#4: Is There a Place for AI in Sales?

Welcome back to another episode of The Junction.

Today we are going to be talking about

is there a place for AI in sales?

So we're unpacking various different industries, how people

and departments are using these AI tools and

not just AI, automation in general.

General, right.

So today we're going to explore the benefits, challenges

and even just maybe dive into some ethical considerations

of integrating AI into the sales process.

So let's just jump into you started, you started here then

in sales or was it more of a technical capacity?

Was Scott selling the work and you were

he sold the work, I did the work.

And then we just circle back around and

we do it over and over again.

It was great.

Then he got tired of selling.

Maybe that's not the best way to

say like he enjoyed interacting with people. Sure.

Scott's a relationship guy and I tend to be sometimes.

But when you build the things that you end

up selling it becomes really easy to sell.

So naturally I started selling what Scott was selling

and it ended up working out really great.

And you were a team of two at the time.

Team of two.

We've grown a lot since then, manoimano.

And now we're at 26 or 27 or 28.

I forget we're in the 20s.

Yeah, so it's working out really well.

So as we've evolved as an organization, obviously

sales is really I hate to say it

because I sit in marketing, sales is king. Right?

Because if we're not selling any work, we're not

making any money, we're not delivering any work.

So obviously that function has evolved a lot.

And as a company that focuses on integrating technology

and automation to help people do more faster, what

has that evolution looked like for us?

What automation have we built into our sales process?

And then let's maybe look at how we're exploring

what AI layered on top of that automation can

do for sales teams, small and large.

One of the things that I found is that

from an automation perspective at least, sales guys are

the least detailed people in the room.

They're not going to fill out that required field

and if they are forced to fill it out,

they're going to fill it out with what they

think they know and sometimes you don't.

And that's a fair ask, right?

Like, I don't know but you're

requiring me to fill it out.

Well, a lot of that data we

want from a closed loss perspective.

Or why did we win this deal, why

did we win that deal over this deal?

Or why was this deal much smaller?

Like a lot of those things we're gathering

from insights that we're importing or putting in

manually by hand on that record, on that

account, on that contact, on that opportunity.

And when you ask a non detailed, oriented person

to do that over and over again, if you

make it required, they can't skip it. Right.

So they're going to naturally put in stuff

that maybe they're not going to dig into.

Or I also want to give there might

be someone who could live in the details,

but their pipeline is so big, right.

They've got tons of deals. Oh, absolutely.

And they're working with various prospects

in different parts of different timelines. Right.

You might be talking to someone that's six months out,

so you're going to focus your efforts on someone who's

rearing to go, ready to sign that sow tomorrow.

Oh, for sure.

And I swear there's like a law for this.

When you throw an unlimited amount of something

into a variable, into a mathematic problem, there's

a theorem for something like this.

But if you give somebody a million opportunities

right, like, they're going to miss something.

And where you want them to focus

on is on the most important things.

So those are the fields that we make required

and where automation plays into this is, well, if

I can figure out the industry automatically or their

revenue automatically, let me automatically go get those things

so that person doesn't have to fill them out.

Not to mention if they're filling them

out, then we risk having some duplicate.

Or subjective.

It's subjective, right.

Like this is kind of nonprofit, but they do a

little bit of it's subjective and or they misspell something

or it's in a especially if you've got free form.

Good night.

Those free form fills.

I lost all of those spelling bees and the only time I

won was when I took a peek at the answer sheets.

You know what, we were talking about where I

was from before we started recording this episode.

Oh, yeah.

Federal Washington shout out.

I actually did have a time where I was on the stage.

The spelling be champion. Wow.

Third grade, Mel Bell, Sherwood Forest. Wow.

Check you out. Yeah.

However, I am not a fan of the still. Right.

Like, I can go up against the best of them

on spelling, but if we can maybe do it better

with a pick list or a multi select and it's

not so much that somebody would type it wrong, it's

that they would get it wrong naturally.

Like, well, I picked the wrong one.

Or they didn't know, or they misunderstood.

And so if we can go back to a source

that is a truth on the matter, or they have

the same way of thinking about it every time. Right.

Well, now they picked that for a reason.

They might be a nonprofit, but they might also be SaaS.

Well, that company is going to delineate.

And these are the Rolodexes, right.

