Make The Most Of Your Data Using Coefficient AI

0:00

[MUSIC]

0:05

Hello everyone and welcome to Go to Market AI.

0:07

It's YouTube and Go to Market Techstack.

0:09

I'm your host, Sarah McConnell.

0:11

These days it seems like every company has AI,

0:13

but on this show we want to go a level deeper so you can see

0:16

first-hand how businesses are actually applying AI to solve your business

0:19

challenges.

0:20

We're going deep into the use cases and showing you live demos of

0:23

the latest and greatest in AI technology.

0:25

Today, I'm joined by Ben Davis,

0:27

head of revenue at Coefficient.

0:29

Ben, welcome to the show.

0:30

Sarah, great to be here.

0:32

That was an awesome intro that gave me goosebumps.

0:35

That was great. That was powerful.

0:38

Now I'm excited. I'm coming to you from our new Coefficient Global Sales

0:43

Headquarters here in Austin, Texas.

0:45

We just got a new office space here,

0:47

so I'm taking one of the conference rooms for a spin.

0:51

Let's hope it goes well. I'm really happy to be here.

0:54

Very cool. Thank you so much.

0:56

First of all, Ben, can you tell me a little bit about who is Coefficient,

0:59

what do you guys do, and then who are you hoping in the market?

1:03

Yeah, Coefficient is a Series A SaaS company.

1:06

Classic venture back SaaS company. We're about 50 people now.

1:10

We did our Series A about a year and a half ago.

1:12

Our product is really, really simple at the core,

1:15

which is aiming to get data from your business systems in a spreadsheet.

1:19

That's the deceptively powerful workflow,

1:22

and I think anyone who works in spreadsheets for their job,

1:24

which as many people, will immediately see the value that Coefficient can

1:27

provide to them.

1:29

We work with the range of companies all the way from startups.

1:31

Qualified has been a big, happy, long-term customer of ours.

1:35

Shout out to Kieran on your RevOps team for that.

1:38

But we work with everyone from venture-backed startups all the way up to large

1:43

Fortune 100 companies.

1:45

We see this problem as a universal problem and an opportunity for us to really

1:49

insert ourselves in that space.

1:51

We've also really invested in AI over the last year, in particular, alongside

1:55

our initial spreadsheet capabilities.

1:57

Amazing. Well, you mentioned it. We are happy customers.

2:01

I know our RevOps team gets a ton of value out of Coefficient,

2:04

but I have not seen the demo in a long time,

2:06

so I'm really excited to jump into the demo and see for myself and show all of

2:10

our viewers what Coefficient can do.

2:12

Yeah, let me hop into it here. I think I have my screen already up.

2:16

And to your point, we have really grown a lot as a product in the last year.

2:21

The website now mentions at Veronica's Cell.

2:23

We were on Google Sheets entirely up until about a month ago.

2:27

It's a really exciting for us to launch for them on Excel.

2:30

Almost half a million happy users in the Google and Microsoft App Store are

2:34

there.

2:34

We launched them product hunt with our new Excel launch a month ago.

2:37

We're number one product of the day.

2:38

So lots of good momentum here.

2:40

But today I'd love to dig into AI in particular.

2:44

I've got a hopefully great demo here for you all.

2:47

I'm going to walk into some of our core features here and ultimately tying

2:51

together

2:52

a few things in how we do AI.

2:53

Just to set the stage for us and what I'm about to show you here,

2:58

our product lives in spreadsheets.

3:01

And so we live as an extension to either Google Sheets or Excel.

3:05

In this case, I'm really comfortable with Google Sheets.

3:07

So I'm going to demo from Google Sheets here.

3:09

Going to show you a few things.

3:10

So first off, I want to show you how we've enabled people who don't know SQL to

3:16

query

3:16

their SQL databases.

3:18

In this case, so for like using natural language, it's a really powerful

3:21

workflow for people who want to access data from a SQL database, but don't know

3:25

how to write SQL.

3:26

From there, we'll go into some of our Salesforce importing functionality and

3:30

how we plot an ICP.

3:31

And then we'll also tie it together with some of the Slack alert functionality

3:35

that we've been

3:35

investing in lately.

3:36

But let me kick it off here.

3:38

So I'm in a blank or a slightly populated Google Sheets.

3:42

And to launch Go Fish, I'm just going to go up here to the sidebar

3:46

and to the extensions tab and launch the tool here.

3:49

First thing we want to do is actually pull from our snowflake demo database

3:56

using natural English.

3:58

And again, one of the things we see frequently, we serve business users who may

4:02

or may not be

4:02

technical, our primary personas are RevOps data teams, marketing and finance

4:08

teams.

4:08

Many of those people do know how to write SQL, but a lot of them don't.

4:11

And so one of the things we wanted to do with AI was enable people to pull from

4:15

a SQL database

4:16

using plain English.

4:17

So let me go to snowflake here.

4:18

You'll see you can worry with custom SQL query directly from your DB.

4:22

But I'm going to select this cheap PT SQL builder here.

4:25

Let's let it load here.

4:27

I've already got the copied query that I want to run, which I think is helpful

4:32

here.

4:33

But I really just want to know, very simple, I'm a RevOps person, I want to

4:36

know,

4:36

tops and accounts by number deals one.

4:38

That's a relatively basic SQL query Sarah, but I think for someone who doesn't

4:42

know SQL,

4:43

doesn't know the syntax of the database, that can take a little bit long to

4:47

learn.

4:48

So I'm just going to click generate SQL here.

4:49

And our tool is going to do its thing.

4:52

Hopefully spit out the right answer here.

4:54

And from here, you see a provided a great response.

