Learning path · Level 9 · Tutorial 1

Choose Marketing Tasks for AI Assistance

Choose suitable marketing tasks for AI assistance by weighing likely time savings, quality requirements, information risks, and current free-tier limits. Make a defensible choice without purchasing software or using real customer data.

  • Level: Intermediate
  • Duration: 13 min
  • 7 steps
Start tutorial Back to the path
Choose Marketing Tasks for AI Assistance

What you will go through

  1. Separate AI Assistance from Marketing Judgment Learn to distinguish useful AI assistance from the human judgment required for responsible marketing work. 1 min
  2. Identify Tasks That Are Easy to Bound and Check 2 min
  3. Route Tasks by Risk and Available Safeguards Route marketing tasks by their consequences, available safeguards, and the authority needed to act. 2 min
  4. Check Which Inputs May Be Used Identify information to exclude and select safe fictional replacements before considering an AI tool. 2 min
  5. Estimate Net Time Saved Estimate total effort for an AI-assisted marketing draft and see how review time can change the recommended route. 2 min
  6. Compare Free-Tier Tools Against the Task Compare fictional free-tier tool profiles with a defined marketing task, then identify the conditions that require current verification before real use. 2 min
  7. Make and Defend Your Task Selection Integrate task fit, input eligibility, risk, review, time, and fictional free-tier conditions into a concise Trailside Gear selection record. 3 min

What you will learn

  • Classify marketing tasks as suitable for AI assistance, better handled manually, or requiring specialist review.
  • Estimate whether preparation and review effort could outweigh the time saved on an AI-generated draft.
  • Compare free-tier tool options against a task's required features, usage limits, and data-handling conditions.
  • Identify information that must be excluded or replaced with fictional material before using an AI tool.

Before you start

  • Write a Focused Marketing Strategy Brief
  • Build an Editorial Plan from a Strategy Brief
  • Create Readable Marketing Visuals with Free Tools

Step 1 of 7

Separate AI Assistance from Marketing Judgment

Learn to distinguish useful AI assistance from the human judgment required for responsible marketing work.

Use AI for working material

Assistance is not judgment

Generative AI can help inside an established marketing task. It can produce a first draft, summarize supplied notes, or organize ideas into a usable format.

That help is not marketing judgment. A fluent response is not evidence that every statement is accurate, complete, or current. More output—or faster output—does not guarantee stronger marketing results.

Treat an AI-assisted draft as working material. A person remains responsible for deciding whether it fits the goal and for giving human approval.

A diagram showing a marketing brief becoming an AI-assisted draft, followed by human review and approval.

AI can create working material; people check it and decide whether to use it.

Apply the distinction

Example

A bounded drafting example

Established task: Draft 3 social caption options for the Trailside Daypack, with a maximum of 30 words per caption.

Approved public product facts: 20 L capacity; 2 external pockets.

AI may turn those supplied facts into an AI-assisted draft, organize alternative wording, or summarize notes already provided. It does not independently establish new product facts, prove that the captions will perform well, or authorize publication.

A marketer still uses judgment about the task and gives human approval before use.

Tip

Keep the boundary clear

Plausible language can still be wrong. For example, an attractive caption could imply a benefit that was never supplied. Speed only changes how quickly working material appears; it does not remove the need for human judgment.

Choose the accurate boundary

Check your understanding

Which statement correctly describes the boundary between AI assistance and marketing judgment?

Step 2 of 7

Identify Tasks That Are Easy to Bound and Check

Check the shape of the task

Four suitability criteria

Before investing effort in AI assistance, test whether a marketing task is easy to define and check. Use these four suitability criteria:

  1. Bounded inputs — the material the draft may use is defined.
  2. Clear output requirements — the deliverable and its requirements are stated, so quality can be assessed.
  3. Ease of human verification — an appropriate reviewer can check the draft using available information and their competence.
  4. Reversibility — the output remains an editable draft that can be changed or discarded before it affects customers, budgets, or publication.

A task does not become suitable merely because generated wording sounds convincing.

See the four checks separately

Diagram linking an unfinished caption draft to a bordered source card, checklist, magnifying glass, and undo arrow.

A stronger AI-assisted draft task has defined inputs, assessable requirements, feasible human verification, and a reversible outcome.

Apply the matrix to a caption draft

A practical suitability matrix

Use the matrix before generating. A specific deliverable belongs under clear output requirements: it tells you what must be produced and how you can assess it.

