Cuestionario vinculado a una guía · Medio Nivel

Prueba de IA en marketing

Prueba los objetivos de la audiencia, las afirmaciones veraces, los proxy sensibles, los resultados posteriores, las versiones de los experimentos, el consentimiento y la aprobación humana.

Rutas de guía relacionadasMarketing de IA
Pregunta 1 de 8

A campaign team has not decided whether it wants awareness, leads or purchases. What should it do before choosing an AI model?

Vista previa de cada pregunta8 preguntas

Las respuestas y explicaciones aparecen a medida que realiza el cuestionario.

  1. A campaign team has not decided whether it wants awareness, leads or purchases. What should it do before choosing an AI model?
    • Optimize impressions by default
    • Define the audience need and intended outcome
    • Select the newest model
    • Write copy before identifying the audience
  2. A generated ad claims a product “cuts costs by 40%.” What review is needed?
    • Verify the claim against current evidence before publication
    • Keep it if the language is persuasive
    • Add a disclaimer instead of checking
    • Ask the model to make the number more specific
  3. Why can a campaign with more clicks still be unsuccessful?
    • Clicks always equal customer value
    • Impressions are the only valid outcome
    • A high click-through rate proves accessibility
    • Sensational wording may attract unqualified users, complaints or unsubscribes
  4. A model selects audiences using a sensitive proxy unrelated to the offer. Which concern applies?
    • A proxy always improves relevance
    • Sensitive attributes guarantee consent
    • The targeting may be irrelevant or unfair and should be reviewed
    • Audience selection does not affect people
  5. Which set of outcomes better evaluates a retention campaign?
    • Qualified conversions, refunds, complaints, and unsubscribe rates
    • Impressions and clicks only
    • Model size and generation speed
    • Number of words in the ad
  6. Why version the audience, creative and landing page used in an experiment?
    • To prevent users from seeing an ad
    • To know which combination produced the observed result
    • To make the model deterministic
    • To remove the need for consent
  7. A marketing assistant generates a testimonial from a prompt with no customer source. What should the team do?
    • Use it because the model writes naturally
    • Add a star rating and publish
    • Ask a second model to invent a name
    • Do not present it as a real customer experience without verifying its basis
  8. Who should approve sensitive or public-facing AI marketing claims?
    • The model alone
    • The ad platform’s ranking algorithm
    • A human reviewer with authority under the organization’s approval process
    • The customer segment itself