AI Customer-Value Two-by-Two
Prioritize AI where customers gain value and humans dislike the work
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 7
- Confidence
- 99%
The AI Customer-Value Two-by-Two plots candidate work on two axes: whether humans love or hate doing it, and whether it creates low or high value for the customer. The recommended starting quadrant is work that customers value highly but humans dislike, find tedious, or are poorly equipped to perform consistently. Examples include complex personalization, continuous prediction, or real-time pattern detection. Automating this quadrant can improve the customer experience while freeing people for work they find more meaningful. Leaders should be cautious when AI moves into high-value work that people love, because resistance, identity loss, and differentiation risks increase. The matrix is a prioritization device rather than a complete governance model, so selected projects must still pass safety, privacy, quality, and strategic-differentiation reviews.
Origin
Kip Bodner proposed the matrix to prioritize practical AI investments, and Kieran Flanagan endorsed the high-value, human-hated quadrant. Extracted from Marketing Against The Grain.
Core principles
- 01Customer value and human desirability are separate evaluation axes.
- 02The strongest starting quadrant combines high customer value with work humans dislike.
- 03AI should initially complement people by removing undesirable, difficult, or computer-suited work.
- 04Automating loved, high-value work creates greater resistance and strategic risk.
- 05Multiple small improvements can compound into a superior customer experience.
How to run it
- 1
Inventory candidate work
List specific tasks, decisions, and customer interactions that could be supported or performed by AI.
Pro tip Keep each item narrow enough to score independently.
Watch out Do not score an entire department as one task.
- 2
Score human desirability
Assess whether the people responsible love, tolerate, or hate performing each item and whether humans are well equipped to do it.
Pro tip Ask the people who perform the work rather than relying only on management assumptions.
Watch out A manager’s view of tedious work may differ from the practitioner’s source of mastery or satisfaction.
- 3
Score customer value
Estimate how strongly each item affects customer outcomes, convenience, trust, or experience.
Pro tip Use customer evidence and operational metrics where possible.
Watch out Internal cost reduction is not automatically customer value.
- 4
Plot the matrix
Place every item into one of four quadrants based on human desirability and customer value.
Pro tip Document borderline scores so teams can challenge assumptions.
Watch out False precision can hide weak evidence behind numerical ratings.
- 5
Prioritize the target quadrant
Start with tasks that humans dislike but that create high customer value, especially where computers can handle complexity more consistently.
Pro tip Look for several complementary tasks that can stack into one better experience.
Watch out Do not select a task solely because employees dislike it if customers gain little.
- 6
Apply strategic gates
Check differentiation, safety, privacy, quality, and human-oversight requirements before implementation.
Pro tip Pilot reversible, bounded use cases first.
Watch out The matrix does not replace governance or risk analysis.
- 7
Measure both sides
After deployment, measure customer value and employee experience to confirm that the task remained in the intended quadrant.
Pro tip Watch for displaced work that is more frustrating than the original task.
Watch out Automation can shift burdens rather than eliminate them.
In the wild
A marketing team plots manual send-time analysis as work employees dislike but customers value because it reduces irrelevant interruptions. It plots campaign concept development as work creatives love and customers also value. The team automates send-time prediction first while retaining human leadership over campaign concepts.
→ The company improves relevance and frees marketers without prematurely automating identity-defining creative work.
Sales managers cannot listen to every live call, while representatives find constant self-monitoring difficult. Customers benefit when representatives adjust pace and explain the right points. The company pilots AI prompts for those narrow signals while leaving relationship judgment and final recommendations to the salesperson.
→ AI handles continuous pattern detection while humans retain high-value interpersonal responsibility.
Common mistakes
Prioritizing low-value busywork
Removing disliked work may save time, but it is not the strongest AI investment when customers receive little benefit.
Automating loved work first
Replacing meaningful, high-value work can create resistance and erode the human strengths customers appreciate.
Ignoring differentiation
A high-value quadrant score does not justify technology that makes the customer experience generic or interchangeable.
Is it for you?
Best for
It is best for leaders selecting initial AI and automation investments across marketing, sales, service, and operations.
Not ideal for
It is not ideal for safety-critical decisions where risk and regulatory analysis must override simple value and preference scores.
From the transcript
“I would like the vertical access to be jobs humans love to do and jobs humans hate to do. And I would like the horizontal…”
“That's the quadrant that I think is the most valuable, most interesting place to start is how do you think about and evaluate technology that…”
“And where we would start going forward is helping AI solve the problems in our businesses that humans either don't want to do or aren't…”
From the episode
The Impact of AI in Marketing (Friend or Foe?)