01 / PROCESS IMPROVEMENT & AI ADOPTION
Give the team a complete brief before the work begins.
Incomplete requests were costing days of clarification. I built a Claude skill that helps requesters work through the requirements before submitting a brief.
Previous back-and-forth2–4 days
Guided brief creation~30 minutes
I observed typical project turnaround improve from about two weeks to one, giving designers more time to produce the work and develop their craft.
Approximate timings based on my experience across campaigns, event brand kits, web design, logo assets, static ads, and video.
Read the case study
The problem
Requests often lacked asset counts, aspect ratios, delivery channels, business objectives, deadlines, or dependencies. Designers and requesters had to resolve those gaps before production could move forward.
What I built
I designed a reusable brief generator in Claude that requires requesters to answer the questions needed for a complete handoff. It helps people organize their thoughts and makes missing requirements visible earlier.
How I introduced it
After discussing process improvements with my CMO, I introduced the skill at a marketing town hall and rolled it into our marketing help channel. I also saw colleagues begin meeting with stakeholders and building skills for other problems.
What changed
More complete briefs reduced clarification cycles. Stakeholders received work sooner, and the creative team gained room for additional projects and craft development. The time estimates describe brief completion and overall delivery separately.
Intake designStakeholder alignmentAI adoption
02 / RESEARCH & RESOURCE PLANNING
Three industries. Five features. One production week.
A campaign needed plumbing, electrical, and HVAC companies that each used five specific features. I researched the right candidates and concentrated filming within one region.
Three shoots completed in one week for approximately $12,000 in shoot expenses. All campaign videos delivered.
How the research shaped production
My contribution
I used company data through Omni and Snowflake connected to Claude to identify companies that fit the feature requirements, industry, company size, and location. I organized candidate information into sheets and coordinated the budget and filming schedules.
The decision that mattered
Geography was part of the research criteria from the start. Finding qualified companies close together made a single production week possible and reduced repeated travel.
Budget scope
The approximately $12,000 covered production rentals, contracted help, travel, lodging, and per diem. I recall previous outsourced shoots costing substantially more; the figure here reports this project’s shoot expenses, without treating editing or internal labor as included.
ResearchBudget ownershipDelivery
03 / AI WORKFLOWS & CONTENT OPERATIONS
Find the customer evidence already in the footage.
Editors needed interviews that supported three features in a Housecall Pro bundle campaign. The search covered roughly five companies in each of 12 industries.
Relevant material surfaced in minutes, compared with an estimated 30 minutes of manual transcript searching per interview.
Inside the workflow
The problem
Finding the right evidence meant listening to 45–60-minute interviews or searching transcripts manually. The team also needed to know whether the available material could support a particular industry’s video.
What I built
I built a Claude workflow to review VTT transcripts and produce a feasibility report with relevant quotes, feature references, and possible video directions. Instructions required preserving the customer’s wording, while timestamps made the material traceable to the interview.
From research to production
I incorporated Daydream into the process to help assemble video from a brief and let editors watch the source clips. Editors were the primary users. The same approach could also support testimonial research for other teams.
What the estimate means
The comparison describes locating relevant material in an individual interview. It is not a measurement of final editing time or a calculated total across the archive.
Information retrievalWorkflow designEditor enablement
04 / COACHING & PEOPLE DEVELOPMENT
Help a designer gain a voice in the decisions.
A designer was receiving direction secondhand. I advocated for his inclusion in stakeholder meetings and coached him to prepare questions and communicate his solutions clearly.
He took greater ownership, made his work more visible, and was promoted within months.
The coaching behind the change
Access and preparation
He was the sole designer working on the brand refresh. Being in the meetings gave him direct access to the context behind requests. In our 1:1s, I encouraged a regular set of questions and coached clear communication.
What I observed
He began proposing and explaining solutions, joining relevant conversations, and reporting his work in shared channels. Greater visibility brought more opportunities to contribute.
My role
My contribution was advocacy and coaching. The design work, growing ownership, and promotion were his achievements. This is the kind of development I enjoy supporting.
CoachingCommunicationOwnership
05 / COLLABORATION & EXPERIMENTATION
Let customer testimony lead the ad.
Working as creative project manager alongside a Senior Manager of Performance Marketing, I helped deliver geo-test campaigns that evolved from voiceover-led scripts to testimonial-led ads.
The team observed improved click efficiency, with cost per click a primary signal.
How the work evolved
The observation
We saw stronger engagement when customer testimony entered the ad. The team brought testimonials earlier in the script, then tested ads built entirely around firsthand customer accounts.
My contribution
I partnered on the campaigns as the creative project manager. Performance analysis and testing decisions were collaborative with the performance marketing lead.
The result
I recall lower cost per click, with click-through rate and CPM also part of the discussion. The improvement is unquantified here; these measures do not establish a conversion-rate lift.
Project managementTestingPerformance partnership
