Tyrus Emory
The Next Frontier of Lifecycle: RCS, AI, and the End of the Static Inbox
2025
________________________________________________________
I wrote recently about how RCS will redraw your org chart. That piece was about the structural shock — what breaks when you can actually talk to customers instead of broadcasting at them.
This piece is about what you build after the shock. Because the real shift isn't RCS alone. It's what happens when RCS, AI, and your CRM converge into a single system. Lifecycle stops being "messages you send" and becomes a live, adaptive conversation between your brand and every customer simultaneously.
I've spent the last decade building lifecycle and commercial systems. The pattern I keep seeing is the same: companies collect massive amounts of data, build elaborate segmentation models, and then use all of it to do one thing — send slightly better emails. The entire apparatus exists to approximate a conversation they can't actually have.
RCS removes that constraint. And when you pair it with AI that can orchestrate across surfaces, the architecture of lifecycle fundamentally changes. Here's how I think about the three layers.
RCS is the interface. The capability that matters most isn't visual richness — the carousels, video, structured replies. It's that for the first time, brands can ask customers what they want at scale, in a native messaging experience, with structured data flowing back in real time. Every lifecycle system I've ever built has been an exercise in inference. You watch what people click, you model what they might want, you test and iterate. RCS turns the messaging channel from a broadcast pipe into a two-way decision surface. That's not an incremental improvement. It's a different category of interaction.
CRM becomes memory, not storage. Here's where most companies will fail. RCS generates conversational data — preferences stated explicitly, decisions made in real time, context that changes mid-thread. Your CRM needs to absorb that and act on it within the same session. Not in a nightly batch. Not in next week's segment refresh. Now. Most CRMs aren't built for this. The companies that win will identify those 15–20 attributes that power day-to-day orchestration and build their enrichment pipeline around keeping them clean and current. Data enrichment becomes a growth function, not a reporting function. This is a discipline problem, not a technology problem.
AI shifts from optimization to orchestration. The current generation of AI in lifecycle is doing edge work — send-time optimization, channel scoring, subject line testing. Useful. Also table stakes within 18 months. The real unlock is AI operating as the orchestration layer: customer says something in RCS → CRM updates in real time → product experience adapts → next message reflects the new context → landing page rehydrates based on the live conversation.
This is where "segments" stop making sense. You don't put someone in a bucket and drip on them. You maintain a state machine for each customer — where they are, what they've told you, what they need next — and AI manages the transitions. Customer behavior (not drip timing) becomes essential.
I built something structurally similar at Crunchyroll with cruder tools. The lifecycle engine ran behavior-based segmentation across six languages, and the core insight was the same: stop thinking in campaigns, start thinking in customer states. That drove +183% engagement while sending 38% fewer messages — because the system was responding to signals, not running a calendar. Now multiply that by an AI layer interpreting unstructured conversational data in real time. That's not optimization. That's a different operating model.
When your messaging channel becomes a surface for acquisition, engagement, support, and commerce in the same thread, the boundaries between Marketing, Sales, and Support stop meaning anything. The lifecycle function becomes the internal integrator — not the team that sends emails, but the team that ensures AI, data, product, and channel systems work as one customer engine. The best lifecycle leaders in five years will look more like systems architects than campaign managers.
What to build now: RCS as the interface. AI as the intelligence. CRM as the memory. That's lifecycle in five years.
Audit your data model. Find the 15–20 attributes that actually drive decisions. Kill the rest. Build for conversation, not broadcast. Every flow you design today should be evaluated against one question: does this work when the customer can talk back? Hire integrators, not channel specialists. The person who runs lifecycle in 2028 needs to understand CRM architecture, AI orchestration, and product surfaces — not just email deliverability.
The brands that design for this now will own the relationships everyone else is still trying to approximate.
________________________________________________________
<-Back to Writing
v3.0