lanternA CRM my brother‑in‑law could trust
Pete was not ready to let an AI near his clients' details. So his AI learned to keep a spreadsheet instead.
My brother-in-law Pete runs his own insurance business in New Zealand. Earlier this year he told me what he was paying for his CRM. It was a lot.
When we looked at how he used it, it was mostly a contact list. He opened it a few times a week to find a client's details. He rarely added anything. He did not want more fields or more features. He wanted his clients in one place, and he wanted the details to be right.
So I asked him the obvious question. Why not let an AI look after it?
Not yet, and he was right
He was not ready for that, and I think he was right. An AI can get things wrong and sound sure of itself. In his work, one made up email address could cost him dearly. His clients' details are not the place to experiment.
So the question changed. How could his AI do the work without asking Pete to trust its memory?
For me, the point is to keep the human at the centre.
Something in between
What we built sits in between. His clients live in a Google Sheet, the kind of spreadsheet people have always kept. His AI looks after it, and it follows the same steps every time. It reads the sheet before it answers anything about a client. If it does not know something, it leaves it blank and says so. Every change is logged, so Pete can see what changed and when. It never sends anything on his behalf.
His AI does
- Keeps the client list in a Google Sheet
- Reads the sheet before it answers anything about a client
- Leaves a blank, and says so, when it does not know
- Follows the same steps every time
- Logs every change
Pete keeps
- Every message to a client: his AI never sends anything
- The checking: the log shows him what changed and when
- The time face to face with his clients
Pete has been running it for three months and he enjoys it. His AI looks after the back end, and he spends the time with his clients.
For me, the point is to keep the human at the centre. Every change leaves a record, and that is how he trusts it a little more over time.
I liked it enough that I now run my own work the same way.
Why we shared it
We turned it into a spark on Lantern. A spark is a small piece of know-how that your AI can read and put to work with you.
This one came from a real problem in a real business, and I think that is where the best ones come from. Most of what I use every day I did not learn at university. I learned it from people, in the day to day. Sparks are a way to share that kind of know-how, with people and with their AI.
If you are paying for a CRM you mostly use as a contact list, this spark might be worth a look. Your AI can read it on Lantern and set it up with you.
This is the first of a few stories we will share about where our sparks come from. Thank you, Pete, for letting me tell yours.
How this spark came to life
- 01
The snag
A paid CRM, used mostly as a contact list
- 02
The worry
Client details are no place to experiment
- 03
The build
A sheet his AI keeps, every change logged
- 04
The proof
Three months of use, and he enjoys it
- 05
The spark
Vetted by a person, published on Lantern
Keep the CRM in a sheet your agent maintains
A skill and a weekly audit agent that set up and run a relationship sheet in Google Sheets, Excel or .csv: IDs, a log, a standing instruction per person, days since contact. Never sends.
Free account. Your AI reads the spark and sets it up with you.
Already connected to Lantern? Tell your AI:
Read the Lantern spark crm-in-a-spreadsheet and set it up with me. Every spark has a story. What's yours?
Tell your AI:
Help me propose a spark to Lantern.