Assumptions in the open.
A practical research method built around buyer questions, evidence boundaries and accountable decisions.
What is evidence?
We separate primary documentation, user-supplied information, editorial judgement and synthetic examples. A vendor’s API page establishes a research lead, not a verified integration in your account. External sources and review dates appear on the relevant briefings.
How route scoring works
Tool: evidenced native feature +3, bounded workflow +2, modest volume/budget +2, low complexity +1. Build: named builder +3, verified API +2, maintenance cover +2, available time +1. Expert: multiple-system complexity +3, accountable support +2, funded integration +2, comparable references +1. Unknown answers earn no points. These are editorial fit heuristics, not predictions.
Gates take priority
No demonstrated spare capacity weakens the call-recovery case. No booking-rule owner or usable data means prepare first. No builder and cover blocks Build; unverified integration blocks confirmed-booking automation. Safety diagnosis is never permitted.
How we calculate
Eligible enquiries = missed calls × suitable unique share × (1 − callback recovery). Additional completed jobs = the lower of capacity and eligible enquiries × incremental recovery × booking × completion. Net contribution subtracts recurring fees, usage and supervision. First-year results subtract setup and assume the entered month repeats, unless separately modelled.
Independence and limitations
No paid ranking affects the comparison. No measured savings, qualifications or supplier relationships are claimed without evidence. Working tools use local deterministic rules. Later-stage tools are authored implementations of the supplied concepts, not attributed missing research. Photographs show the working context and do not establish customer outcomes.