The Problem: Chasing "Cool AI"
Last week I spent 90 minutes with a group of SMB C-Level executives wrestling with the same question: "How do we pick AI use cases that actually move the needle?" The conversation reminded me of the early SaaS days - everyone chasing the shiny object, few asking the hard questions first.
Most teams do this backwards. They see a cool AI demo, get excited, and start trying to find problems it can solve. That's like buying a Ferrari and then figuring out you need milk from the corner store.
Based on surveying these leaders, the two biggest barriers to AI adoption in their organizations were:
- Lack of AI expertise - teams don't know how to evaluate tools, write effective prompts, or measure results
- Resistance to change - employees fear displacement or simply default to familiar workflows
Both are solvable. But not by buying more tools.
What Is the 6 Ps Framework for AI Use Case Selection?
The 6 Ps is a scoring system designed to evaluate AI use cases before a single dollar is spent. It forces teams to define success in business terms - not technology terms - before committing to a project.
What is the specific blocker you're trying to remove? Is it tied to a KPI you already track? Name it. Vague problems produce vague outcomes.
List at least three ways to solve the problem - including non-AI options. Spoiler: AI is not the right answer to every problem. Process improvements often beat AI pilots.
Define success in business metrics only: revenue lift, cost reduction, cycle time improvement, or risk reduction. If you can't quantify it, you can't prove it worked.
Be honest about the level of effort required: data quality, budget, people's time, and your team's bandwidth. Most projects underestimate this.
Given your actual constraints - not best-case scenario - what are the realistic odds this succeeds? Score it honestly. Low-probability projects shouldn't get priority funding.
Run the math: weighted business impact × probability ÷ effort. This produces a ranked list of your use cases. Start with high-payoff, high-probability, low-effort pilots.
Should You Start with Optimization or Transformation?
The counterintuitive answer: always start with optimization.
Transformative AI - reinventing your business model, building new products with AI at the core - is compelling. But it carries enormous execution risk, requires significant data infrastructure, and is the wrong starting point for most organizations.
Companies like Salesforce, Klarna, and Dow are seeing 30–70% efficiency gains by focusing first on concrete, measurable improvements to existing workflows. They're not trying to reinvent their business on day one. They're making what they already do faster, cheaper, and more consistent.
Quick wins in existing processes build something that transformation projects rarely do: organizational trust. When employees see AI delivering real results in Week 3 of a pilot, resistance drops and adoption accelerates.
How to Think About AI: Three Personas That Unlock Productivity
One of the most useful mental models I've developed - and one that cuts through employee fear about AI - is framing AI through three distinct personas depending on the task:
The Assistant
Automates repetitive, time-consuming administrative tasks that drain energy and focus. Data entry, scheduling, summarization, email drafts, report generation.
The Strategist
Challenges your assumptions and surfaces alternative perspectives you wouldn't have considered. Red-teaming decisions, stress-testing plans, generating counterarguments.
The Creator
Generates first drafts - not finished work, but a strong starting point that beats staring at a blank page. Proposals, presentations, SOPs, training materials, content.
Framing AI this way during training and change management reduces fear. Employees stop hearing "AI will replace me" and start hearing "AI handles the grind so I can do higher-value work."
How to Run Your First AI Pilot in 30 Days
If I were dropping into your team next week, this is exactly what I'd do:
- Deep discovery interviews with employees across roles to map real workflows and surface high-effort, repetitive tasks
- Pick 3 real problems tied to hard metrics - not cool demos, actual KPIs
- Baseline the "before" - document current time, cost, and error rates so you can prove impact
- Score with the 6 Ps and select the highest-probability pilot you can ship in 30 days
- Instrument it - measure everything, publish the results internally
- Socialize the win, then pick the next use case and repeat
"The companies winning with AI aren't the ones with the best technology or the biggest budgets. They're the ones that pick the right problems and actually drive adoption across their teams. Culture is usually the bottleneck - not the tools."
Frequently Asked Questions
Most AI projects fail because companies choose tools before defining problems. They get excited by demos, sign contracts, and then try to find a use case - backwards from how it should work. Without clear success metrics defined upfront in business terms, there's no way to prove the project worked or justify continued investment.
The 6 Ps Framework is a scoring system: Problem (what KPI is being blocked?), Possibilities (list 3+ solutions including non-AI), Payoff (business metrics only: revenue, cost, cycle time, risk), Perspiration (honest effort assessment), Probability (realistic odds given current constraints), and Prioritization (weighted score of the above). It produces a ranked list of AI use cases to tackle in order.
Based on direct research with SMB leaders, the two top barriers are: (1) lack of AI expertise - teams don't know how to evaluate tools, prompt effectively, or measure ROI; and (2) resistance to change - fear of job displacement or attachment to existing workflows. Both require leadership modeling, AI fundamentals training, and a clear communication strategy, not just tool purchases.
Always start with optimization - improving existing workflows - before pursuing transformative innovation. Companies like Salesforce, Klarna, and Dow are achieving 30–70% efficiency gains by focusing on concrete, measurable process improvements first. Quick wins build organizational trust and reduce resistance, making larger transformation efforts far more likely to succeed.
Target 30 days for your first instrumented pilot. Select a high-probability, low-effort use case with a clear baseline metric. Measure before and after. Publish results internally. The goal isn't perfection - it's a visible win that builds momentum and organizational confidence in AI as a tool, not a threat.
Ready to Pick an AI Use Case That Actually Delivers?
A focused 30-minute working session is often all it takes to apply the 6 Ps to your backlog and identify a pilot you can ship this month.
Schedule a Free 30-Minute SessionNo commitment. No pitch. Just a focused conversation about where AI can deliver real ROI in your business.