Thirteen percent, and the strategy behind the workflow

My portfolio records a 13% improvement in email opens, clicks, and click-through performance over two years, with performance reported as 7% above the industry average.

That is an engagement result from my work, not proof that a particular model or automation caused the entire improvement. The public summary does not supply the underlying campaign-level data or benchmark definition, so I do not turn it into a claim of percentage-point lift, revenue growth, or a controlled experiment.

What it does illustrate is why marketing judgment belongs inside an AI implementation, not at the end of it.

The audience is not an input field

The useful question is not only how quickly we can produce an email. It is why this person should receive it, what they need from it, and what action makes sense next.

A dealer, a designer, and an end user can encounter the same product with different questions. My preference is to build the audience and relationship into the workflow before generating another variation of the headline.

Automation is valuable when it preserves that context. It is less valuable when it helps distribute an irrelevant message more efficiently.

Use AI to expand the options, not outsource the decision

AI-assisted development and content work can support iteration: alternate framing, different levels of detail, segment-specific drafts, and reusable components.

The strategic choices still need an owner. What is the offer? What evidence supports the claim? Does the call to action fit the relationship? Is there a reason to send this now?

My role spans that decision and the system carrying it out. The point is not to protect manual work. It is to spend human judgment where it changes the outcome.

Brand consistency should not require a queue

My more recent sales-enablement newsletter-builder project applies that principle to sales enablement. Marketing-approved content and rep-level customization belong in one usable application, with delivery and reporting connected to it.

The application gives reps room to act without making the brand team approve the same structural decisions repeatedly. It also makes the implementation more than a writing assistant: content, interface, permissions, and the operating process have to fit together.

That project is documented separately from the earlier email-engagement result. I do not combine them into one causal story merely because both involve email.

Choose the measurement before the victory lap

For future work, I want the audience definition, baseline period, comparison window, and metric calculation decided before assessing the result. I also want to know what changed besides the automation.

Opens, clicks, qualified inquiries, and revenue answer different questions. A system can improve one without establishing the others. Reporting should keep those differences visible.

The commercially useful version of AI is not just a cheaper draft. It is a better relationship between the strategy, the workflow, and the customer experience.

Revision note: Updated September 20, 2026 to clarify the scope of the owner-reported engagement figures and connect the strategic lesson to current implementation work. The original June 25 publication date is preserved. No new campaign results are claimed.