How is your website ranking on ChatGPT?
A practical playbook to make your marketing emails parsable and actionable so inbox agents can prioritize, schedule, and convert without manual work.
Vicky
Sep 29, 2025
[MEETING] Q4 plan review - propose 3 slots or [OFFER] Team plan upgrade - approve quote by Oct 7.\n- Keep tokens consistent across campaigns: [MEETING], [OFFER], [EVENT], [INVOICE], [NEWSLETTER], [ACTION NEEDED], [REMINDER], [SECURITY].\n\n### 3) Make scheduling machine readable\n- Always attach an iCalendar file for any meeting ask. Include DTSTART, DTEND, SUMMARY, ORGANIZER, ATTENDEE, UID, LOCATION per the iCalendar RFC 5545 standard.\n- Mirror availability in the body as explicit options the agent can parse.\n- Example lines: Option A: Tue Oct 7 10:00 10:30 PT Option B: Wed Oct 8 13:00 13:30 PT Option C: Thu Oct 9 09:00 09:30 PT.\n\n### 4) Provide a structured offer block\n- Summarize pricing and specs in compact table-like lines, not just prose.\n- Use canonical labels that are easy to learn: Plan, Seats, Unit price, Currency, Term, Total, SKU, Start date, End date.\n- Example line: Plan: Team Seats: 25 Unit price: 12.00 Currency: USD Term: 12 months Total: 3,600.00 SKU: TEAM-25 Start date: 2025-10-15.\n- If you need tiers, list one per line with the same label order.\n\n### 5) Write reply‑ready choices\n- Offer unambiguous buttons as text that map to intents the assistant can execute.\n- Pattern in body: Reply with ACCEPT to schedule Option B. Reply with ADJUST to request new times. Reply with APPROVE to confirm the Team plan quote.\n\n### 6) Add List‑Unsubscribe signals for safety and trust\n- Include List-Unsubscribe and List-Unsubscribe-Post: List-Unsubscribe=One-Click so assistants and MUAs can perform trusted one‑click unsubscribes. See the one‑click List‑Unsubscribe standard for required fields.\n- DKIM sign headers and include both unsubscribe fields in the h= list of the signature.\n- This reduces friction signals and helps assistants treat your domain as well behaved.\n\n### 7) Use assistant‑friendly headers and footers\n- If your ESP allows custom headers, add X-Assistant-Intent with a controlled vocabulary such as schedule, approve_quote, register_event, pay_invoice.\n- Repeat the same intent in the opening TLDR and subject token to give models multiple anchors.\n- Place a compact metadata footer for key fields, one per line, prefixed with labels. Keep it plain text for fallback parsing.\n\n### 8) Respect how assistants get invoked\n- If the workflow supports CC‑based invocation, place the assistant address on the thread when you want action. Use the TLDR and intent tokens to make the ask unambiguous.\n- Include who is the decision maker and who should be scheduled. Example: Decision maker: Dana Wong Schedule with: dana@company.com.\n\n### 9) Keep layout simple and semantic\n- Prefer clean HTML with a single column, logical heading order, and real text for tables.\n- Avoid image‑only pricing and avoid nested tables that hinder parsing.\n- Ensure alt text describes the action or value, not marketing fluff.\n\n### 10) Strengthen deliverability signals assistants use as priors\n- Authenticate with SPF, DKIM, and DMARC. Align From with your primary domain. Publish BIMI if eligible.\n- Consistency of domain, friendly from name, and stable send times help models build trust features.\n\n### 11) Provide event and workflow intents for campaigns\n- Events: invitations, webinars, office hours. Attach ICS and include registration intent lines: Register: yes or no. If yes, state seat hold until a date.\n- Sales motions: quotes, renewals, invoices. Use the structured offer block and reply tokens APPROVE, CHANGE, DECLINE.\n- Support motions: ticket updates. Include Ticket ID, Status, Next action, SLA, and tokens ACKNOWLEDGE or ESCALATE.\n\n## Assistant‑friendly email example\nSubject: [MEETING] Q4 plan review - choose a slot\n\nTLDR: Goal: confirm a 30 minute review Action: choose a time Deadline: Oct 7 5 pm PT Priority: high\n\nHello <First name>, here are three times for a 30 minute review.\n- Option A: Tue Oct 7 10:00 10:30 PT\n- Option B: Wed Oct 8 13:00 13:30 PT\n- Option C: Thu Oct 9 09:00 09:30 PT\n\nAttach: q4-review.ics with the same times.\n\nReply with ACCEPT plus the option letter to schedule. Reply with ADJUST to propose a new window.\n\nAssistant metadata footer\n- Intent: schedule\n- Meeting length: 30 minutes\n- Participants: you and Dana Wong\n- Priority: high\n\n## Measurement plan\n- Placement: percent of sends that assistants pin, star, or mark important.\n- Action: meetings created, approvals returned, registrations completed, daily summaries that cite your email.\n- Speed: time to first assistant action and time to outcome.\n- Safety: one‑click unsubscribes completed and spam complaints.\n\n## Governance and risk\n- Test intent tokens and TLDR variants with A B tests and track assistant actions, not just open rate.\n- Avoid over‑tagging subjects. If every subject says [ACTION NEEDED], models will discount the token.\n- Keep personally identifiable information narrow in footer metadata. Do not include secrets or payment data in machine‑readable blocks.\n\n## What to implement this week\n- Add TLDR, subject intent tokens, and reply choices to your top 3 workflows.\n- Attach ICS for all scheduling asks and mirror options in text.\n- Ship List‑Unsubscribe and List‑Unsubscribe‑Post with DKIM coverage.\n- Standardize the offer block labels and roll into your ESP templates.\n\n## How this connects to your AEO stack\n- If you also publish through Perplexity’s browser, apply these patterns alongside the sidecar tactics in our Comet optimization playbook.\n- For Microsoft stacks, see cross‑assistant patterns in Model‑Aware AEO for Microsoft 365 Copilot.\n- To train subscribers to expect machine‑readable actions daily, adopt the TLDR and reply tokens in ChatGPT Pulse Optimization for briefings.\n\n## Bottom line\nAssistants reward clarity, structure, and safety. If you make the next step machine readable and low friction, agents will prioritize, schedule, and convert on your behalf. The brands that adopt Assistant Inbox Optimization first will earn top placement in AI triage.