The Zoom Infos, the Apollos, all of

those constant contact, even we use HubSpot.

Even HubSpot to an extent has some sort

of database that they're pre populating some of

these fields whether or not you are overriding

them with another data provider. Yeah.

Well, shout out to apollo because their API is

great and they have great insight into these companies.

But for the most part, and I feel

for a lot of them, things change, right.

You're a $10 million company today, but tomorrow

I just closed a $5 million deal and

now we're a $15 million company.

And some of those are just inherent challenges

and you have to deal as far as

automation goes, you have to deal with that.

Are you speaking to the data integrity?

Yeah, data integrity, right.

If you want to go back and have

insights into the data from last year, well,

that was last year got you, right? Yes.

They either grew or they died.

They probably didn't stay the same.

So taking it back to the sales process

and at least how we've run it and

grown and evolved with our automation.

So before we even get down the road to what

AI can help us do, where do you think, whether

we're talking about Ven or the clients that we work

with, where are those key moments that automation can take

that salesperson from good to great?

Because you have eliminated the manual data entry or

the search function that now these tools are.

Yeah, I mean, think about it this way, the more

context you have, the more likely you're going to make

better decisions, you're going to ask better questions.

One thing that really has stand out even among

some of the things that we're doing mel internally

is, well, what systems are they using, right?

From an integration perspective?

Well, they're using intact, not NetSuite.

So if I come into the phone call and say,

man, I hope you all love NetSuite, they're going to

be like, that's really awkward because we're on intact.

Having that insight and that context on the front

end gives you a lot of not necessarily expertise,

but you come in knowing who you're talking about,

goes back to that recruiting front.

Like if you call out like, hey, I saw

you did some time at Capgemini and saw you

worked at Tesla, man, how was know that's?

Just like that, hey, they know who

we are, they know what we like.

Hold on, how did they know that we

use know like they know about us.

And when you provide that context to

a, like they're going to utilize that.

It's instant connection material. Yeah.

So to kind of paint a picture for those

listening, very basic example, a lead comes in through

the website and we take it in through HubSpot.

We kick it back into Slack so that our

sales team can see first name, last name, company.

We can pull in really any

field, any property that we want.

So talk about like if we enrich the

company data with some technographic information from Apollo,

we want to pull that in.

Obviously we can see where they came from, the last

page that they visited, all of that within Slack and

that's just out of the box integration right.

Between HubSpot and slack.

I can just tell it what fields I want to

send so that y'all are a little bit more informed.

And when the salesperson goes and picks up the lead

to your point, they can now click out into a

LinkedIn profile or we can pull that information all right.

Into Slack. Yeah.

It's all about the context and giving the

individual an opportunity to make a connection.

We talked about relationships.

If I can relate to you on an individual

level with something that you've done in the past

or something that you've done or you want to

do now, we have something to connect about.

People want to work with or be around.

Individuals that look have done do

the same things as they do.

And if we can connect about kayaks or automation

or, I don't know, softball, I know you ebike.

Do you kayak? No.

Well, I mean, maybe like, one time.

You don't, like, own a kayak.

You're not like, an avid kayaker.

That's exercise like keyboard, you're waiting

for that to be automated. Yes.

Let me know when that happens.

But when you provide that context right.

That ability to connect and you do it in

an automated fashion, that sales guy just now is

one step closer to making that connection and avoiding

that awkward silence of like, hey, where are you?

Oh, okay. You're from canada.

I've never been there. Okay.

Where did you grow up?

Yeah, victoria.

Where is that? There's.

If you don't have any kind of way to connect

and you're not really good at connecting or finding those

connections, it makes that relationship much harder to build.

And you've got maybe two minutes at the front

end of a phone call to make that quick

connection and then pull that into other topics of

discussion as you're talking about business. Right.

Well, not to mention we also

automate the outrage to an extent. Right.

So if someone comes through, we're kicking out

an email to them with that rep's calendar.

And if they book, we have some parameters around it.

You can't book like, 30 minutes from receiving the email,

but if that rep is back to back when they're

in calls and they're coming up against that meeting, oh,

wow, someone booked on my calendar for this afternoon.

Again, servicing that information

within Slack absolutely.

Where we live, that only helps some of this stuff.

I think we've been talking about the

front end and building that relationship quickly,

having some of those insights.

