pattern is an interface or communication design that steers people toward an outcome they might not choose under conditions of clear, balanced, and easy choice. Common forms include:

• Forced continuity: a free trial silently becomes a paid subscription.
• Roach motel: joining is easy but cancellation is difficult.
• Confirmshaming: refusal is framed as foolish or irresponsible.
• Hidden costs: charges appear late in checkout.
• Misdirection: visual emphasis favors the firm’s preferred option.
• Disguised advertising: commercial content resembles neutral information.
• Privacy steering: accepting extensive data use is simple while declining is obscure.
• False urgency: a countdown or scarcity claim is fabricated or reset repeatedly.

The ethical issue is not persuasion itself. Persuasion is legitimate when it presents reasons and preserves agency. Manipulation exploits cognitive limitations, asymmetrical information, or emotional vulnerability to undermine reflective choice. A practical design test asks whether the customer can identify the options, understand the consequences, select a less data-intensive alternative, and reverse the decision without unreasonable friction.

6. Responsible Data Use and Measurement

Data can improve targeting, relevance, forecasting, and service, but more data does not automatically produce better decisions. Responsible data practice requires a documented lifecycle:

1. Collection: define purpose, necessity, source, and permission.
2. Quality assessment: identify missingness, errors, stale records, and representativeness problems.
3. Use: restrict access and prevent incompatible secondary uses.
4. Sharing: assess vendors, contracts, security, transfers, and onward use.
5. Retention: set deletion or review periods rather than retaining indefinitely.
6. Disposal: securely delete or irreversibly anonymize information.
7. Monitoring: investigate complaints, incidents, bias, and unexpected outcomes.

Pseudonymization replaces direct identifiers with codes but does not necessarily make data anonymous. Re-identification may be possible when datasets are combined. Anonymization claims should therefore be supported by a realistic assessment of linkability and auxiliary information.

Acquisition metrics also require ethical interpretation. Conversion rate can be written as:

The metric is meaningful only when “eligible prospects,” “desired conversion,” time period, and channel are defined consistently. A high conversion rate produced by deceptive scarcity or targeting vulnerable people is not evidence of healthy performance. Firms should supplement conversion with complaint rate, opt-out rate, refund rate, retention, accessibility outcomes, fairness indicators, data incidents, and customer-reported trust.

7. Governance for AI and Automation

AI may support lead scoring, content generation, conversational selling, pricing, recommendation, forecasting, and customer-service triage. Governance must address both the model and the wider sociotechnical system in which it operates.

7.1 A practical AI governance architecture

A mature governance system assigns:

• Business ownership: who is accountable for the use case and its outcomes?
• Data stewardship: are data provenance, quality, permissions, and retention controlled?
• Model oversight: are performance, robustness, explainability, and bias assessed?
• Human accountability: who reviews exceptions and handles appeals?
• Security: can prompts, customer records, or model outputs be exposed or manipulated?
• Vendor management: what claims, audit rights, service levels, and incident duties apply?
• Monitoring: how are drift, error, disparate outcomes, and misuse detected?
• Exit and remediation: how can the system be paused, corrected, or retired?

A risk-tiered approach is appropriate. A system drafting general product descriptions may need basic review, while a system influencing credit, insurance, employment-related purchasing, health decisions, or access to essential services requires much stronger controls and may be restricted by law.

7.2 Human oversight is not a ceremonial approval

A human-in-the-loop process is weak if the reviewer lacks time, authority, information, or expertise to challenge the system. Effective oversight specifies escalation triggers, review standards, documentation, and decision rights. Automation should not be used to conceal responsibility. “The algorithm decided” is not an acceptable account of organizational accountability.

7.3 Model risk and drift

A model can perform well during development and poorly after market conditions, customer behavior, or data definitions change. Monitoring should compare current performance with approved thresholds and examine subgroup outcomes. When a lead-scoring model changes the allocation of sales attention, managers should assess not only predictive accuracy but also whether the resulting distribution of attention is fair and commercially defensible.

8. Global Selling Challenges

Global customer acquisition involves more than translating a campaign. Firms must address:

• divergent privacy, marketing, consumer-protection, competition, and anti-bribery rules;
• cross-border data transfers and data localization requirements;
• local expectations about consent, authority, gender, negotiation, and privacy;
• currency, tax, refunds, delivery, and dispute resolution;
• language nuance and culturally specific meanings;
• sanctions, export controls, and restricted-party screening;
• local accessibility and digital infrastructure;
• different enforcement intensity and regulatory maturity.

