AI workflow audit
The map of every place AI genuinely belongs in your marketing operation today, scored by impact and feasibility. You leave the first two weeks with a prioritized backlog, not a list of ideas.
07 · AI Enablement
AI your team actually uses.
Most marketing teams have Claude accounts and no workflow around them. We come in, find the jobs where AI genuinely belongs, and ship the AirOps workflows, custom Claude skills, project files and automation that make it operational. The deliverable is working AI inside your stack, not a slide deck about it.
The problem
Your team has Claude. Your team has AirOps. Your team has no workflow.
Buying the seats was the easy part. Three months in, two people on your team use Claude every day, three log in once a week, and the rest forgot the password. AirOps has one workflow that runs and four that were built and abandoned. Nobody knows what good looks like.
This is what happens when AI gets rolled out as a tool, not as a workflow. Tools become useful when they are wired into how work actually moves through your team: the brief request that triggers a Claude skill, the on-brand draft that comes back from an AirOps workflow, the cron job that catches the metric drop before anyone notices.
The work of AI enablement is the boring part: writing the project files that tell Claude how your brand sounds, building the skills that handle the repeatable jobs, configuring the AirOps brand kit and knowledge base so output stays on voice, and rolling the new habits out so the new way of working survives the first hard week.
We do that work alongside your team. By the end of the engagement, you have shipped AI artifacts your team uses every week, a documented playbook for the next set of builds, and a measurement framework that shows the time and cost the new workflows actually save.
Symptoms we see
- Claude usage concentrated in two people who figured it out themselves.
- AirOps workflows built once, abandoned, never measured.
- Drafts that need so much editing the team gives up and writes from scratch.
- Brand voice drift because there is no project file telling Claude what good sounds like.
- Nobody can answer "how much time has AI saved us this quarter?" in numbers.
What's included
Six deliverables, all operational.
Every engagement ships the same six artifacts. They are not slides. They are configured tools, written project files, and live workflows your team uses on day one of the next month.
AirOps implementation
Brand kit set up against your actual brand. Knowledge base built from your source-of-truth. Workflows wired to produce on-brand content at the scale your team is actually trying to hit.
Custom Claude skills
Packaged skills installed in your Claude account that handle the repeatable jobs: content briefs, sales-call summaries, weekly status drafts, blog refresh specs. Your team triggers them by name.
Project files (CLAUDE.md)
The source-of-truth file that tells Claude how your brand sounds, what conventions to follow, which files matter and what to never do. Committed to your repo so it travels with the work.
Workflow automation
The detect-suggest-notify loops that run without a human in the loop: traffic anomaly alerts, content refresh queues, citation-share trackers, brief-ready notifications. Cron, hooks, integrations.
Team rollout & embedding
Working sessions with each role on your team, a documented playbook for the new way of doing the job, and a measurement framework that shows time saved, output increased and quality held.
Why this works
Built by someone who has done the work.
This is not a freshly-minted AI consultancy. We have built and shipped this stack inside our own operation and inside client engagements. The credentials are real and the artifacts are running in production.
AirOps Distinguished Content Engineer
Recognised by AirOps for shipping production content workflows. Active in the AirOps Builders Slack and Champions Hub.
Production agents in our own stack
We run our own internal Claude agents, a Foundation Audit autonomous agent on Haiku, a live WebMCP tool on this site, and a visitor chat agent on Cloudflare. We build before we recommend.
Chrome Built-in AI Early Preview Program
Enrolled in Google's early-preview track for WebMCP, multimodal Prompt API, Proofreader and hybrid AI. We see the next generation of browser-native AI before it ships.
10+ years in B2B SaaS marketing
A marketer who learned to build AI inside marketing. The workflows reflect what the day-to-day actually looks like.
How it runs
Six to eight weeks, in working increments.
Engagement is fixed-scope with weekly working sessions. Every increment ships something your team can use that week, not a deliverable at the end.
Workflow audit
Working sessions with each role on your marketing team to find where AI genuinely belongs in their week. We score every candidate by impact, feasibility and reliability. You leave with a written audit and a prioritized backlog.
AirOps foundation
Brand kit, knowledge base and at least one production workflow stood up against your real brand and content needs. If you are new to AirOps we handle procurement and standup. If you are on AirOps, we accelerate or replace what is already running.
Claude skills & project files
The custom skills that handle the repeatable jobs, installed in your Claude account. The CLAUDE.md project file committed to your repo so every conversation starts with the right context. The conventions documented so the next engineer can extend them.
Automation builds
The detect-suggest-notify loops that run without humans: anomaly alerts, refresh queues, content trackers. Built where they belong, in AirOps, in cron, in hooks, in Slack. Tested against your real data, not a demo set.
Rollout & measurement
Working sessions with each role to embed the new workflows in their week. A documented playbook so the new way survives turnover. A measurement framework that shows time saved, output increased and quality held, with a dashboard you can read in five minutes.
Who it's for
Built for teams ready to make the tools work.
- Your marketing team has Claude or AirOps access and limited workflow around either.
- You want AI to ship inside repeatable jobs, not bolted on as a draft-helper.
- You measure success in time saved and output shipped, not in tool adoption.
- Your team can absorb new habits if the workflow is genuinely better than the old one.
- You want operational AI you can audit, not a black box you have to trust.
Pairs well with
What sits around it.
AI enablement amplifies the rest. Strategy decides what the workflows should produce. Production runs through the workflows you build. Visibility measures whether they are working.
Frequently asked
Questions we hear.
What is AI enablement for a marketing team? +
AI enablement is the work of taking AI tools your team already has access to, like Claude, ChatGPT and AirOps, and turning them into a working part of your marketing operation. It covers the audit of where AI genuinely belongs, the build of custom skills and workflows that fit how your team actually works, and the rollout that makes the new habit stick past week three.
Do I need to be on AirOps to work with you? +
No. AirOps is one of the tools we build inside, and we are a recognized AirOps Champion and Distinguished Content Engineer, but the broader engagement covers Claude, the wider AI stack, and the workflows that connect them. If you do not have AirOps and we recommend it, we help you procure and stand it up. If you do, we accelerate what you are already running.
What is a Claude skill or project file? +
A Claude skill is a packaged set of instructions and tools that turns Claude into a specialist for one job, for example a content brief generator or a sales-call summarizer. A project file, CLAUDE.md, is the source of truth that tells Claude how your team works, what conventions to follow, and what files matter, so that every conversation starts with the right context instead of from zero.
How long does an AI enablement engagement take? +
A full first-build engagement runs six to eight weeks. The first two weeks are workflow audit and prioritization. The next four are the actual builds, AirOps workflows, Claude skills, project files and automation, shipped in working increments. The last weeks are team rollout and embedding. After that, most teams move into a lighter monthly support arrangement.
How is this different from hiring a Claude consultant? +
Most Claude consultants stop at prompts and training. This engagement ends with shipped artifacts: live AirOps workflows your team uses every week, custom skills installed in your Claude account, project files committed to your repo, and automation that runs on its own. The deliverable is operational AI, not slide-deck strategy.
Next step
Let's see where AI belongs.
A 30-minute discovery call, no pitch, no pressure. We look at where your team uses AI today, where it should, and whether an enablement engagement makes sense for the quarter ahead.
Book a discovery call →Booking goes straight to Gemma's calendar.