But what about when we get past that?

We're moving them through the pipeline.

We're demoing now we've moved to close one

or close loss, the insights piece of it.

We were just talking about it before the recording

and how even today we're doing some analysis on

our own data in salesforce and running reports.

And wouldn't it be great if we had the ability to

ask a tool, hey, what were in q Three of 2022.

What was our best selling service or integration?

Pre built integration by rep, I mean, there's so

many things that if we had the ability to

mine our data in that way and I know

there's tools out there like we talked about.

Tableau, I think was one of them.

But these AI tools are changing the

conversation around that and how quickly we'd

be able to go find that information.

I'm going to have to give a shout

out to the company you're referencing to. It's.

Ask Edith young Startup.

They've got quite a bit going on, but what they

have done is mastered the large language model to interpret

a natural they've interpreted a natural prompt into SQL.

So SQL is a way to it's a database, right?

And you can run a SQL query

to query the data that you need.

And this is all happening in the background.

So you ask a natural question, how

many opportunities did I close in 2022?

It will then generate the correct query, the code

almost, if you will, to grab the correct data.

It will then deliver the data in a

table and paint a picture in a graph

without you having to go into your CRM.

Oh, it's magical.

Build a report. Yeah.

And you think about from I'm going to quote air

quotes here, traditional, like, you go into Salesforce, you have

to think, well, okay, it's opportunities right now.

I got to pick the right fields,

I got to pick the right filters.

And 30 minutes later, you finally kind

of whipped out with Ask Edith.

They are taking the natural question, translating

all of that, and within, I don't

know, call it a minute, right?

They're delivering the results that you're looking for

and painting that graph for you automatically.

Okay, so we're a salesforce shop.

We've been a salesforce partner for a long time.

We implement it, customize it, automate it.

Yeah, we've got folks, some really smart people

here that do this all day long.

You can't can't be telling me though

that Salesforce is not investing in, right?

Like, to what extent are we

we've talked about this before.

Before we go build it and before we go

start connecting apps and using OpenAI, we don't want

to replicate functionality that all of a sudden Salesforce

or HubSpot is going to roll out. Oh, absolutely.

What do you know about Salesforce

and what they're doing with AI?

It's funny that you're bringing this up during

our recording session because I literally have one

of the pages pulled up from one of

our team members who talked about this.

And Salesforce is quickly iterating on the GPT

large language model to pull it out and

deliver functionality across all of their products.

And so it's only a matter of time

before they perfect it and make it really

good and deliver something like Ask Edith.

Maybe the core difference there is they're going to

charge a whole lot more than maybe Ask Edith

will you have to use Salesforce, obviously, if you're

going to be wanting to use those things.

The other challenge that Salesforce is going to

run into is that people are going to

want to look across multiple databases. Right.

So a lot of our organizations, including

us, we use HubSpot Salesforce intact.

Maybe they're using Stripe, maybe they've got 70 of

the other tools that a lot of these studies

are saying that everybody's using X amount 242.

Yeah, a lot.

That's what we learned at Inbound.

I saw Inbound last year was the CEO got up and said,

hey, the average scaling company has 242 SaaS apps a lot.

Really?

I think we do a fairly decent job of kind

of auditing our tech stack just from a that's a

good practice for any business, especially growing and scaling.

You want to understand user count, you

can understand which licenses aren't being used.

But I think that blew everyone's mind.

And even today when we continue to kind

of leverage that stat out there and I'm

sure it's changed even since then.

Well, Salesforce is the 900 pound gorilla,

the billion dollar company that a lot

of people have never heard of.

And unless you build your entire business on Salesforce,

they're going to come out with these capabilities.

But it's really for the average sized company.

They're only going to be able

to look at their salesforce data.

I'm sure they'll come out with functionality

that allows them to connect to more

things, but it will be prohibitively expensive.

Right, and that's why companies like Tableau

and Microsoft, Bi and all these different

bi companies are proliferating because they can

look across all of these different databases.

Well, Ask Edith can do the same thing, but

all you have to do is type A.

When I, when I think about running reports in Salesforce,

I'm like, well, I have to have the exact data,

I need to know the exact right, and I need

to talk to that in person and what their expectations

are because otherwise I'm going to miss it.

I'm going to deliver something that they're not.

You just barely missed it.

Yeah, I know.