Four separate checks for an AI-assisted draft task.
CriterionAsk before generatingStrong signal
Bounded inputsWhat material may the draft use?A small, defined source set
Clear output requirementsWhat deliverable and requirements must it meet?Count, format, length, and allowed facts are stated
Ease of human verificationCan an appropriate reviewer check it with available information?The reviewer can compare the draft with reliable source material
ReversibilityCan the output be changed or discarded before it has an effect?It remains an editable draft before human approval

Example

Worked task shape

Candidate task: Draft 3 social caption options for the Trailside Daypack, with a maximum of 30 words per caption.

Defined source facts: 20 L capacity; 2 external pockets.

The inputs are bounded by the two supplied facts. The output requirements name three captions and a maximum length. A reviewer can check each AI-assisted draft against the supplied facts and requirements. The result remains a draft: it can be edited or discarded before human review and human approval.

By contrast, “Improve Trailside Gear marketing” does not state a deliverable or requirements that can be checked.

Tip

Do not skip a check

A task can be clear in one respect and weak in another. For example, a named deliverable may still be unsuitable if nobody can verify its information, or if it immediately sends something to customers. Check all four criteria rather than relying on one positive signal.

Practice: diagnose task fit

Diagnose task fit

Match each candidate-task weakness, or the strong candidate, to the best description.

Tap an item and then its matching pair.

Step 3 of 7

Route Tasks by Risk and Available Safeguards

Route marketing tasks by their consequences, available safeguards, and the authority needed to act.

Assess the consequence before the speed

Start with consequences

A task can be quick to generate and still carry serious risk. Start with the consequence of an error: could it mislead customers, make an unsupported product claim, exclude people, or reach someone without a human check?

The suitability criteria you already know still matter. But assess risk and available safeguards before considering speed or time saved. A plausible draft is not a reason to lower oversight.

Risk comes before productivity

A decision map showing consequence and safeguard checks before three possible task routes.

First ask what an error could affect. Then ask whether suitable safeguards and authority are available. Consider productivity only after those checks.

Route according to risk and authority

Three routes

Choose a provisional route based on the task’s consequence and the safeguards that are actually available. These routes are not rankings of importance; they are ways to keep responsibility aligned with risk.

Warning

Important boundary

A manual approach is not a workaround for an unsafe task. If a task requires expertise or authority you do not have, doing it manually does not make it acceptable. Seek specialist review instead. Specialist review can lead to rejection: it is not permission to publish an unsupported claim.

Route examples

Examples of provisional routes for Trailside Gear

Each route depends on the stated conditions, not on how fast a tool can generate words.
DecisionWhen it fitsExample
AI assistanceBounded, reversible draft; human review and human approval are availableDraft 3 social caption options, up to 30 words each, using approved public facts
Manual approachAccountable marketing judgment is centralChoose a campaign priority when the team must weigh conflicting business priorities
Specialist reviewClaims, sensitive uses, or requirements exceed the marketer’s competence or authorityAssess a proposed product claim that could affect customer wellbeing

Check the safeguard before using a tool

Check the safeguard

A colleague proposes using a generator to write replies to every customer complaint and publishing them automatically, with no human review or human approval. Which factor makes AI assistance unsuitable as proposed?

Choose the best answer.

Apply the decision map

Make a provisional route decision

For each task below, select one label—AI assistance, Manual approach, or Specialist review—and give one concise reason based on consequence and safeguards, not speed.

A. Draft 3 social caption options for the Trailside Daypack. Each caption is no more than 30 words and uses only the approved public facts: 20 L capacity and 2 external pockets. Human review and human approval are available.

B. Decide whether the next campaign should prioritize awareness or conversion when the strategy team has conflicting priorities.

C. Decide whether a proposed claim that the Trailside Daypack “prevents injury” can be used before publication.

Write at least 120 characters (0/120).

Step 4 of 7

Check Which Inputs May Be Used

Identify information to exclude and select safe fictional replacements before considering an AI tool.

Screen the material first

Start with input eligibility

Before considering an external AI tool, screen every input—not only the task.

For this tutorial, eligible practice inputs are:

  • Fictional material that is wholly invented.
  • Public material that is permitted for the intended use.

Use the labels Public, Fictional, Personal, Confidential, and Restricted. Labels can overlap rather than forming exclusive categories. For example, the Trailside Gear practice context is Fictional, while its approved product facts are also Public.

