AI Lab

Less repetitive work. More intelligent systems.

Practical AI for marketing operations. Every workflow on this page states what it does, what a person still reviews, and whether it is live, in use, offered as a service or still a prototype.

Discuss an AI Project AI automation service

Workflows

What exists today.

Four workflows, each with the same shape: an input, research, an AI draft, a human review and an output.

Live tool

SEO research and audit workflow

Problem: Site audits take hours and the findings arrive as an unprioritised list.
Workflow: Crawl a URL → check on-page signals, metadata and Core Web Vitals → rank fixes by likely impact.
Human review: A person confirms each recommendation before it reaches a client or a developer.

Open the tool

In use on this site

AI-assisted content operations

Problem: Publishing consistently needs research, briefs and drafts, every week.
Workflow: Topic queue from real search queries → structured draft → automatic checks for invented numbers, copied text and broken links → saved as a draft.
Human review: Nothing is published automatically by default. Every draft is read, edited and approved by a person.

Offered as a service

Automated marketing reporting

Problem: Monthly reports are copied and pasted from five dashboards.
Workflow: GA4, ad platforms and CRM data → one dashboard → a written summary of what changed.
Human review: Interpretation and next actions are written or approved by a person.

Prototype

Lead research and qualification

Problem: Sales teams spend time on enquiries and accounts that will never buy.
Workflow: Target account or enquiry → company research and enrichment → fit score against agreed criteria → routed to the right person.
Human review: Scoring rules are set and reviewed by the sales team; no one is contacted automatically.

Ground rules

AI with human judgement.

The same rules apply to the tools used on this site and to every workflow built for a client.

A person owns every output

Each workflow names who checks the result before it reaches a customer, a client or a live website.

No invented facts

Drafts are checked for numbers, claims and quotes that cannot be traced to a source. Anything unverifiable is removed.

Your data stays scoped

Workflows use the minimum data they need, in accounts you control, with AI providers agreed against your data policy before anything is connected.

Measured like any project

Before building, we agree what the workflow should save or improve, and we check it afterwards.

Fit

Where automation helps, and where not.

Good candidates

Repetitive research: competitors, keywords, accounts
First drafts of briefs, outlines and reports
Moving data between GA4, ad platforms, sheets and CRM
Routing and enriching inbound enquiries
Monitoring: rankings, broken pages, budget pacing

Kept with people

Final decisions on strategy, pricing or positioning
Publishing claims without a human check
Contacting people automatically without consent
Replacing expertise your buyers expect to talk to

Discuss an AI Project Read about AI in marketing