I've run into the same thing, building custom reports

in HubSpot where it's a lot of trial and

error, trying to figure out exactly I know what

I want to find or the question.

But it's just like pulling in the right

properties and make sure you're slicing it the

right way and you have the right filters.

Yeah, well, that's what is really cool in

some of the demos that I've seen.

And we're starting to figure out

what this partnership might look like.

But they're utilizing this large

language model to quickly interpret.

Okay, I've got this set of data and these set of

fields and of then to answer this question, I need column

one, three and five and boom, here's the answer. Right.

For the average person that's going to

take 30 minutes to figure out.

I mean, I think this is going

to help our sales teams, right?

I don't think we're replacing people

with these bots anytime soon. Absolutely not.

At least not on our side.

It's not vetting strategy.

We're certainly not going to replace our

sales team with any of these AI. I don't know.

Are they bots robots?

I mean, it depends on how you think about it, right?

I would think about it as a bot if

you're going to chat with it, but if it's

preemptively providing me answers, then it's advanced analytics.

Sure.

All right, so let's jump into what is in the headlines.

Headline number one, sales

Industries always be Closing.

Mantra could get boost from AI.

I mean, I think that's just attention grabbing.

Everybody likes to say that always be closing.

You got to click the link right?

To get the ad revenue. Yeah.

I'm actually going to revert this back to you

because you can't close unless you got a lead.

How do you think AI is going to

pump up that lead funnel at the beginning?

Because if we're being totally honest, I can't close a

deal unless we have a lead in the door.

And lead in the door only

comes from the marketing front.

So how is AI going to help me close something?

By getting more leads in the door.

Yeah, well, I think some of these tools,

especially if we are able to go implement

them, let's mine our data quickly.

Let's understand exactly where our leads

are coming from and double down.

This is an exercise that marketers

do today have done historically manually. Right.

You're going into your various databases.

Usually it's a CRM.

And hopefully the lead source is filled out if

at one point in time, because at one point

in time, previous roles, you're relying on that salesperson

to tell salesforce that that account and that contact

came from a trade show.

Remember all the trade shows they go to?

Did you put the lead in, did you upload the list?

I mean, these are the types of things that

as we continue to mature our marketing muscle, right,

and understanding exactly how we bring leads in.

Because historically we've been and still today our

strong referral base, but still trying to understand

how to reach those people that maybe aren't

finding us via the referral networks.

So I would use it to first mine the data we

have we're getting by asking it of each of our segments,

what's our strongest performance, where are we getting the most organic

search leads, what pages are they looking at?

We can do some of this in HubSpot, right?

And we can go build some reports in salesforce, but

I would start there by quickly asking those questions and

then of course we would go, how do we reach

them, we're going to generate more content.

I'm not going to go to Chat GPT and ask it

to write me 15 blogs on stripe and intact integrations?

No.

I'm going to go interview

our subject matter experts here.

After doing a little bit of research, I might

ask it, hey, what are people searching on?

What are stripe users?

What's the biggest functionality? I don't know.

You might do some preliminary research on using chat

JPT but then I'm going to go do some

internal analysis, do some interviews, maybe talk to some

clients that are using it that it's been deployed.

And then I'm going to pump those

transcripts into OpenAI into our own box.

That's what we're using, right? Right.

And then I'm going to go create content.

And that right there, it's a volume game

because that's where and I'm kind of jumping

ahead because we're not talking about marketing yet.

But when you ask me how am I going

to deliver more leads, my challenge is always velocity

volume and then the variation of it.

And so how am I going to do that quicker?

Being able to get those pieces out faster.

And you've heard me say it.

Big fan of utilizing your

own transcripts for this stuff.

So it exists today.

Go take your case studies and your blog content now

and go pump it in there and then figure out

how to optimize it or iterate on it.

And I think that will we've seen content bring

more leads in the door for our business.

So that's how I would start.

Just kind of like scratching the surface of it.

I have a question but the next headline

really kind of pushes into this question.

It's from government technology marketing professor uses

AI to coach sales professionals, right?

So this idea is we use AI to teach sales professionals

how to deliver better, you know, so on and so forth.

I want to take it a step further though.

I'm thinking about HubSpot, right?

And these drip emails, the sequences that run maybe

you're using salesforce marketing, cloud marketo, one of those

technologies to do some level of drip campaigns.