A global policy should establish a high baseline of honesty, privacy, security, and anti-bribery while allowing local legal and cultural adaptation. The firm should map data flows, identify responsible entities, appoint local expertise where necessary, and avoid assuming that a customer’s physical location is obvious from an IP address alone.

Cultural adaptation must not become an excuse for stereotyping or unequal respect. Localization should be based on research and participation by local experts, with testing for unintended meanings and exclusion.

9. Emerging Developments

9.1 Autonomous selling agents

Autonomous agents can discover prospects, qualify leads, send messages, negotiate within limits, schedule meetings, and update customer records. Their benefits include speed, continuity, and scalability. Their risks include unauthorized commitments, fabricated claims, excessive contact, identity deception, prompt injection, data leakage, and unclear liability.

Controls should include clear disclosure that an agent is being used where material, approved communication boundaries, contact-frequency limits, restricted authority to negotiate, human escalation, reliable records, identity verification, and immediate shutdown capability. Agents should not invent evidence, imply human judgment where none occurred, or make commitments beyond delegated authority.

9.2 Predictive personalization

Predictive personalization uses observed or inferred attributes to estimate what content, product, price, or timing may be relevant. It can reduce information overload, but it can also create filter bubbles, exploit vulnerability, or produce unequal offers. The firm should distinguish personalization that improves relevance from personalization that manipulates susceptibility.

A responsible program tests whether the inference is necessary, accurate enough for its purpose, expected by the customer, and proportionate to the impact. Sensitive inferences should receive heightened scrutiny, and customers should not be denied meaningful alternatives merely because a model predicts low profitability.

9.3 Trust-based growth

Trust-based growth treats credibility as a strategic asset rather than a slogan. It is built through accurate promises, reliable fulfillment, transparent data practices, fair recovery after failure, and consistency between internal conduct and external claims. Trust is not blind confidence; it is confidence supported by evidence and accountability.

A useful strategic principle is to optimize for durable customer value, not merely immediate conversion. This means considering retention, advocacy, complaint burden, customer lifetime relationships, and the social cost of acquisition tactics. Short-term growth that creates regulatory penalties, employee turnover, public criticism, or customer betrayal may destroy enterprise value.

10. An Integrated Decision Framework

Managers can evaluate a proposed acquisition practice through the following sequence:

1. Define the objective: What customer and business outcome is sought?
2. Identify affected parties: Include prospects, customers, employees, suppliers, non-users, and vulnerable groups.
3. Map information and power: What does the firm know, and what can the customer realistically understand or refuse?
4. Test legality and policy: Which jurisdictions, contracts, standards, and internal commitments apply?
5. Test autonomy and fairness: Is the choice informed, reversible, accessible, and non-discriminatory?
6. Assess proportionality: Is the data, persuasion, or automation necessary for the benefit obtained?
7. Design controls: Specify permissions, limits, monitoring, human review, records, and remedies.
8. Pilot and measure: Examine commercial, customer, fairness, privacy, and trust outcomes.
9. Decide and document: Record the rationale, dissenting views, approval, and review date.
10. Learn and remediate: Treat incidents and complaints as evidence for system improvement.

This framework converts abstract values into managerial decisions. It also encourages “pre-mortem” analysis: imagine that the campaign has caused harm or public criticism, then identify how that failure could have occurred.

Applied Case: Illustrative Scenario—NexaHome’s Smart-Device Launch

Background

NexaHome, a fictional entrepreneurial firm, sells connected home-security devices through online advertising, retail partners, and a conversational AI sales assistant. It plans to launch a premium monitoring subscription. The marketing team proposes using browsing behavior, household location, device ownership, and inferred “security anxiety” to target households with personalized urgency messages.

Problem

The proposed campaign would use a third-party data broker, create a model that predicts willingness to pay, offer different prices to different households, and make cancellation available only through telephone support. The AI assistant would not initially disclose that it was automated. A distributor also suggests a large “relationship payment” to a public-sector housing manager who could recommend the product.

Analysis

The plan presents several risks. Broker data may lack valid permission or contextual appropriateness. “Security anxiety” is a sensitive and potentially exploitative inference. Differential pricing may be unlawful or unfair if based on protected characteristics or proxies. A difficult cancellation process resembles forced continuity. Failure to disclose automation may mislead customers. The proposed payment creates bribery risk, especially because it is linked to access and influence rather than documented services.

Approach and solution

NexaHome redesigns the campaign as follows:

• It removes the broker data and uses first-party information collected for clearly stated purposes.
• It prohibits sensitive vulnerability inferences and reviews location-based targeting for disparate effects.
• It uses transparent, evidence-based product claims and displays all material fees before purchase.
• It discloses the AI assistant and offers easy transfer to a trained human.
• It provides online cancellation with no greater friction than enrollment.
• It establishes one standard price, with documented discounts available on objective, publicly stated criteria.
• It rejects the relationship payment and conducts due diligence on the distributor.
• It creates an approval record, a complaint dashboard, and a post-launch review across customer groups.