For Trailside Gear, the approved public product facts are 20 L capacity and 2 external pockets. They can support an AI-assisted draft of 3 social caption options, with a maximum of 30 words per caption.

Screen inputs before drafting

Diagram showing fictional and permitted public inputs moving toward an AI drafting workspace, while personal, confidential, and restricted inputs are blocked.

Eligible practice inputs move forward only after they pass an input check.

Exclude, then replace safely

Inputs that do not belong in an external tool

Exclude Personal, Confidential, and Restricted material from external AI inputs unless an authorized arrangement explicitly permits that use.

Examples to exclude include:

  • customer names, email addresses, or contact details;
  • customer records or identifiable stories;
  • confidential plans, unpublished results, or internal pricing discussions;
  • material marked restricted or subject to organizational controls.

Do not assume that public availability solves every issue. Public material still needs permission for the intended use, and personal information does not become appropriate to upload just because it appears online.

Example

A fictional replacement changes the whole example

Removing a customer’s name is not the same as creating a fictional example.

Suppose a source says that a customer bought a Trailside Daypack on a particular date, lives in a small named town, and described a specific delivery problem. Even without the name, the remaining details may identify the person or expose confidential service information.

For practice, replace the entire situation with a wholly invented one. Do not try to repair real customer material by deleting a few details.

Warning

Do not force an AI route

If removing ineligible information makes the task impossible, do not upload it to make the task work. Choose an approved Manual approach or seek authorized guidance. AI assistance is not a reason to override information restrictions.

Choose a safe route

Use an input gate

Use this input gate before evaluating a tool:

  1. Label each input with every label that applies.
  2. Keep only wholly Fictional material or Public material that is permitted for the intended use.
  3. Exclude Personal, Confidential, and Restricted material unless an authorized arrangement explicitly allows the use.
  4. If the remaining material cannot support the task, use a Manual approach or seek authorized guidance.

This check comes before any possible time saving. A fast draft is not useful if its inputs should not be used.

Tip

Keep the practice source bounded

For the practice caption task, a safe source can be limited to the Trailside Gear name, the product name Trailside Daypack, and the approved public facts: 20 L capacity and 2 external pockets. A later tutorial covers how to structure source material and constraints in a prompt.

Practice: classify before use

Classify inputs before use

For each case, write every applicable input label from this list: Public, Fictional, Personal, Confidential, Restricted. Then state the handling: eligible for practice, exclude from an external AI input, or use a Manual approach or seek authorized guidance.

A. The Trailside Daypack fact sheet supplied for this tutorial: 20 L capacity and 2 external pockets.

B. An invented hiker and an invented comment created solely for practice.

C. A customer service story with the name removed, but a distinctive order date, a small named town, and a specific delivery problem retained.

D. An unpublished internal launch plan marked for restricted internal access.

Write at least 120 characters (0/120).

Step 5 of 7

Estimate Net Time Saved

Estimate total effort for an AI-assisted marketing draft and see how review time can change the recommended route.

Compare equivalent work

A fair comparison

Compare the same deliverable

A time comparison is useful only when both routes aim for the same output and quality standard.

For Trailside Gear, compare drafting 3 social caption options, with a maximum of 30 words per caption, using only the approved public facts: 20 L capacity and 2 external pockets. An AI-assisted draft is working material, not approved copy.

Do not compare a fast unreviewed draft with a careful manual result. Include the effort needed to make either route ready for human approval.

See the hidden work

A solid manual-time bar compared with an AI-assisted bar divided into four stages, showing that generation is only one part of the work.

Compare total effort, not just the short generation segment.

Calculate the main scenario

Count every stage

Use these formulas for a Scenario estimate:

AI-assisted time = preparation + generation and waiting + checking + revision and approval

Net time saved = manual time - AI-assisted time

Checking is human review of the AI-assisted draft. Approval is human approval to authorize its use. Count both, along with preparation and any necessary revision.

Scenario estimate

Scenario estimate for the same caption task and quality requirement.

The AI-assisted route saves time only after all four stages are included.
Route or stageTime
Manual approach25 min
Preparation5 min
Generation and waiting2 min
Checking8 min
Revision and approval5 min
AI-assisted total20 min

Example

Work the estimate

Scenario estimate

Manual time: 25 min

AI-assisted time: 5 + 2 + 8 + 5 = 20 min

Net time saved: 25 - 20 = 5 min

This estimate supports AI assistance for a draft only if the earlier input, risk, and review safeguards are also satisfied.