Maybe it's the sequences where the sales rep has

predetermined some level of kind of drip campaign that

is specific to him or her to that prospect.

And you were talking about variation.

What if the AI took what the salesperson wrote and

had a B and C and it automatically generated variations

of the content without in this scenario I'm going to

go without oversight and it is self testing variations of

the email to get more clicks, to get higher click

through rates, whatever that ends up being.

Where are you on that? Is that a good thing? I hate it.

I love it.

Let's do it.

Should we code it right now?

Where do I sign?

No, I think it's a great I mean I'd be

lying if I said I haven't been experimenting with some

of the features around rewriting for tone, right?

So I think if as long mean look, we've got

emails that I've partnered with sales to you know?

Hey, I'm Mike.

I got your inquiry from the website.

You're looking for this integration book time.

Where do I sign? Right?

Like, some of these things are and it's

we're converting, we're getting meetings booked, right?

If they raise their hand, we don't need to

oversell in that first email, that first touch.

Now, if you're talking about losing, like

someone goes radio silent, this happens.

We kind of call that lead

going into purgatory, so to speak.

I do think that I've heard Marketers last week.

I was on a webinar marketers talking about

how, hey, my third touch was not it

was just falling second, you know, killing it.

Third not.

So put.

That marketer put in all four touches and

asked Chat GPT, why is this not working?

It recommended a rewrite and her clicks went up.

Her opens went up. Really? Yeah. Interesting. Yeah.

I connected with her on LinkedIn

because I was like, that's cool. Yeah.

I was like, I need to learn again.

I think in general, we're doing this podcast because people

have said we want to know more about AI.

We want to know more, too, right?

Let's open up the conversation.

Let's get people talking about it.

So, anyway, that's one example

that I've not personally deployed.

Just came off that webinar last week.

But I'd love to take our existing Drips and

do the same thing to better understand that.

You know what's really cool about this time in history?

And I'm going to go back to my gaming days

when a new game came out, you wanted to be

not the first to play because it just came out.

You wanted to be the first to play because nobody else

had any insight on how to get to Level 50. Right.

Or to my husband's, still like this, to be clear.

Oh, totally.

I mean, the new Zelda right now, he's all over it. Yeah.

There's no cheat guide.

Nobody knows how to do it.

And we are in that moment in time in

history where nobody in the business world has like,

oh, yeah, you got to do this.

And right here's the master guide.

And for 90 grand a week, we'll tell you how to do it.

Everybody is on the same page.

And that's what I love about this particular

moment in history, because everybody's trying to figure

out, well, what if you did this?

Or what if you did that?

And like, oh, check this out.

And I really love this moment in time.

It'll be something that I look back on when

I'm retired, if I ever get to retire. You will.

That people were like, that was that one

moment in history where it wasn't passed down. Right?

It was brand new. Right.

Nobody knew what they were doing.

I do think this is a perfect time to plug.

So we try to get down to, like, what

is a use case that you as a person

listening to this podcast can take away today.

So if you're a sales professional listening to

this, don't sleep on rewriting the emails.

I have seen your emails, friends. Okay? Literally.

Literally.

And I appreciate your passion, but

sometimes we don't use complete sentences.

At bare minimum 101.

Yeah, go right.

You and I actually, we've gone back and forth on Slack.

You've had some prospects that you're trying to know

some activity out of and talking through different approaches

and we're screen sharing and you're pulling up Chachi

PT and you're asking it, what about ashamed? Yeah.

Here's what I'm trying to do.

I want to create some sense of urgency.

Here's my deadline.

When I'd like for this to close or to make

movement, I want to get them to do this action.

Can you write it in a friendly, authentic way? Yeah.

That's a huge starting point versus sitting

down and writing that from scratch.

So if you're listening today and you're in

sales, like I said, don't sleep on taking.

I know how long it takes to write

content and emails, especially if you're trying to

solicit an action, a call to action.

I think that that is a perfectly good use case

to get your feet wet if you're not using it.

Absolutely, I'll tack on to that.

Don't be afraid to iterate too.

From a manager to your direct report, it's difficult

to be like, missed the mark just a little

bit because you don't want to be chintzy, right?

Like, yeah, it's good enough, that's fine.

But with these large language models, they don't care.