Outcome

The redesigned campaign produces a lower initial conversion rate but fewer complaints, refunds, and opt-outs. Customer-support costs decline because expectations are clearer. Retail partners report greater confidence in recommending the product, and the firm can explain its acquisition practices to regulators, investors, and customers.

Learning points

The case demonstrates that ethical governance can alter targeting, pricing, interface design, automation, and channel management simultaneously. It also shows why compliance review should occur before launch: once a model, campaign, or partner practice is embedded, remediation becomes more expensive and reputational damage may already have occurred.

Challenges and Limitations

Ethical governance involves unavoidable trade-offs. Personalization can improve relevance but requires information. Fraud prevention can protect customers but may generate false positives. Human review improves accountability but increases cost and may itself be biased. Global consistency can simplify governance but may conflict with local law or context.

No framework eliminates judgment. Privacy notices can be technically complete yet incomprehensible. Fairness metrics may conflict, and a model can satisfy one metric while failing another. Customer surveys may underrepresent people who were excluded before they could respond. Audit results can become outdated as systems and markets change.

Accordingly, governance should be iterative, evidence-based, and proportionate to risk. Firms should invite independent review for high-impact systems, publish meaningful explanations where appropriate, preserve records, and provide remedies rather than treating governance as a one-time certification.

Best-Practice Checklist

Before launching an acquisition initiative, an entrepreneurial firm should be able to answer “yes” to the following questions:

• Is the value proposition accurate and substantiated?
• Are purposes, permissions, and data sources documented?
• Can customers refuse, withdraw, cancel, and complain without unreasonable friction?
• Have vulnerable and underrepresented groups been considered?
• Are automated decisions explainable enough for their consequences?
• Are human escalation and remedy mechanisms real and resourced?
• Have vendors, agents, and distributors been screened and contractually controlled?
• Are retention, deletion, security, and incident procedures defined?
• Are success metrics broader than immediate conversion?
• Is there an accountable executive and a scheduled review?

Chapter Summary

Ethical customer acquisition is the practice of creating legitimate, informed, and fair exchanges rather than maximizing conversion by any available means. Consent must be meaningful, privacy practices must respect context and purpose, and consumer claims must be accurate and transparent. Anti-bribery controls must extend to intermediaries, while discrimination analysis must address direct, indirect, and algorithmic effects. Dark patterns undermine autonomy by making refusal, cancellation, or privacy protection unnecessarily difficult.

Responsible data use requires minimization, purpose limitation, security, retention controls, and attention to inference and re-identification. AI governance assigns ownership, tests data and models, provides meaningful human oversight, monitors drift and disparate outcomes, and establishes remediation. Global selling requires legal mapping, cultural competence, and responsible cross-border data practices. Autonomous agents and predictive personalization offer major efficiency gains but require disclosure, authority limits, fairness controls, and human accountability.

The strategic conclusion is that trust is not opposed to growth. Accurate promises, fair design, responsible data use, and effective remedies support durable relationships, lower reputational risk, and create trust-based growth. Governance should therefore be treated as a capability for sustainable entrepreneurship.

Key Terms

Consent: A voluntary, informed, specific, and revocable indication of agreement to a defined action or use.

Consumer protection: Laws and practices that prevent deception, unfairness, unsafe products, and abusive selling.

Dark pattern: A design that manipulates or steers users against their informed interests.

Data minimization: Collecting and retaining only data necessary for a defined purpose.

Discrimination: Unequal or unjust treatment based on protected or proxy characteristics.

Fairness: The equitable treatment and impact of people or groups in a decision or process.

Governance: Structures and controls used to direct, monitor, and hold an organization accountable.

Inference: A conclusion or attribute generated from data rather than directly supplied by a person.

Purpose limitation: Restricting data use to specified, compatible purposes.

Pseudonymization: Replacing direct identifiers with codes while retaining possible re-identification risk.

Privacy by design: Building privacy safeguards into systems and processes from the outset.

Predictive personalization: Using predictions about a person or group to tailor content, offers, or timing.

Responsible data use: Data practice that is lawful, necessary, secure, fair, transparent, and accountable.

Trust-based growth: Growth founded on reliable performance, honest communication, and customer confidence.

AI governance: Policies, roles, controls, and monitoring for responsible development and use of AI.

Autonomous agent: Software that can pursue delegated tasks and take actions with limited direct human intervention.

