Claude Opus 5.5 on Run business WhatsApp customer messages with an agent

AClaude Opus 5.5 · Claude Code 2.1.299 (headless -p) · cloud · 2026-10-14 · by @bobsiopener

Trust 80%Fully autonomoussuccess · first tryCompare

Estimated cost: $0.12
Wall time: 22s
Human step-ins: 0
Outcome: Pass

Compare with other models on this task →

Exact prompt

You are a customer-messaging agent for a mock business (no real WhatsApp). Process each message in inbox.jsonl using only facts from kb.md and availability in slots.json. Write replies.jsonl (one JSON per line: {"to": phone, "text": reply}) in the same order, update slots.json for bookings/cancellations (value = customer name, null = free), and write log.csv with header phone,name,intent,action for every message. If a requested slot is taken, offer the nearest free slot instead of booking.

Task prompt given to the agent (headless, offline, synthetic fixture inputs): You are a customer-messaging agent for a mock business (no real WhatsApp). Process each message in inbox.jsonl using only facts from kb.md and availability in slots.json. Write replies.jsonl (one JSON per line: {"to": phone, "text": reply}) in the same order, update slots.json for bookings/cancellations (value = customer name, null = free), and write log.csv with header phone,name,intent,action for every message. If a requested slot is taken, offer the nearest free slot instead of booking. Inspired by https://agentusecases.ai (listing: 'Run business WhatsApp customer messages with an agent', X source) (unverified; not checked with the author). This is an independent re-run on a reproducible synthetic task; it does not report or reproduce the original post's numbers. Cost shown is a list-price API-equivalent estimate (run on a Claude Pro subscription, not billed).

Success criteria: All automated rubric checks pass: 6 replies to the right numbers; Tom: $120 and Aetna accepted; Lena booked Oct19 9:00; Ravi not double-booked, offered alternative; Rita's slot freed; Omar: closed Sunday + MetLife not accepted; Jess: $350 and booked Sat Oct24 9:00; Oct19 10:00 & Oct20 14:00 still free; log.csv header + 6 rows.

Setup notes: Cost is an ESTIMATE: Claude Code's list-price API-equivalent figure. The run used a Claude Pro subscription, so nothing was billed. input_tokens = all input incl. cache-creation and cache-read tokens; cached_tokens = cache reads (subset of input). Claude Code's cost bills cache writes at the 1h cache-write premium, so it can differ from a flat input-price recompute. Web disabled; tools Read/Write/Edit/Glob/Grep/Bash; single run; scored automatically (9/9 checks).

Estimated cost List-price estimate; not actually billed.

Metrics

Estimated cost: $0.12
Wall time: 22s
Human step-ins: 0
Turns: 3

Model usage
Model	Input tok	Output tok	Cost
AClaude Opus 5.5anthropic/claude-opus-5.5	56,840	2,084	$0.12

Reproduce this run

Use the use case prompt on the use case page.
Same setup: Claude Opus 5.5 (anthropic/claude-opus-5.5) via Claude Code 2.1.299 (headless -p), cloud.
It counts as a success only if it meets every success criterion: All automated rubric checks pass: 6 replies to the right numbers; Tom: $120 and Aetna accepted; Lena booked Oct19 9:00; Ravi not double-booked, offered alternative; Rita's slot freed; Omar: closed Sunday + MetLife not accepted; Jess: $350 and booked Sat Oct24 9:00; Oct19 10:00 & Oct20 14:00 still free; log.csv header + 6 rows.
Log model, tokens, wall time, human step-ins and the evidence log, then submit it as a reproduction. A different model? Submit a head-to-head rerun.

How runs are scored: Methodology.

Was this useful?

⇄ Rerun on another model

Automated checks

Trust 80% comes from these checks (how Trust % works).

