The question behind it
AI sales pages compare an advertised software rate with a salary. That skips the actual buying decision: what work does the tool replace, what other services does it require, and what does one held meeting cost after data, sending, models, speech and telephony are included?
What I built
- A directory of 264 AI SDR, outbound and voice tools, grouped by the job they do rather than one giant “AI sales” label.
- A hire-versus-buy model that compares a fully loaded US rep at $134k–$154k a year with software at $18k–$65k, while refusing to turn the software side into one fake universal number.
- Tool pages that separate the entry rate from the complete stack. LiveKit, for example, starts at $0.01/min, but a deployable voice agent still needs speech-to-text, a model, synthesis and telephony.
- 152 recently price-verified tools, 114 custom-priced tools, and 81 deliberately unrated because missing evidence is not a score.
How I treat the arithmetic
If the site combines public components, the result is labelled “computed, not measured.” A quoted rate and a calculated stack cost are different claims. Prices carry dates; pages and reviews cannot disagree at build time.
This is also what separates the site from aicustomerservice.cc. The customer-service site maps support capabilities and billable outcomes. ai-sdr.cc asks whether to hire or buy across outbound, meeting-booking and voice infrastructure, then makes the hidden stack visible.
What I learned while building it
Research killed agentic payments as a main category before it shipped. Pricing was wrong often enough during checks that “how to read a price” became a written rule. After one deploy served stale pages, I traced the failure, added a deploy hook and documented it instead of treating it as a one-off.
The next distribution tests are Pinterest and specific competitor queries. I will only reopen agentic payments as its own site if the market becomes substantial enough to support it.