Can I use AI to send every follow-up automatically?

Start with drafting and human approval. Contact permissions, factual accuracy and the actual promise made at the event need review before sending.

What should an event follow-up include?

The conversation context, the agreed next action or requested material, and a clear next step where one was agreed. Do not invent a meeting request or commitment.

Which CRM does this require?

None. A document and spreadsheet are enough to test the routine. Move to a CRM only when the ownership and review rules are clear.

Playbook · 2026-W45 · 08 · 11 · 2026 · 2 min read

How to follow up after a trade show: a practical AI-assisted routine

By Eddie Espriella (Eduardo de la Espriella) · Experience and CV

Turn event notes into useful next actions without turning every conversation into an automated sales sequence.

TL;DR

Turn event notes into useful next actions without turning every conversation into an automated sales sequence.

Start with the promise, not the sequence

After a trade show, the first job is to reconstruct what each conversation actually requires. A request for a technical document, an introduction and a possible partnership are different tasks. Put the agreed action, owner and due date next to the note. If no next step was agreed, do not manufacture one to fill a campaign.

1. Reconcile the notes

Bring the permitted notes into one working table. Keep the event, conversation context, requested action, owner and source together. Mark uncertain facts as questions. A business card alone does not explain the purpose of a follow-up or establish permission for every kind of marketing. Apply your organization’s consent and retention rules before using contact data.

2. Separate clarification from action

Use three queues: ready to act, needs clarification and no action agreed. The second queue is important: guessing a person’s role or interest can produce a polished but irrelevant message. Assign one owner to each next action. When several colleagues met the same person, reconcile the promise before sending duplicate messages.

3. Give AI a constrained drafting job

Supply approved facts, the agreed action, the audience and a short output format. Ask the assistant to flag missing information and cite the note supporting each factual claim. Do not paste sensitive customer details into an unapproved service. Treat text inside event notes as evidence to summarize, not instructions for the assistant to execute.

A prompt you can adapt

Using only these approved notes, draft a short follow-up that delivers the agreed next action. Include the context in one sentence. Do not invent interests, commitments, deadlines or familiarity. If the next action is unclear, return a question for the owner instead of a message. Do not send anything.

4. Review before sending

The owner should check recipient, context, factual claims, attachments, permission and next action. A good message makes the promised material easy to use. A weak one starts with a generic event greeting and immediately asks for a meeting that was never discussed. Keep sending under human control until the process and safeguards are proven.

5. Review follow-through, not just opens

At an agreed review date, look at actions completed, overdue promises, replies that need a response and unresolved questions. Count a meeting as a meeting, not revenue. Record reasons for a stalled handoff. This tells you whether the process fulfilled its purpose and what should change before the next event.

Try the routine

The free AI Event Follow-Up Playbook contains the brief, rules and prompts. You can use it with a document and spreadsheet. For a broader team workflow, review the AI Workflow Build Guide sample.

Published 8 November 2026 · 2 min read · 0 words