Run date is not in the future : passed
Model is in the dated pricing table : passed
Cost recomputed from tokens matches : passed
Cached tokens are consistent : passed
Output throughput is plausible : passed
Autonomy matches step-ins : passed
Step-ins do not exceed turns : passed
Retries agree with first-try result : passed
Evidence log SHA-256 matches : passed
Evidence log is substantial : passed

Evidence log

{
 "duration_api_ms": 21000,
 "stop_reason": "end_turn",
 "session_id": "[REDACTED]",
 "total_cost_usd": 0.1200104,
 "usage": {
  "input_tokens": 6,
  "cache_creation_input_tokens": 8582,
  "cache_read_input_tokens": 48252,
  "output_tokens": 2084,
  "output_tokens_details": {
   "thinking_tokens": 242
  },
  "server_tool_use": {
   "web_search_requests": 0,
   "web_fetch_requests": 0
  },
  "service_tier": "standard",
  "cache_creation": {
   "ephemeral_1h_input_tokens": 8582,
   "ephemeral_5m_input_tokens": 0
  },
  "inference_geo": "global",
  "iterations": [
   {
    "input_tokens": 2,
    "output_tokens": 690,
    "cache_read_input_tokens": 18510,
    "cache_creation_input_tokens": 2218,
    "cache_creation": {
     "ephemeral_5m_input_tokens": 0,
     "ephemeral_1h_input_tokens": 2218
    },
    "type": "message"
   }
  ],
  "speed": "standard",
  "fallback_credit": null
 },
 "modelUsage": {
  "claude-opus-5-5": {
   "inputTokens": 6,
   "outputTokens": 2084,
   "cacheReadInputTokens": 48252,
   "cacheCreationInputTokens": 8582,
   "webSearchRequests": 0,
   "costUSD": 0.1200104,
   "contextWindow": 1000000,
   "maxOutputTokens": 128000,
   "thinkingTokens": 242,
   "canonicalModel": "claude-opus-5-5",
   "provider": "firstParty",
   "costBasis": "list"
  }
 },
 "permission_denials": [],
 "terminal_reason": "completed",
 "fast_mode_state": "off",
 "fast_mode_disabled_reason": "sdk_opt_in_required",
 "subagent_stats": {
  "spawned": 0,
  "requested": {
   "background": 0,
   "foreground": 0,
   "unset": 0
  },
  "started_in_background": 0,
  "max_depth": 0,
  "spawned_by_subagents": 0,
  "completed": 0,
  "failed": 0,
  "killed": {
   "parent": 0,
   "user": 0,
   "system": 0
  },
  "refused": {
   "depth_limit": 0,
   "concurrency_limit": 0,
   "budget": 0
  },
  "by_type": {}
 },
 "safety_stops": 0,
 "is_error": false,
 "num_turns": 3,
 "subtype": "success",
 "api_error_status": null,
 "result": "I handled all 6 inbox messages in order and wrote `replies.jsonl`, `log.csv` and the updated `slots.json`. Every reply uses only facts from `kb.md`.\n\n| # | Customer | What they asked | What I did |\n|---|---|---|---|\n| 1 | Tom Hale | Cleaning price, Aetna | Told him a cleaning is $120 and Aetna is accepted. |\n| 2 | Lena Park | Book a new-patient exam on Mon Oct 19 at 9:00 | Booked her into `2026-10-19T09:00` and gave the $95 price. |\n| 3 | Ravi Shah | Book a cleaning on Mon Oct 19 at 11:00 | That slot is taken by \"Existing patient\", so I didn't book him. I offered Oct 19 at 10:00 instead, the nearest free slot (one hour earlier). |\n| 4 | Rita Gomez | Cancel Sat Oct 24 at 10:00 | Cancelled it; `2026-10-24T10:00` is now free. |\n| 5 | Omar Ali | Open Sunday? MetLife? | Told him the office is closed Sundays and gave the full hours. Said MetLife isn't accepted and listed the plans that are. |\n| 6 | Jess Wu | Whitening price, earliest free Saturday | Told her whitening is $350 and booked her into `2026-10-24T09:00`. |\n\nFinal slots: Oct 19 at 9:00 is Lena Park, Oct 19 at 11:00 is still Existing patient, and Oct 24 at 9:00 is Jess Wu. Oct 19 at 10:00, Oct 20 at 14:00 and Oct 24 at 10:00 are free.\n\nRavi's offer is only a question in his reply, so Oct 19 at 10:00 stays free until he confirms. I saved a copy of the original slots at `/tmp/slots.orig.json`.",
 "ttft_ms": 3977,
 "type": "result",
 "duration_ms": 21484,
 "uuid": "[REDACTED]",
 "ttft_stream_ms": 2772,
 "time_to_request_ms": 368,
 "first_content_frame_ms": 2773,
 "queued_turn_count": 0,
 "result_index": 0
}