Turn a messy lead list into an automated pipeline using Teable.
Managing clients in a spreadsheet falls apart the moment you have more than a handful of them. In this video I build a lead management system in Teable from scratch: a pipeline that tracks every prospect, hot and cold categorization, project details in one place, and AI generated follow up emails so nothing goes cold while you are busy.
If you have ever lost a deal because you forgot to follow up, this is the fix.
Faith + Money: "In all toil there is profit, but mere talk tends only to poverty." Proverbs 14:23 (ESV)
Transparency: this video is sponsored by Teable, and the link above is an affiliate link. I only recommend tools I would use myself.
So this is a client management system.
It's got a lead pipeline, hot and cold lists,
every projects, details, and follow-up emails
that get written for me and sent after I approve them.
I didn't build any of this by hand.
I started with a messy spreadsheet of leads
and then I talked my way with AI to get here.
So this is where it started
and let me show you how we got to where we're at.
And we're gonna be looking at tbl.ai today
and there's an affiliate link down in the description.
If you sign up through it, I do get a commission,
you get 10% off your first payment, so a win-win.
I tested T-Bull on my own workflow for one of my businesses
so you can judge from the result
whether it's something that'll work for you.
So and T-Bull calls itself an AI spreadsheet for business.
So it looks like a spreadsheet, it behaves like a database,
and there's an AI inside it that will build the thing
for you if you describe it.
It's open source, it has over 20,000 stars on GitHub
and it hit number one on Product of the Day.
So this is definitely something worth checking out.
So here's my actual problem.
So I run a small web design business on the side,
I talk about it in a lot of my videos.
Leads come in from different places,
referrals, Google, a form, people just contact me
and they all end up in a spreadsheet
and that spreadsheet tends to be a mess.
Lots of times the same company is in there twice,
there's duplicates, it's not all in the same format,
it's just, it's kind of a mess.
Half the budgets are typed as text,
the dates are in different formats,
some rows don't have email addresses.
Let me know in the comments if this looks similar
to spreadsheets that you have in your business.
But underneath this mess of leads,
there's some actual money.
Because some of these people are actually looking to buy,
some have gone cold, and it's hard for me to tell
what's what by this mess of a spreadsheet that I have.
So I'm not going to ask AI to invent
a pretend business for me,
I'm gonna hand it the mess I already have
and see if it can turn it into something that can operate.
And this is what I love testing AI with,
is I have actual problems, can AI solve those problems?
So I wanted to do eight things, eight steps here.
I wanted to import the spreadsheet, clean it up,
sort the list from hot leads to cold leads,
put a form on the front, have it schedule
the follow-up emails, automate the sending,
build an internal app on top,
and then actually work a lead through it.
Okay, that's what I'm gonna aim to accomplish here.
By the way, I did change the names and the emails
and the phone numbers.
I swapped those out with the real ones,
I don't want the real ones on here.
So just note that.
Here's the starting point.
You notice I've done this a couple times already,
I've been playing around with it.
But we're gonna start by create a base.
And you'll notice this doesn't start with a blank grid,
it actually opens a conversation.
That's how you talk with AI, is a conversation.
So I am taking my spreadsheet and I am putting it in here.
And there's no export wizard, there's no field mapping
or anything like that, no deciding in advance
what every column is, I am just dragging the file in.
And I'm gonna say turn this web design lead data
into a client, see clean and standardize the fields,
fix the inconsistent capitalization.
By the way, favorite thing about AI,
you can misspell things and I'm not even gonna fix it,
it'll know.
Fix the capitalization, budgets, dates,
flag the duplicates, identify hot leads,
cold leads, and leads that need follow up.
And then I am starting it.
So that's all I'm gonna do, that's all the prep
I'm giving it.
The way that AI has evolved over time
is the idea is you should just be able to talk to it
like you would a person.
So just if you were giving this task to an employee,
what do you want done?
How would you explain it to them?
That's how you can talk to AI.
The assistant here is called Cuppy,
so that's kinda cute.
So we'll let this run.
Again, another thing I love about AI,
I can let this run, I'm gonna work on other stuff.
All right, so that took about four minutes,
so not too long.