Bribery: Offering, giving, requesting, or accepting an undue advantage to influence a decision.

Human oversight: Meaningful human capacity to understand, challenge, and intervene in an automated process.

Review Questions

Short Answer Questions

1. What distinguishes lawfulness from ethical conduct in customer acquisition?
2. Identify four characteristics of meaningful consent.
3. What is a dark pattern? Give two examples.
4. Why can removing protected attributes fail to eliminate algorithmic discrimination?
5. Name three controls relevant to autonomous selling agents.

Long Answer Questions

1. Explain how privacy, consumer protection, and trust interact in a digital acquisition strategy.
2. Develop an AI governance architecture for a lead-scoring system used by a growing enterprise.
3. Compare legitimate persuasion with manipulative personalization.
4. Discuss the principal anti-bribery risks arising from international sales intermediaries.
5. Evaluate the claim that ethical governance reduces rather than increases entrepreneurial competitiveness.

Analytical / Application Questions

1. A retailer uses purchase history and location to send different discounts to customers. Design an analysis to determine whether the practice is acceptable, considering privacy, fairness, transparency, and commercial necessity.
2. An AI sales agent repeatedly contacts prospects who have not responded and claims that an offer expires in ten minutes, although the offer is permanent. Identify the ethical and governance failures and propose controls.
3. Design a pre-launch review for a global campaign that uses predictive personalization, including stakeholders, evidence, approval criteria, monitoring metrics, and remediation steps.

Multiple Choice Questions with Answers

Question 1

Which feature most clearly distinguishes meaningful consent from a mere record of acceptance?

A. The consent box is displayed in a bright color

B. Consent is freely given, informed, specific, and practically revocable

C. The customer accepts a long document without reading it

D. Consent is bundled with an essential service and cannot be declined

Correct Answer: B. Consent is freely given, informed, specific, and practically revocable

Question 2

Which practice is an example of a dark pattern?

A. Displaying the total price before payment

B. Offering a clear online cancellation route

C. Making subscription enrollment easy but cancellation difficult

D. Explaining how marketing preferences can be changed

Correct Answer: C. Making subscription enrollment easy but cancellation difficult

Question 3

Why can an algorithm discriminate even when protected attributes are excluded?

A. Algorithms cannot process numerical data

B. Proxy variables may reproduce the effects of protected attributes

C. Removing attributes always makes predictions random

D. Discrimination occurs only through human decisions

Correct Answer: B. Proxy variables may reproduce the effects of protected attributes

Question 4

Which control is most directly associated with anti-bribery governance?

A. Allowing undocumented commission payments

B. Risk-based due diligence on sales intermediaries

C. Removing approval thresholds for hospitality

D. Paying agents into accounts unrelated to their services

Correct Answer: B. Risk-based due diligence on sales intermediaries

Question 5

What is the primary purpose of data minimization?

A. To collect as much data as possible for future use

B. To retain all data indefinitely

C. To collect and retain only data necessary for a defined purpose

D. To make every dataset publicly available

Correct Answer: C. To collect and retain only data necessary for a defined purpose

Question 6

Which statement best describes meaningful human oversight of an AI system?

A. A manager approves every output automatically

B. A reviewer has authority, information, and capacity to challenge the system

C. The organization assigns responsibility to the algorithm

D. A human is named in policy but cannot intervene

Correct Answer: B. A reviewer has authority, information, and capacity to challenge the system

Question 7

A conversion rate is 240 conversions from 3,000 eligible prospects. What is the conversion rate?

A. 0.8%

B. 8%

C. 12.5%

D. 80%

Correct Answer: B. 8%

Question 8

Which practice is most consistent with responsible predictive personalization?

A. Using inferred vulnerability to create pressure purchases

B. Using necessary data transparently while testing for unfair outcomes

C. Hiding personalization so customers cannot question it

D. Denying all alternatives to customers predicted to be less profitable

Correct Answer: B. Using necessary data transparently while testing for unfair outcomes

Question 9

What is a major risk of an autonomous selling agent?

A. It cannot communicate with any customer

B. It may make unauthorized commitments or fabricate claims

C. It eliminates the need for governance

D. It always produces identical messages

Correct Answer: B. It may make unauthorized commitments or fabricate claims

Question 10

Which approach best supports trust-based growth?

A. Optimizing only immediate conversion

B. Using hidden fees to increase initial revenue

C. Combining accurate promises, fair design, reliable delivery, and remedies

D. Avoiding complaint measurement because it may reduce reported performance

Correct Answer: C. Combining accurate promises, fair design, reliable delivery, and remedies