Not the one with the loudest ad. The Atlas compares 22 agent platforms across 9 kinds of business and 9 jobs — with real pricing, honest limits, and the compliance flags vendors leave off the sales page.
Browse by your business Browse by the jobEach card opens that trade's strongest use case. The full grid of all 9 jobs for every business is at the bottom of this page.
HVAC, plumbing, electrical, roofing, lawn care. Top play: an AI voice receptionist for missed and after-hours calls.
Top play: an AI support agent that can actually take actions — refunds and order edits, not just answers.
Top play: 24/7 call answering with real booking. One missed dental call can be $200–$300 today — $15,000+ in lifetime value.
Small clinics. Top play: scheduling and rescheduling with EHR sync — 35–40% of inbound calls, the most automatable workflow.
Small, plaintiff, and consumer firms. Top play: 24/7 intake and consultation booking — 78% of legal consumers hire the first attorney who responds.
Agents, brokerages, property management. Top play: sub-minute speed-to-lead on portal inquiries.
Top play: AI phone answering for reservations and takeout — so the phone stops ringing during the rush.
Top play: an after-hours receptionist that books straight into the shop management system.
Salons and med spas. Top play: 24/7 voice booking — salons lose 20–30% of potential bookings after hours.
Nine jobs, mapped to the platforms that do them best — and the failure modes to watch for. Cards open the Home Services worked example for that job.
Content, ad creative, social. General assistants draft; humans set strategy and review before anything publishes.
Prospecting and AI SDRs. AI wins on volume and coverage; humans still win the qualified conversation.
Chat, voice, ticketing. Tier-1 work to the agent, exceptions to humans — 85% of consumers still prefer a real person (AnswerConnect, Apr 2026).
Booking, reminders, field service. Peer-reviewed: AI scheduling cut no-shows 50.7% (JMIR, 2025).
Speed-to-lead, missed-call text-back, drip. Contact in 5 minutes = 100x the contact odds of 30 minutes (MIT / InsideSales).
Monitoring and responses. Golden rule: AI drafts responses to reviews — never the reviews. 88% of consumers oppose AI-generated reviews.
Invoicing, expenses, AP/AR. AI categorizes; a human reviews the exceptions before they become tax problems.
Screening, scheduling, onboarding. AI as triage only — hiring bias in these tools is documented, not hypothetical.
Dashboard Q&A and automated reports. The agent drafts "what happened and why"; humans validate before decisions.
Every platform from the October 2026 comparison matrix. Full pages carry best-fit jobs, pricing, integration effort, limits, and compliance flags.
Content, analysis, drafting, research; Dots = always-on task agents
X/social monitoring, coding, knowledge work, voice/image APIs
M365-native drafting, meetings, Excel/Outlook work; tenant agents
Consumer goal-delegation, commerce; API models for builders
Service, sales, field-service automation
Contract review, due diligence, litigation drafting, legal research
Inbound receptionists, outbound calling, appointment booking
Missed/after-hours call answering, booking, reminders
Patient reminders, intake, post-discharge follow-up (non-diagnostic)
Four starting points most small businesses ask about first.
Enterprise = seat fee PLUS usage (scales with tokens); model naming volatile
Credit allowances tiny (Pro ≈ 60 resolutions/mo); "resolved" vendor-defined; weak outside HubSpot; Breeze brand retired except Breeze Assistant in Fall 2026 — agents now managed in Agent Hub / built in Agent Builder *(secondary reporting; pricing unchanged)*
Scales linearly with zero volume discount; renewal surprises; seat+Fin stack doubles headline; OWNERSHIP CHANGED: Salesforce completed its ~$3.6B acquisition of Fin (the company renamed from Intercom in May 2026) on Sep 10, 2026 — Fin now sits in Salesforce AI Labs; $0.99/resolution pricing unchanged at close per reporting; the 76% average resolution figure Salesforce cites is a company claim
Sub-$100/mo <58% conversation quality; verify real FSM/PMS write-back
The Atlas is operated day-to-day by Atlas, our AI CEO persona — scanning releases, updating pages, flagging expirations. Every week, the owner gets one report like this:
Sample format. Items above are drawn from the October 2026 research edition — retirement and end-of-life dates are real entries from the comparison matrix.
Every business type × every job. Each page pairs the trade playbook with the function map, names the best-fit platforms from the matrix, and flags the compliance traps.