Again, I was working on other stuff,
so it doesn't really matter.
But all my 34 rows made it in here,
and that's the first thing I cared about
because a migration that quietly drops records
is worse than no migration at all.
So when it did normalize the data,
so like budget, these were all a mess before,
and now they're in there nice and standard,
which is nice, because now I can sort it
based on these values that are numbers.
Again, if the data is not good,
if it's all in different formats,
you're not able to do that.
So I can highlight the row,
it gives me an average, gives me a sum,
that's nice to see.
It also figured out on its own,
I did not tell it to do this,
but it figured out if these were meant to be
one time or monthly,
kinda based on the notes that I had
about the customer in there.
Last contact, it did the same thing,
it normalized the data,
so these are all in standard date format now.
And once these are dates,
then really anything's possible, right?
That's when you can start doing follow-ups
based on the dates and things like that.
The dates are all in different formats,
you can't do anything with them.
And it categorized all of these
by what they're needing,
which this is great,
because then I can group them as I need to.
So I'm gonna put here all leads,
and then I have this duplicate review.
This was something I asked it to do,
is find potential duplicates.
So amazingly, it just went through
and said, hey, these are probably the same,
which they are.
And so then I can clean those up.
And I like it didn't just silently delete them,
but it made a duplicate review.
That's actually a really smart way of doing it,
because they might not be duplicates.
I want a human eye on it before deleting anything.
And it did the part that I actually wanted.
It didn't just tidy the data up,
but it made a judgment about them.
So it split these apart into cold leads and hot leads.
So I have cold leads and hot leads.
And then it also made a list of customers,
hey, here's who probably needs a message soon.
They need a follow-up now.
So the hot leads,
these are the ones with an actual budget on them,
a real budget, a recent conversation,
and a status that says that they're moving.
So these are the ones that I would follow up on,
probably today.
Cold leads, some of these haven't been contacted
in quite a while.
Some of them don't have a budget.
So these are like, hey,
they might turn into something, but probably not.
So in this follow-up queue,
this is the one that costs me money.
So these are the people who are interested,
who've not been written off,
and I simply haven't gotten back to them.
So that's a list I didn't have before I should have,
and I should be following up with these people.
So this was a qualified customer.
They're a hot lead,
and I've not followed up with them for a couple months.
Okay, that's bad.
Guys, I'm glad this found that.
So now cleaning up the leads that I already had,
that fixes today.
But the real solution,
the next one has to be,
don't be a mess in the first place.
Okay, that's what I want going forward.
So let's put down here,
create a form for new lead submissions
that writes into the leads table.
Ask for company, contact name, email, phone,
service needed, budget, and notes about the project.
Go.
All right, cuppy, get going on that one.
So this one only took about a minute,
but here the form is now built.
Every field on it maps to a column that already exists,
so nothing comes through this that creates a new stray field
or anything like that on the other side.
Oh, and I can access this from the web.
So here, you know, in just a minute,
created a form that goes onto this spreadsheet.
So again, if you're needing like a CRM
to keep track of leads,
here's, you know, in five minutes,
you can get it done with tbl and cuppy.
So, and I can put this exact thing on my website
and people can fill it out themselves.
You know, I could obviously change a few things.
Actually, you know, this is pretty nice.
I was gonna say I could change service needed
into a dropdown, but it already is there.
So, and I just wanna test, make sure it works.
We'll put Yankee Hill, sign works.
Budget is, we'll say 6,000.
Notes, old site is not mobile friendly.
So we'll just submit this and see how it goes.
Okay, submit succeeded.
Then we'll go back to our leads here,
and here is the lead.
So it did arrive through the contact form.
So I have this list of people to follow up with,
but the reason I have not followed up with people
is not just because I forgot.
It's because that writing a decent followup
takes 10 minutes and I, you know, I'm busy.
I'm doing a lot of other stuff.
This business as a whole is kind of a side project for me.
It's not a main focus.
So I just, it's fallen off the table.
Okay, I haven't done it.
Create an autofill AI field called followup email
that writes a personalized followup
using each lead's service needed, budget, and project notes.
Run it for the existing leads that need followup
and generate it automatically for any new leads.