Test a higher-review scenario

Review can erase the gain

A fast first draft does not ensure a time advantage. Strict quality requirements or material that is difficult to verify can raise the human review time.

Use Scenario estimates as a starting point, not proof of productivity. If checking becomes much longer, recalculate before choosing a route.

Scenario estimate — higher review effort

Scenario estimate — higher review effort: The task and manual estimate stay the same, but checking requires more human review.

When net time saved is zero or negative, there is no estimated time advantage over the Manual approach.
Route or stageTime
Manual approach25 min
Preparation5 min
Generation and waiting2 min
Checking14 min
Revision and approval5 min
AI-assisted total26 min
Net time saved-1 min

Apply the calculation

Make the time-based choice

For the eligible Trailside Gear task, draft 3 social caption options of no more than 30 words each. Assume safeguards are available.

Scenario estimate A: Manual approach: 18 min. AI-assisted stages: 4 min preparation, 2 min generation and waiting, 5 min checking, and 3 min revision and approval.

Scenario estimate B: The task has stricter quality requirements. Keep all figures from A, but checking takes 10 min.

Calculate the AI-assisted time and net time saved for A and B. Then state which route you would choose for each Scenario estimate.

Write at least 55 characters (0/55).

Step 6 of 7

Compare Free-Tier Tools Against the Task

Compare fictional free-tier tool profiles with a defined marketing task, then identify the conditions that require current verification before real use.

Define the fit before comparing tools

Start with the task requirements

Free-tier fit is task-specific. Start with what the task must produce—not with the largest quota.

Candidate task: Draft 3 social caption options for the fictional Trailside Gear product, Trailside Daypack. Each caption may contain up to 30 words. The AI-assisted draft must contain all 3 options in one text response, so it needs capacity for up to 90 words.

Eligible practice inputs:

  • Public: 20 L capacity; 2 external pockets
  • Fictional: Trailside Gear and Trailside Daypack

The task also needs text generation, at least 1 request today, and an account/access route that is acceptable for the intended use.

Compare requirements before quotas

A checklist of output, quota, account, and data conditions is compared with two fictional tool cards; one has a mismatch and the other needs verification.

A tool is a possible fit only when its feature, limits, access conditions, and data conditions match the specific task.

Tip

Decision rule

A high quota cannot compensate for a missing required feature or an unacceptable data condition. Treat every unknown as a condition to check, not as a benefit.

Read the profiles as constraints, not promises

Fictional tool profiles — not current provider policies

Fictional tool profiles — not current provider policies. These profiles are only a practice comparison; do not treat them as facts about any real provider.

Compare each profile against the Trailside Daypack task requirements.
ToolOutput limitQuotaAccountData profileAccess
Tool A40 words per request10 requests per dayEmail sign-in requiredRetained 30 days; may be used to improve systems; no free-tier controlBrowser where offered
Tool B120 words per request3 requests per dayEmail sign-in requiredRetained 7 days for abuse prevention; not used for training in this profile; no custom controlBrowser where offered

Compare, then verify current conditions

For the defined task, Tool A's 40-word output limit cannot hold up to 3 captions of 30 words in one response. Its larger quota does not solve that mismatch. Tool B has enough output capacity and quota in this fictional comparison.

Before real use, verify current conditions from the provider's pricing or free-tier page, help documentation, terms, and privacy information. Check the actual feature and output limits, requests per day, account and access requirements, retention, possible training use, available controls, and compatibility with organizational restrictions.

Conditions change. If a required condition is unknown or incompatible, use a Manual approach or obtain authorized clarification; do not assume the tool is suitable.

Check your decision rule

Handle an unresolved condition

A provider's current privacy information does not clearly show whether its data handling is compatible with organizational restrictions. What is the sound next step?

Apply the comparison to the task

Make a task-specific tool selection

Using the fictional profiles from the previous screen, choose Tool A, Tool B, or reject both for the Trailside Daypack caption task. Defend your choice using the output feature, requests per day quota, account/access condition, and data profile. Then name two current conditions that must be verified before real use.

Write at least 180 characters (0/180).

Step 7 of 7

Make and Defend Your Task Selection

Integrate task fit, input eligibility, risk, review, time, and fictional free-tier conditions into a concise Trailside Gear selection record.

Work from a bounded case file

Work from a bounded case file

Use this fictional case file to make one defensible selection—without opening an AI tool.

For the caption task, use only the approved Public facts for the fictional Trailside Daypack: 20 L capacity and 2 external pockets. The Trailside Gear context is Fictional, so the usable practice inputs carry both labels: Public and Fictional.

