Controlled use. Human accountability. Maintained competence.

Why I use it

Choosing not to use AI is no longer a neutral decision. Large firms already build these tools into much of their work. For an independent consultant, using them does not put me ahead of anyone. It puts me on level ground. Declining them would mean asking you to pay more and wait longer for the same result.

AI can draft, summarize, organize, model, and cross-check faster than I can. In some areas it can do things I could not do on my own, and I won't pretend otherwise.

What I bring that the tool does not

The tool carries no responsibility for the result, and it sounds equally confident whether it is right or wrong. I bring more than 20 years in energy infrastructure, PMO design, project controls, and executive leadership. I have not lived your exact situation, but I have worked through the same kinds of problems in other organizations: schedules that slip, reporting that hides trouble until it is expensive, and growth that outpaces the processes meant to support it. That experience is what lets me ask the right questions about your organization, test what the tool produces against how projects actually behave, and decide what reaches you.

The standard I hold myself to

I borrow it from the fields that have lived with automation the longest: aviation, surgery, and power system operations. In each of them, automation assists the professional. It does not replace the professional's competence or responsibility. My companion piece, Keep the Ability to Do, covers how those fields put that into practice.

For my work, that means two things. In the core of my discipline, I keep the hands-on ability to do the work without the tool. Where AI extends me past what I could do alone, I make sure I understand the work well enough to direct it, test it, and recognize when it is wrong, and I bring in a qualified person when the stakes call for it.

My commitments to you

1. I am accountable for everything I deliver. Nothing goes to a client that I have not reviewed and would defend. There is no "the system said so."
2. You choose how your work is done. Every engagement is offered two ways, and the choice is written into the engagement agreement. The default is hand-built: AI is used only if you specifically select it.
   • Hand-built: performed entirely by me, with no AI tools applied to your materials.
   • AI-assisted: led and finalized by me, with AI used to speed up the work. It is priced lower than hand-built because it takes less time, and you should not pay a hand-built rate for accelerated work. Standard AI-assisted rates are listed on this site; for larger or custom engagements, the difference is built into the statement of work and pricing.
3. The quality guarantee is the same either way. Both options are held to the same standard. Neither is a lesser product.
4. I check myself. Each year, a random sample of AI-assisted engagements also receives a full parallel review done without AI, and I compare the two. Selection is random and not disclosed in advance, so every AI-assisted engagement is held to that standard.
5. I verify rather than accept. AI output is a draft to be tested against your facts, not an answer to be passed along.
6. I am transparent. If you want to know which parts of your work were AI-assisted, ask and I will tell you.
7. Your information is not used to train AI. I use Claude on a Team (business) subscription whose commercial terms exclude client content from model training, and I have turned off the feedback feature that would otherwise share conversations with the provider. At the close of each engagement I delete your material from the AI service, and the provider removes it from its systems within 30 days. Client data is handled under the Vibrant Blue Technology and Data Security Policy.
8. Your email can stay out of AI entirely. Like most businesses, I use AI to keep my own inbox and calendar organized: sorting, reminders, and scheduling. That is not work on your material. If you would rather your email never pass through AI tools, check the box in the agreement. Your messages are routed to a folder those tools never read, and AI scheduling tools use your meetings only to note their date and time.
9. Judgment stays with me. The tool can draft and calculate. Recommendations, risk calls, and what goes in front of your leadership are mine.

Why I hold this line

Dependence builds one reasonable shortcut at a time, until a professional can no longer evaluate what they are approving. That is a bad place for a consultant to be, and a worse place for a client.

Controlled use, human accountability, and maintained competence: that is the standard I hold myself to.