Log a call without inventing fields

Automated call logging is an Agent Skill: a written method your AI applies on demand, in Zealos or in any client that reads SKILL.md.

CRM data hygiene survives AI only if one rule is written down first: a field left blank is a known unknown, while a field filled in by inference is an unknown unknown that will be reported on as fact.

What is in this skill

namecrm-field-conventions
descriptionWhat belongs in each CRM field, and what to leave blank rather than guess. Use whenever writing to a contact, company or deal record, or logging activity after a call or email.
allowed-toolsnone — this skill adds judgement, not reach
compatibilityZealos, Claude Code, Cursor, Codex and any client that reads SKILL.md
licenseFree to use and edit

How it works

  1. 1List the fields that may never be inferred — budget, headcount, timeline, decision maker, competitor — and require a stated source before any of them is written.
  2. 2Define what a next step is: one action, one owner, one date. Follow up is not a next step, and an empty next-step field is a signal worth preserving.
  3. 3Write the recap first and extract fields from the recap, never straight from memory. Extraction from memory is where invented values come from.
  4. 4Separate observation from interpretation in every note, so the next person can tell what the customer said from what a model concluded.

The skill itself

---
name: crm-field-conventions
description: What belongs in each CRM field, and what to leave blank rather than guess. Use whenever writing to a contact, company or deal record, or logging activity after a call or email.
---

# CRM field conventions

## The rule that matters most

**Leave a field blank rather than infer it.** A blank field is a known unknown.
A guessed field is an unknown unknown, and it will be reported on, forecast
against, and used to route a lead. If a value was not stated by the customer or
recorded in a source you can cite, do not write it — say what is missing and why.

Never infer: budget, headcount, timeline, decision maker, competitor, industry
from a company name, seniority from a title alone.

## Fields

- **Next step** — one action, with an owner and a date. "Follow up" is not a
  next step. If there is no agreed next step, leave it empty; that emptiness is
  the most useful signal in the pipeline.
- **Close date** — what the buyer said, not what the quarter needs. If they have
  not said, leave it and note that it is unset.
- **Amount** — the discussed figure. If it is a range, take the low end and note
  the range in the deal notes.
- **Stage** — moved by evidence, not by optimism. Every stage has an exit
  criterion; if you cannot name the evidence, do not advance it.
- **Notes** — what was said, attributed. Quote where the wording matters.
  Separate observation from interpretation.
- **Contact role** — only from a stated org relationship, never from a title.

## After a call or email

Write the recap first, then extract fields from it. Extracting straight from
memory is where invented values come from.

Who uses it

Sales operations

Pipeline reports stop quietly including guessed close dates, and the gaps in the data become visible enough to chase rather than invisible enough to trust.

Account executive

Logging a call takes no time and produces nothing you would be embarrassed to have a manager read back to you in a deal review.

RevOps

Forecast accuracy improves for an unglamorous reason: the model stopped filling in the fields nobody had actually asked about.

Frequently asked questions

How do I stop AI making up CRM values?

Give it an explicit list of fields it must leave blank rather than infer, and require that anything written cites where it came from. This is the single highest-value paragraph in any CRM skill.

Does this work for HubSpot and Salesforce?

Both. The skill describes the convention and the agent's tools do the writing, so the same rules apply to whichever CRM is connected.

What about custom fields?

Name them and say what each one means. A skill is documentation the model reads, so it handles a custom object model as easily as a standard one.

Related skills

Summarize calls and update the CRM

Post-call automation turns a recorded conversation into the three things that outlive it: the commitments each side made, the fields that changed, and the follow-up that has to go out today.

MeetingsCRMZoom/call-recap-format

Brief yourself before every call

A pre-call brief is worth writing only if it changes what you ask: who is in the room, what they are measured on, what happened last time, and the one question you cannot afford to leave without answering.

CoachingMeetingsAccounts/discovery-questions

Run an AI SDR play on new leads

An AI SDR agent is only as good as the definition of a good customer it is working from, which is why the fit criteria belong in a written skill rather than buried in a prompt nobody can find.

ProspectingSDRCRM/our-icp

Triage your inbox by deal impact

AI email triage is worth having only when it sorts by consequence rather than by recency: what blocks a deal, what someone is waiting on you for, and what can be read at the end of the week.

EmailGmailAccount executives/email-rules

Qualify a lead against our ICP

Qualification is a claim about evidence, not a score: a deal is qualified when every criterion has a source you can quote, and unevidenced criteria are named rather than filled in optimistically.

DealsProspectingCRM/meddicc

Run a pipeline review

A pipeline review is useful in proportion to how uncomfortable it is: the job is to find the deals that are slipping quietly, not to narrate the ones everybody already knows about.

DealsCRMAccount executives/meddicc

Skills, in your workspace

Write it once. Every agent follows it.

Twelve built-in skills ship with every workspace, and you can draft your own with AI or import any public SKILL.md from GitHub.

7-day free trial. Set up in minutes. Cancel anytime.