Use the full estimate, not generation speed alone:

AI-assisted time = preparation + generation and waiting + checking + revision and approval

Net time saved = manual time - AI-assisted time

Scenario estimate: 25 min manually versus 5 + 2 + 8 + 5 = 20 min with AI assistance. The estimated net time saved is 5 min.

Do not write or run a prompt in this activity. Your decision comes first.

Connect the checks before choosing

A worksheet shows task boundaries, input eligibility, risk, review, time, and tool conditions feeding into a route decision.

A route is defensible only when the task, inputs, safeguards, time estimate, and tool conditions fit together.

Trailside Gear task shortlist

Assess these supplied fictional tasks before selecting a route.

Use the same quality and safety checks across all three tasks.
TaskOutput requirementsInput labelsScenario estimate
Draft 3 social caption optionsMaximum 30 words per captionPublic; Fictional25 min manual; 20 min AI-assisted
Set final wording for a customer-complaint replyOne final customer-facing replyFictionalNo estimate before routing
Approve a performance claim and audience-targeting categoryApproval decisionFictional; RestrictedNo estimate before routing

Check the tool fit

Compare the free-tier conditions

The caption task needs text generation for all 3 options in one response: up to 90 words total. It also needs at least 1 request today and an acceptable account, browser access, and data-handling condition.

These are practice profiles only. Before any real use, current provider information must be checked. An unknown or incompatible required condition means using a Manual approach or obtaining authorized clarification.

Tool conditions for the case

Compare the profiles against the caption task—not against each other in the abstract.

Fictional tool profiles — not current provider policies
ToolOutput limitQuotaAccount and accessData profile
Tool A40 words per request10 requests per dayEmail sign-in required; browser where offeredRetained 30 days; may be used to improve systems; no free-tier control
Tool B120 words per request3 requests per dayEmail sign-in required; browser where offeredRetained 7 days for abuse prevention; not used for training in this profile; no custom control

Tip

Keep the selection conditional

A larger quota does not repair a missing feature. Tool A cannot provide up to 90 words in one response. Tool B has enough output capacity and an adequate illustrative quota for this task. Its profile is only a provisional fit: current feature limits, requests per day, account/access requirements, retention, training use, controls, and compatibility with organizational restrictions still need verification before real use.

Apply the risk and safeguard checks

Route the three tasks

Match each Trailside Gear task to its most appropriate provisional route.

Tap an item and then its matching pair.

Apply the complete decision

Build and defend the selection record

Create one concise task-selection record from the case file.

First, route all three tasks using exactly one label each: AI assistance, Manual approach, or Specialist review.

Then make the record for the task selected for AI assistance. Use every heading exactly as shown:

Task
Output requirements
Input eligibility
Risk and reviewer
Estimated net time saved
Free-tier fit
Decision
Reason

Include the eligible inputs, the quality limit, the Scenario estimate, human review and human approval, the fitting fictional tool, a condition that would change the recommendation, and a Manual approach fallback if that condition remains unresolved.

Write at least 450 characters (0/450).

Review and complete

Summary

Selection checklist

  • Route by consequences and safeguards before considering speed.
  • For Trailside Gear, bounded caption drafts can use AI assistance only with eligible Public and Fictional inputs, human review, and human approval.
  • Use a Manual approach for accountable direct communication when ordinary AI safeguards are not enough.
  • Use Specialist review for restricted, claim, or targeting decisions outside the marketer’s authority.
  • A free-tier option fits only when feature capacity, quota, account/access, and data-handling conditions fit the task. If a required condition is unresolved, retain a Manual approach fallback.
  • The next tutorial focuses on writing a source-grounded prompt. This tutorial stops at selecting the task and its boundaries.

Tutorial completed

Congratulations! You completed: Choose Marketing Tasks for AI Assistance

You can now select a marketing task for AI assistance with clear boundaries, realistic time estimates, and a defensible fallback.

Download the App now to have access to + 5000 free courses, exercises, certificates and lots of content without paying anything!

  • 100% free online courses from start to finish

    Thousands of online courses in video, ebooks and audiobooks.

  • More than 60 thousand free exercises

    To test your knowledge during online courses

  • Valid free Digital Certificate with QR Code

    Generated directly from your cell phone's photo gallery and sent to your email

Cursa app on the ebook screen, the video course screen and the course exercises screen, plus the course completion certificate