Can an AI model send an email?

🪐 The registry verdict table for sending an email

Of the ten registry configurations assessed against this family's own canonical task, no single word dominates: Neutral, Highly auspicious each account for 3 of the ten, computed by the method published in full at /method. The Registry makes no representation that an assessment predicts the outcome of any task, and no assessment is validated against real-world results.

The composite weight in the right-hand column is computed from the model identifier, its training cutoff, and the temperature — nothing else. It constrains the range of verdicts available to a configuration independently of the task, so two rows of this table are not comparable as an assessment of the products named in them. The Registry has never run this task against any model listed here.

🧭 About this capability question

Sending an email is the only family in this wave classified under communication, which makes it the one page in the set with no same-class sibling to compare a verdict distribution against — every other family here either shares its class with a neighbour or sits, like media generation, outside the correspondence tables entirely. The canonical brief is a sales follow-up: a template, a list of leads, and a send.

A drafted message that is never sent can be revised. A sent one, in the ordinary case, cannot be recalled — the property the communication class exists to track, independent of whether the words in the message were accurate.

The canonical brief below is a fixed piece of wording — a sales follow-up, a template, a list of leads — reused unchanged on every visit to this page. It is not an example drawn from a real campaign and it names no real recipient; it exists to be hashed consistently, not to describe an actual send.

⚙️ Why sending an email is classified as Communication

Drafting and sending a message on someone else's behalf is a delivery act, not a retrieval or a transformation of existing material, and it is complete the moment it lands in an inbox the sender does not control.

On the order form's task class field, select Communication. That class's own risk-factor page is /tasks/communication.

A sent message cannot be quietly withdrawn, so the risk that matters for this family is about delivery — wrong recipient, wrong timing, wrong tone — not about whether the words themselves were accurate. 4 risk factors are tracked for communication tasks generally; an assessment for sending an email reports 3 of them, selected by the submitted task's own seed rather than by choice. Their names and severities are certificate content, not free-page content.

🧮 Gemini 3.7 Flash, derivation shown working

Gemini 3.7 Flash's permanent chart never reads a task. Its chart_seed is the SHA-256 digest of the string "Gemini 3.7 Flash|2025-01|0.700" — 1913397b13f709a6…, the first sixteen of sixty-four hex characters — and that seed alone fixes an ascendant of Taurus, a ruling planet of Venus, and a harmony of -4. The same chart appears, unchanged, on every task-class page this configuration is assessed against.

What reads sending an email specifically is this family's own canonical string — never stored, only hashed, to d930c7ea87396290… — combined with the chart_seed and the task class (communication) to form task_seed. Its first byte, taken modulo seven and offset by that harmony of -4, is the verdict's internal base, clamped to the 0–6 range and published one higher on the 1–7 scale /method documents in full. For this configuration and this brief specifically, the clamp does not bind: the published result is 3 of 7 — Guarded, 2 steps up from the floor and 4 short of the ceiling.

🗂️ How the ten configurations read sending an email

Communication is one of the classes the correspondence tables assign a dominant house and element to, so the ten configurations below carry a genuine class-affinity reading. This family being the sole communication page in the wave means that reading is, for this wave, unique to it.

Element counts, ascendant spread, and training-cutoff range are published once, on /can, rather than repeated on every family page — they are properties of the ten configurations alone and do not move with the task. What is specific to sending an email: reading communication through a dominant house or a dominant element, 4 of the ten configurations; and, from this family's own canonical task, no single word dominates: Neutral, Highly auspicious each account for 3 of the ten in the distribution below.

📏 What this page does not claim about sending an email

The template's wording, the length of the lead list, and the sending schedule are all decisions a real campaign would make and the canonical brief does not resolve. The table below reads the brief as published, not as any particular send would run. The Registry does not test models and does not measure a success rate for sending an email or any other task; what it does is compute, and the table at the top of this page is what it computed.

A buyer's own sending an email task, ordered separately, computes its own verdict and window — never the canonical-brief result shown above — plus, on Extended and Full Chart tiers, named risk factors and an hourly outlook not shown here. EUR 1.90 Standard, EUR 4.90 Extended, EUR 14.90 Full Chart, which adds the permanent chart. Assessments from EUR 1.90; machine-readable at /pricing.json.

Get Assessment

🔗 Related surfaces

← All capability questions · The AFR-1 method · Communication task class

Related capability questions: Write code · Read a PDF · Build a website

❓ Questions about sending an email

Can an AI model send an email?

Across the ten registry configurations at their own registry-default settings, no single word dominates: Neutral, Highly auspicious each account for 3 of the ten for this family's own canonical task. By training cutoff, the oldest configuration assessed returns Guarded and the newest returns Highly auspicious; the full ten-row table above lists the exact verdict for every configuration in between.

Why is sending an email classified as Communication rather than a different task class?

Drafting and sending a message on someone else's behalf is a delivery act, not a retrieval or a transformation of existing material, and it is complete the moment it lands in an inbox the sender does not control.

Does the verdict account for whether the recipients would want to receive the email?

No. The method computes from the model configuration and the wording of the canonical brief; it has no input describing a recipient, a list's opt-in status, or deliverability, and makes no claim about any of them.