After a decade in the agency world, if I were starting from scratch today I wouldn't touch a SEO retainer. Here's the uncomfortable truth I've come to realise.
Three reasons SEO is a terrible foundation for a new agency:
First, organic clicks are being hollowed out by AI overviews. Even when you rank, half the searches no longer generate a click. the effort is higher, the payoff smaller.
second, it's the easiest service for a client to bring in-house. Once they grasp the basics, they hire a junior or subscribe to a tool. Six months of retainer revenue gone.
third, compounding takes six to nine months. The first half-year you're losing money on every retainer because results haven't manifested. Clients churn out of impatience before the work pays off.
i'd lead with voice AI instead because the opposite holds true across all three:
Results show within week one. The agent answers calls, books appointments, recovers missed calls. The client sees a booking in their CRM the same week.
Switching costs become massive after ninety days. Scripts, CRM wiring, number ports, recording archives. The client physically can't leave once it's deployed.
Margins are eye-popping. Two to five thousand per month per client at over eighty percent gross margin once you've built your platform stack.
the build order I'd actually run:
Month one to two: pick one vertical - HVAC, dental, real estate, or law firms. Stand up two or three voice AI deployments at a discount to learn the workflow. Goal is reps and case studies. revenue comes later.
Month three to four: charge full price - two to three thousand per month - starting with deployment number four. Use the case studies from months one and two to land them.
Month five to six: layer additional use cases per existing client - missed-call recovery, after-hours, dead-lead reactivation. That's where you go from two thousand ARPU to four or five thousand ARPU without acquiring any new logos.
Month seven to twelve: only then do you add SEO or AEO as a service line, but as a complement to voice AI clients. It becomes an upsell to existing clients who already trust you.
Honest tradeoff: this only works in verticals where call volume matters. if your target is e-commerce DTC brands or pure B2B SaaS marketing, voice AI isn't the right lead. You'd lead with paid and content.
Platform note since people will ask: Wave Runner AI if you want multi-tenant client setups - white-label, native CRM integrations, workflow builder. Retell AI if you're starting with a single-vertical play and want a simple agent. Either works for the build order above.
i keep meeting agency founders in the wild who started with voice AI and layered SEO or paid on top later. That's the shape i think makes sense for the next couple of years.
