AI Use-Case Prioritization 2x2
Rank AI ideas by customer impact and team-wide productivity potential
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 7
- Confidence
- 98%
The framework evaluates AI use cases on two axes: their potential impact on demand generation or brand awareness, and their breadth of use across the internal team. Ideas are gathered through a lightweight intake process, roughly graded as high, medium, or low on each axis, and plotted to identify the candidates with the strongest combined value. Leaders then stack-rank the top candidates and assign technical resources to execute them. The process deliberately favors speed over false precision, using informed estimates rather than lengthy business cases. A recurring review every two weeks allows the portfolio to respond quickly when new technology, business needs, or macro conditions change.
Origin
Developed at HubSpot by the marketing and marketing-technology teams to prioritize more than 100 proposed AI experiments. Shared on Marketing Against the Grain.
Core principles
- 01Prioritize impact rather than novelty
- 02Use rough estimates when speed matters more than precision
- 03Separate external growth impact from internal productivity gains
- 04Assign dedicated technical resources to the highest-ranked ideas
- 05Revisit priorities as technology and market conditions change
How to run it
- 1
Open a Lightweight Intake
Invite the full team to submit AI use cases through a simple channel such as a form. Optimize for participation and speed rather than a perfect submission system.
Pro tip Ask submitters to state the expected user, outcome, and affected workflow.
Watch out A complicated intake process can suppress useful ideas before evaluation begins.
- 2
Estimate External Impact
Grade each idea's potential effect on demand generation or brand awareness as high, medium, or low. Use directional judgment when reliable forecasts do not yet exist.
Pro tip Define what meaningful impact means before scoring begins.
Watch out Do not treat speculative precision as evidence.
- 3
Estimate Breadth of Use
Grade how much of the internal team could use or benefit from the capability. This axis represents the potential productivity or efficiency gain.
Pro tip Count realistic recurring users rather than everyone who could theoretically access it.
Watch out A broadly available tool may still have little practical adoption.
- 4
Plot and Shortlist
Place the ideas on the 2x2 and identify those that score strongly across both axes. Reduce the full intake to a manageable shortlist for leadership review.
Pro tip Use rough grading to structure the backlog quickly.
Watch out Do not automatically discard a narrow use case with exceptional revenue potential.
- 5
Stack-Rank the Candidates
Review the shortlisted ideas and establish an explicit order of operations. Consider strategic context alongside their matrix positions.
Pro tip Keep the final decision group small enough to move quickly.
Watch out A large committee can turn prioritization into prolonged consensus-building.
- 6
Resource the Winners
Assign dedicated technical and domain experts to the highest-ranked experiments. Ensure the team can execute rather than merely maintain a prioritized list.
Pro tip Pair an AI specialist with a workflow subject-matter expert.
Watch out Prioritization without dedicated capacity produces no learning.
- 7
Revisit the Portfolio
Reconvene regularly to account for new technology, changing conditions, and experiment results. Adjust the ranking when the evidence changes.
Pro tip Use a two-week review cadence during fast-moving adoption periods.
Watch out Do not let the original ranking become permanent policy.
In the wild
HubSpot invited every marketer to submit AI ideas through Slack and a Google form, receiving more than 100 proposals. A centralized marketing-technology team estimated each idea's top-of-funnel impact and breadth of internal use, plotted the strongest 10 to 15 candidates, and reviewed the ranking with marketing leadership before assigning resources.
→ The process elevated high-potential email and website-chat experiments and gave them dedicated execution capacity.
Common mistakes
Perfecting the Intake Process
Building an elaborate submission and scoring system delays the experiments that are supposed to create the evidence. Start with simple collection and rough estimates.
Ranking Without Resourcing
A prioritized backlog has little value if nobody has dedicated time or technical capability to execute its leading ideas.
Freezing the Initial Ranking
AI capabilities and business conditions change quickly, so the portfolio must be reviewed and reordered regularly.
Is it for you?
Best for
It is best for teams with many possible AI projects but limited people, time, and technical capacity.
Not ideal for
It is not ideal when a team has only one mandatory use case or must prioritize primarily by compliance and safety risk.
From the transcript
“we put together a 2 by two that really plotted what we felt were two very important criteria to measure these use cases against”
“on the Y we had the impact that a use case might have on top of funnel metrics”
“on the X we had breadth of use meaning what size of the marketing team could really leverage this use case”
From the episode
The AI Strategy That Doubled His Email Conversion Rate (Step By Step)