Once those fields are filled,
also add an approve to send a checkbox
and a followup sent at field.
All right, copy, get going on that.
I'm guessing this will take about a minute,
but we will see.
Right, so it took about three minutes for that one.
We have this new field and you can see
it put down here a followup email.
So it even mentioned the budget
and what they were looking to do.
It looked over the notes.
It personalized it for each person.
So this isn't an old mail merge
with just their name put in,
but actually takes the details it knows about that customer
and puts it into the email.
All right, now I want the sending to be automatic,
but not in a way that actually, that terrifies me.
I don't want a robot just emailing my clients on its own.
I want it to go out when I say go.
So we're gonna go back to Clippy here,
create an automation.
When approve to send is checked on a lead,
send that lead's generated followup email
to their email address.
Then set followup sent at to now
and change their status to contacted.
So again, just in English,
what I want it to do is when I check approve to send,
it sends it and it updates all their data
to show that it was sent.
So let's do this.
All right, so it created this automation here
and you can see the whole thing.
When we approve a lead with email content,
validate recipient message, parse subject to message,
send approved followup email, record sent timestamp.
So it does it all.
I just love this.
You can just tell it what the workflow you want
and it just creates it.
And you can see in the code that it wrote
that it even is checking, it's validating.
So if there's no recipient, if there's no generated,
don't do it.
If the email send failed, show an error.
So when I tested the automation and it did come through.
So that is working great.
And back in Table, the lead has been marked as contacted.
The followup sent time has been updated down here.
It's all been recorded automatically.
All right, last piece.
So it's done everything pretty well so far.
Everything so far is a database
and a database is fine for me because I built it.
So that's fine, but it's not fine for everybody else.
If I bring somebody in to work these leads,
I don't want to teach them how to do views
and filters and all that.
So let's create an app.
So build a client management app.
And I really misspelled that one,
but again, I think AI will understand.
With a lead pipeline, hot and cold lead views,
a lead detail page showing their project requirements
and their AI generated followup content,
followup status, timestamp, unapproved followup action
and actions to mark a lead as engaged or converted.
All right, let's get going on it.
And after a few minutes, this is the app
that it came up with.
So when it has a link, you can go to it
and I could give this to somebody
and let them mess around with it.
So again, I was working with this list the whole video,
but it turned it into an app and the list is fine,
but the app a lot more easy for me to hand off
to someone else to use.
But it's a pipeline, every deal, what stage it's at
and I can see where my money is sitting
without opening anything.
So again, I can filter by just the hot leads,
just the cold leads, the warm leads.
Actually, it gave me a hot leads view,
a cold leads view and a followups view.
And I can click on a lead
and the whole picture is on one screen.
What they want, what they're spending,
my notes on the project,
the followup that was written for them,
whether it's gone out and when.
And I can approve the followup send from right here.
So, and I didn't lay any of this out.
I could make changes if I wanted to,
but this was just, I just told it what I was looking for
and then it went and did the rest.
So, honest assessment,
after running my own messy data through the whole thing,
the import and the cleanup, in my opinion,
is the most powerful part.
So I have a lot of old spreadsheets
that I'm definitely gonna, can throw into here
and it will really help me out.
It picked up the duplicates a lot better
than I thought it would.
I thought it might automatically remove them,
which I'm always hesitant for AI to automatically do things.
So I like to see that.
So the one thing that I would tell you to be careful with
are the automated emails.
So they're good drafts and they're a lot better
than just a template or something like that.
But there's still a machine writing to a customer
using notes I typed about that customer.
So I would recommend you look over everything.
Don't let AI just send out emails
without you looking at them first.
And again, that's why I built this approve checkbox
rather than just, hey, go do it.
There's temptation to just automate everything,
but that's one automation line
I'm not quite ready to cross yet.
So I would love for you to go try this out,
tbl.ai, link will be down below.
Go show me, put me in the comments
what that you have built with this.
Take whatever messy spreadsheet that you've been avoiding,
drop it in and let me see how it went.
So if you like AI tools and just like to see them reviewed
and see what's out there,
it's such a fast changing space,
please like and subscribe.
I'd be doing this pretty much every day.
Thank you so much for watching.
God bless you all and have a great rest of your day.