The brief. The system underneath it.
Five engagements. Every one started the same way — the account opened, the trigger logic read, the leak found and priced — before anything was rebuilt. The line under each one tells you what it can show you.
01From spam to inbox in six weeks.
A 66,000-contact AI companion platform whose lifecycle had stopped responding to what users did. Authentication rebuilt, the list cleaned, onboarding moved from a timer to behaviour.13.3% → 60.2%Companion creation28% → 0%Gmail user-reported spamCustomer.io and Postmaster captures · client anonymisedRead the case study 02 · IndependentFour signals collected at signup. None used in the lifecycle.
Two audits, a decision model, a Customer.io build and three experiments — what I’d do with the data Notion already has. Nobody asked for it.“A chance LinkedIn post from Isaac Olukoya led to a genuinely insightful discussion on customer lifecycles, production adoption, and AI.”Áine Dundas · Head of EMEA Marketing, Notion25-page document · six artifacts · Notion’s own screensRead the research 
03An assessment that decides what happens next.Built from zero for an aviation mental health practice: a branching clinical assessment, four outcome tiers, a lifecycle campaign for each, and an automated Slack escalation on the two that need a human.Typeform · Zapier routing · Customer.io campaign · Zapier escalationTypeform, Zapier and Customer.io captures · client anonymisedRead the case study 04They asked for five campaigns. I stopped the build.
One audience across two systems, and no rule saying which owned which relationship. I drew the boundary first, then built six campaigns instead of five.Customer.io, Beehiiv and Postmaster captures · project record · client reviewRead the case study 
05Leads arrived. Nothing carried them to a purchase.A full lifecycle engine built in GoHighLevel — dual lead magnets by intent stage, quiz-based routing, post-purchase onboarding and an automated path into the membership tier.“His audit alone was worth the investment.”Monica Anyango · Founder, Simplicity & Productivity Co.Client reviewRead the case study
A 66,000-contact AI companion platform whose lifecycle had stopped responding to what users did. Authentication rebuilt, the list cleaned, onboarding moved from a timer to behaviour.13.3% → 60.2%Companion creation28% → 0%Gmail user-reported spamCustomer.io and Postmaster captures · client anonymisedRead the case study 02 · IndependentFour signals collected at signup. None used in the lifecycle.Two audits, a decision model, a Customer.io build and three experiments — what I’d do with the data Notion already has. Nobody asked for it.“A chance LinkedIn post from Isaac Olukoya led to a genuinely insightful discussion on customer lifecycles, production adoption, and AI.”Áine Dundas · Head of EMEA Marketing, Notion25-page document · six artifacts · Notion’s own screensRead the research 03An assessment that decides what happens next.Built from zero for an aviation mental health practice: a branching clinical assessment, four outcome tiers, a lifecycle campaign for each, and an automated Slack escalation on the two that need a human.Typeform · Zapier routing · Customer.io campaign · Zapier escalationTypeform, Zapier and Customer.io captures · client anonymisedRead the case study 04They asked for five campaigns. I stopped the build.One audience across two systems, and no rule saying which owned which relationship. I drew the boundary first, then built six campaigns instead of five.Customer.io, Beehiiv and Postmaster captures · project record · client reviewRead the case study 05Leads arrived. Nothing carried them to a purchase.A full lifecycle engine built in GoHighLevel — dual lead magnets by intent stage, quiz-based routing, post-purchase onboarding and an automated path into the membership tier.“His audit alone was worth the investment.”Monica Anyango · Founder, Simplicity & Productivity Co.Client reviewRead the case study There is also the work nobody commissioned. Products I took apart in public →