ONGOING · AI ENABLEMENT
Building AI-Enabled Marketing Workflows
An ongoing umbrella of practical AI workflows for marketing work, each with a defined status, connected data, and human review. Not one application, and not a claim that every experiment is production.
- Capability
- Marketing AI enablement & workflow design
- Lifecycle status
- Ongoing · status varies by workflow
- Role & ownership boundary
- Designs the operating requirements, data connections, human-review steps, and repeatable process for each workflow. This is marketing AI enablement and content operations, with a person in the loop on every output.
- Collaborators
- Marketing team, Internal workflow users, Development / technical support
- Tools
- ChatGPT / custom agents, Obsidian, Litmus, GA4 / Search Console, Python-assisted analysis
Overview
This is an ongoing portfolio of practical AI systems for marketing work: an umbrella, not a single application. Each workflow has its own status. Some are implemented and used regularly; others are partial; exploratory work is kept explicitly separate from production.
Problem
A lot of marketing work is repetitive and information-heavy: pulling together competitive research, analyzing content and internal links, checking a site for issues, assembling reports, producing coded emails. Done by hand, it is slow and inconsistent, and the method lives in one person’s head instead of in a repeatable process.
My role
For each workflow I identify the repetitive or information-heavy task, assess whether AI actually fits, design the operating requirements and workflow, connect the data and tools, test the outputs, add a human-review step, and document a reusable process.
People stay in review for factual claims, customer-facing content, recommendations, sending, and any live-system change. The work is the operating model and the enablement around it, not application engineering.
Approach
Identify repetitive or information-heavy work → assess AI fit → design the workflow → connect data and tools → test outputs → add human review → document a reusable process. State the status of each workflow honestly and keep experiments labeled as experiments.
What I built and did
Implemented and actively used: competitive research, SEO and content analysis, internal-link analysis, website QA, sales intelligence, reporting, email production, reusable agents and skills, and agent orchestration.
Mixed status: documentation, social workflows, and API-connected workflows, with status stated per workflow.
Separate experimental work: prototypes and exploratory workflows remain labeled as experiments and are not presented as production systems.
Selected artifacts
Artifact to add: a sanitized workflow map (input, system, review, output), plus two or three concrete workflow examples underneath it, e.g. competitive intelligence, SEO and content analysis, and email production.Outcome
Several workflows are implemented and in regular use, including competitive research, SEO and content analysis, internal-link analysis, website QA, reporting, and email production, each with its data connected and a human-review step built in. The result is a repeatable operating model that other people can run, with the method documented rather than held in one person’s head. Measuring time and quality impact per workflow is a current focus.
Tools
ChatGPT and custom agents, Obsidian, Litmus, GA4 and Search Console, Python-assisted analysis. Salesforce Account Engagement is a destination for email production, not the identity of this work.
Reflection / key takeaway
The value is not “AI did it.” The value is a workflow other people can run, with the data connected, the review step built in, and the method written down.