Just be open and be like, let's revise

that and replace this word with this.

Can I say I found myself

being overly polite to Chat GPT? Am I the only one?

Please write this for me.

I also actually want to thank it and

I don't know if there's a good way.

I found myself interacting with it as though

it was a person and that's just in

my like, naturally, I would say they're providing

me this intense response in 2 seconds.

Thank you so much.

Can we just maybe tweak this thing?

And you're right, you can just tell

it make it yeah, make it witty.

I've heard and seen in different articles that we should

be polite to these bots because they will train on

that and they will then naturally be polite back.

I don't know which way that's going to go.

It's like, should I thank it for doing

something that it was designed to do?

And it A, doesn't really care, and B, it's

actually costing you more money to write thank you.

Because with the context, right, these tokens that

it's digesting, that it literally costs you more

money to say thank you, period.

But at the same time, all of

this, it's like emulating a human.

So you're naturally inclined to be like, hey,

thank you so much for helping me.

Would you please change this just a hair.

And there's this. Natural inclination.

And meanwhile, you're like, now I'm never going to

forget that when I'm messing around with you. Right?

We're getting charged tokens because

we're using that open API. Yeah.

Okay, well, fair enough.

I don't think it hurts to be polite,

but hey, we'll keep the tokens in mind.

All right, let's move on.

We've talked through our headlines.

Let's do a hot take.

These are always fun.

How does AI add value to sales intelligence?

I guess we've kind of been talking about this.

We kind of let off with that, right? Yeah.

I feel like this is one of those

areas where people have already been doing this.

Salesforce has Einstein HubSpot has multiple

scores that you can look at.

What is being left behind here is

that, how do I do this?

For use cases that are not top of mind.

Salesforce has opportunity scoring.

It's like, okay, well, it's 90.

What does that mean again, for the

average Ae that's just stepping in. Okay, it's great.

90, is that an A?

Does that mean they're going to close?

Where I think this really will play into is

providing intelligence to answer questions that people naturally have,

like, hey, how did that call go? I don't know.

It went pretty well. Right.

But this was my first call, so I have no idea.

But where we have these transcripts and we can

look into that and we can do the sentiment

analysis, we can look at the activity and we

can answer these questions that people naturally have that

they're answering to their direct reports or vice versa,

we're going to have that intelligence level to be

able to call out these really unique things.

And ultimately that's going to make you more productive

because you can either close, lose this opportunity really

fast because you didn't have I'm stepping back.

One of the questions you can ask is, did it

sound like we had the decision maker on the phone?

Well, the chat GBT doesn't really know, but

you can tell based on sentiment that this

person was really nervous about making decisions. Right.

Somebody that's the decision maker is not going to be

nervous about making a decision for the most part.

So there are questions that we naturally ask in

our head or our bosses ask of us. Right.

And those are the areas where

we should be focused on it's.

All of those sales questions on, like,

was this going to close this month?

Is it going to close next quarter?

When is this going to close, Mel?

And that's like, everybody asks that question because everybody

wants to know what the number is going to

be and you don't know until it's too late. Yeah.

Am I on pace for the year, too?

Or there's probably like quota things too.

I know there's reports and dashboards that

you can build out, but just questions.

You were talking about this, like,

the percent to close, right?

Like, how many do we close when versus close loss.

Push rate by. Yes. Thank you.

That's what I was talking about.

Close rate by.

Sales rep by quarter.

Maybe they were having a great year,

and they're more likely to close.

Well, that should influence our

forecast or our predictability.

That's more of just maybe slight calculation, but

we can utilize that to pair people up.

Like, hey, these two guys sell the same things, right?

Let's use some AI to determine, well,

who's going to be better, right?

Or we can do some natural investigation of

the data to determine and predict things that

we weren't otherwise going to be able to

do without this level of intelligence. Yeah.

All right, I think that wraps us up for time

today, but I am still super curious about how other

people are using this in a sales function.

So if you're a sales professional out

there, please send us an email.

It's okay if you use Chat GBT to write it.

Super interested in all the ways that you

are either using it, thinking about using it,

maybe your organization is already rolling it out.

I know there's lots of tools out there, too,

that we haven't even scratched the surface on.

So shoot us an email at thejunction@bentechnology.com,

and we'll catch you all next time. Ciao.