Pilot Scope
Before filling this out, complete the Quick Check and Situation Map. They ensure you've identified your pain point, mapped your workflows, and confirmed you have the foundation to make this actionable.
Why a Scoped Pilot
A well-scoped pilot is the most important thing a business owner needs to start their applied AI journey. It's not a proposal. It's not a wishlist. It's a concrete, bounded experiment that proves (or disproves) whether AI can deliver real value for a specific part of your business.
Without a scoped pilot, you're either paralyzed ("where do we even start?") or reckless ("let's just try AI on everything"). Both waste time and money. A pilot gives you a focused first step with clear success criteria so you know whether to keep going.
The Honesty Filter
Before you scope a pilot around a specific problem, pressure-test it:
- Is the data clean enough? If the information you need is scattered across inboxes and someone's memory, you have a data problem to solve first.
- Is the process stable? If the workflow changes every month, automating it will create a maintenance burden, not savings.
- What's the cost of AI errors vs human errors? Some domains (medical, financial, legal) have asymmetric downside. Factor that in.
- Do you have the infrastructure to deploy this? Can your team actually run and maintain the result?
Most problems will fail this filter. That's fine. You want the ones that pass.
This document is the output of that scoping work. Once complete, it does three things:
- Gives you clarity. You'll know exactly what you're testing, what success looks like, and what resources it requires.
- Makes you ready for a practitioner. Any skilled practitioner can read this and immediately assess whether they can help and how.
- Sets shared expectations. Both sides have a common reference point from day one. No ambiguity about scope, timeline, or what "done" means.
Who fills it out: The person leading AI adoption inside your company, ideally with input from the operational team closest to the problem. Best completed with a practitioner guiding the conversation.
Who reads it: The practitioner who will execute the pilot. Also useful for getting internal buy-in from leadership.
Length target: 1-3 pages. Keep answers concise. Brief bullet points are preferred over paragraphs.
1. Company Overview
Company name:
Industry:
Size (employees / revenue range):
What your company does (2-3 sentences):
The person leading this:
- Name and title:
- Role in the company:
- Decision-making authority (can you greenlight this, or does someone else approve?):
2. The Problem You're Solving
You identified the area with the most pain and mapped the current workflow in your Situation Map. Now be specific about what the pilot will address.
What is the specific problem this pilot will solve? (Not "we want to use AI." What's not working, what's too slow, or what's costing you the most?)
What does this problem cost you today? (Time per week, dollars lost, errors made, delays caused. Even rough estimates help.)
Why this problem first? (What makes this the right starting point vs. other problems you have?)
3. What Success Looks Like
What would a successful pilot look like in 30 days?
What metric will you use to measure success? (Time saved, error rate reduced, revenue increased, cost eliminated, throughput improved, etc. Pick one primary metric. If you can't measure it, you can't prove it worked. And you need proof to get budget for project two.)
What's your baseline today? (How does the current process perform on that metric? If you don't know, that's worth noting. You need a starting point to prove improvement.)
Who benefits most from solving this? (Specific roles, teams, or customers)
4. Technical Environment
Summarize the key details from your Situation Map. Keep it brief.
What systems and tools are involved? (CRM, ERP, spreadsheets, email, proprietary software, etc.)
Can data be exported from these systems? (APIs, CSV exports, or is it locked in a vendor platform?)
Are there compliance constraints? (HIPAA, SOC 2, GDPR, financial regulations, etc.)
5. People and Ownership
Who owns this pilot day to day? (The person who will manage the AI-assisted workflow, troubleshoot issues, and report on results)
Who needs to approve this? (If it's not you, who is the decision maker?)
Who will be most affected if the workflow changes? Any concerns about adoption or resistance?
How much time can your team realistically spend on this per week? (Be honest. If the answer is "very little," that shapes what's feasible.)
6. Constraints and Budget
Budget range for this pilot: (Under $5K, $5-15K, $15-50K, $50K+)
Timeline: When do you want the pilot running? A good pilot ships in 8 to 12 weeks. If it can't be executed in that window, the scope is too broad. Narrow it.
Deal-breakers: Any constraints the practitioner must know up front? (No cloud storage, must be self-hosted, can't change the core workflow, etc.)
7. After the Pilot
If the pilot succeeds, what happens next? (Scale it across the company? Expand to other workflows? Build a longer-term engagement?)
Are you open to this project being referenced publicly? (As a case study, at a community event, in a practitioner's portfolio, etc.)
What Happens Next
Once your Pilot Scope is complete:
- Review with a practitioner. Walk through it together. They'll pressure-test your assumptions and sharpen the scope.
- Execute the pilot. A focused 2-4 week sprint with clear deliverables and a defined success metric.
- Evaluate and decide. Did it work? If yes, scope the next phase. If not, you learned something valuable for a fraction of what a full engagement would have cost.
The pilot is the proof. Everything else follows from it.
If you want hands-on help building your Pilot Scope, the Applied AI Society runs workshops where practitioners guide you through the entire process. See upcoming events → or reach out at appliedaisociety.org.
Maintained by the Applied AI Society. This template evolves based on real practitioner and business owner experience.