How to keep your CRM and pipeline clean with AI.
I use this to keep the Airtable ATS-2026 base current and to answer questions off it without building views by hand - log call outcomes and next steps, add and update records, and ask things like total pipeline value or what is shipping this week. I reach for it after every call to log the outcome (the back half of the call-notes skill), and any time I want a number or a list out of the pipeline.
How I use this, and what for
I use this to keep the Airtable ATS-2026 base current and to answer questions off it without building views by hand - log call outcomes and next steps, add and update records, and ask things like total pipeline value or what is shipping this week. I reach for it after every call to log the outcome (the back half of the call-notes skill), and any time I want a number or a list out of the pipeline.
What the AI actually does
You talk to the CRM in plain English instead of clicking through it. The AI finds the right base and table, reads the schema, writes and updates records, and answers questions across the data. Logging that used to be a manual chore becomes a sentence; reporting that needed a custom view becomes a question.
Reading the schema first is not optional plumbing - it is what keeps records accurate. Airtable compares single and multi-select values by exact string match, so the AI has to read the existing field options before it writes a stage or status, or it risks creating a near-duplicate. The Airtable MCP handles that order for you: read the schema, then filter and write.
The tools and APIs you need
The right prompts
Log a call outcome
```
Log this to the ATS Airtable base: contact [name] at [company], call
outcome [one line], next step [action] due [date], stage moves to [stage].
Update the record if they already exist, otherwise create it.
```
Report off the pipeline
```
From the ATS base, give me total pipeline value and a breakdown by stage,
plus anything with a close date this month. Post the summary to #sales
in Slack.
```
How to do it
Point it at the base
It searches for the ATS base, lists tables, and reads the schema so it writes to the right fields.
Log outcomes
After a call, give it the summary and the next step - it creates or updates the contact and opportunity record.
Keep stages honest
Ask it to update stage, owner and next-action dates as things move.
Ask the pipeline questions
Total value, deals by stage, what is closing this month, roles shipping this week - it queries and answers.
Share the answer
to [Slack](https://slack.com/) if the team needs it.
See the difference
Doing it by hand: keep your crm and pipeline clean with ai the manual way - slow, and the first thing to slip when you are busy.
With AI: you describe what you want in plain English and it does the work, on-brand, in minutes.
Let AI carry the heavy lifting; you keep the judgement and the final say.
How you'll know it's working
- The pipeline reflects reality because logging takes one sentence.
- You get numbers and lists by asking, without building a view.
- Records go to the right fields because the schema is read first.
- The team sees the same picture you do.
Questions people ask
What is AI CRM management?
Using an AI assistant to run your CRM in plain English - you tell it what happened and it writes the record; you ask it a question and it reports off the pipeline. The finding-the-base, reading-the-schema and writing-to-the-right-fields all happen for you, instead of by hand in [Airtable](https://www.airtable.com/).
Will it write to the wrong fields?
It reads the schema before it writes, so it maps your sentence to the real field names and existing select options. [Airtable](https://www.airtable.com/) matches select values by exact string, so reading first is what stops it creating a near-duplicate stage or status.
How is this different from building the CRM or running routines?
This skill keeps an existing CRM accurate through plain-English logging and reporting. The build skill (build-crm-with-ai-airtable) stands up the tables and fields in the first place. The routines skill (ai-crm-routines-automation) runs this same upkeep on a schedule so you do not have to.
Hire with Assessment.
£7-10k flat fee. The methodology, delivered.
See Hire with Assessment