4:58

Looking at county count number deals one from opportunities where one equals

5:03

true,

5:03

grouping by county sort by the number deals one looks pretty accurate to me.

5:08

Let's look at this preview here.

5:10

Looks like we're getting what we want.

5:12

I'm actually going to go ahead and import that.

5:15

So let's just query that and import that into the.

5:19

The tabs here, if I wanted to go one step further,

5:24

I think I demoed it here, I had to pull up to demo here, but adding a column

5:29

for some of the

5:29

accounts one, if one has changed the first column to the account name, you can

5:33

really get crazy with it.

5:34

Let me go back to this example tab here.

5:37

The whole goal of our product is to make key data accessible in spreadsheets

5:41

here.

5:42

I had another query that I ran right before this where I was pulling in the

5:45

number of deals one.

5:45

I was grouping it by country, summing the number deals one.

5:49

And because it's in your spreadsheets there, we're seeing that people really

5:52

get access

5:53

to that data in the medium that they're most familiar with.

5:56

You get a lot of people in Salesforce, Snowflake, HubSpot, who don't like the

6:00

reporting capabilities

6:00

of those tools and they want to really get that data in spreadsheets.

6:04

So that's the first part here.

6:06

I really wanted to kind of just show how we can have immediate time to value

6:09

with a GBT to SQL and PowerBuilder query, Snowflake directly isn't coefficient.

6:14

Yeah, that's amazing.

6:16

To your point, if I, you know, I'm just dangerous enough in an Excel in spread

6:21

sheets and just

6:22

staying enough in Salesforce to break things, but not to really build anything

6:26

good,

6:26

but how fast are you able to do that?

6:27

I'm like, oh, if I wanted to build that report in Salesforce, like, yeah, I

6:31

could do it,

6:31

but it's not going to have it in the best output and it's not going to be in

6:33

the best place for me.

6:34

And to get it into a Google sheet, I would have to know SQL, which I don't.

6:38

That's not, you know, you have to manage on perspective.

6:41

It's not on skill.

6:42

But like, I've been able to do stuff like this using co-efficient.

6:45

It's so fast and it's simple to bring in this data that I need to know for my

6:48

day-to-day job.

6:50

So yeah, it's cool to see in how fast it is.

6:53

Yeah.

6:54

And then once you have it in here, obviously, there's a power of co-efficient

6:56

behind it.

6:57

You could do things like set the refresh schedule, maybe want this report to be

7:01

live every day

7:02

at 6 a.m. or weekly.

7:04

You can obviously go and edit that custom SQL.

7:07

We do also allow you to write back to your source tool if I wanted to actually

7:12

maybe edit

7:13

some of those rows and snowflake.

7:15

That's a less common workflow, but people are editing things in Salesforce, Hub

7:18

Spot,

7:19

with their tool all the time.

7:20

You can configure right backs.

7:22

Watch the show does the second year with another tool.

7:24

But let's just start this query.

7:26

Let's kind of back out.

7:27

Again, the main value prop of co-efficient is being able to find data from

7:32

multiple tools here.

7:33

Just showed you all a snowflake demo.

7:35

Wanting to take this one step further and use our AI features to do a little

7:40

bit of data hygiene

7:41

and data analysis.

7:42

I think it's very common flow for both marketing and sales to take a data set,

7:47

scrub it,

7:48

clean it, make it usable, clean it up a bit.

7:50

That's historically a very manual process.

7:52

What I went ahead and did was pull in from Salesforce again using our tool,

7:57

just a dummy set of contacts.

8:00

We have imaginary people at accounts.

8:02

So we've got recognizable names here, but completely made up roles in those

8:07

accounts.

8:07

And for this part of the demo, I actually wanted to show our GPT powered

8:12

formulas.

8:13

And in this case, I'm trying to, let me exit out of the sidebar so we can see

8:17

better.

8:18

Trying to kind of scrub this billing address.

8:20

So I have this column.

8:21

It's kind of messy.

8:22

It's not formatted right.

8:24

Like everyone's different.

8:26

It's just a little clunky.

8:27

And I really want to tie that to a region.

8:29

In this case, the city, the country, and then I want to summarize the details

8:35

of those contacts.

8:36

So I'm using our GPT powered formula here.

8:38

Pull up the region name, pull up the country.

8:41

And then the cool step here that I think is really powerful for anyone in Rev

8:45

Ops who's watching

8:46

is pouring these leads.

8:48

And so actually you have this ICP score.

8:50

If we go to our tab over here, we're powering that ICP score in the other tab

8:55

with actually GPT formulas.

8:57

In this case, we're referencing the website.

9:00

We're looking up and we're telling these GPT formulas to write a value

9:03

proposition for that website.

9:05

Nothing on here is inputted by me or Frank.

9:07

It's all pulled by AI from this website.

9:10

So it's querying coefficient.io pulling the value prop.

9:13

Also pulling the ICP, which is pretty amazing.

9:16

So it's going and summarizing what it sees on that website,

9:19

pulling in information about this.

9:21

This is pretty accurate.

9:22

It looks like finance, marketing, BI, data scientists, business owners,

9:26

operations,

9:27

really accurate profile.

9:28

And then we go back over here and it's taking that input and it's trying to fit

9:33

this person.

9:33

In this case, Stinthie Hesselwood at Acme.com.

9:37

She's a BI manager trying to fit Cynthia into a ICP.

9:42

Gives her a score seven.

9:43

Let's go ahead and drag these formulas down.

9:45

And again, we're pouring these GPT formulas there,

9:48

pulling from our tool into the GPT database and trying to fit these contacts

9:54

into an ICP.

9:56

See, we get some summarized details, some countries, some ICP scores.

10:00

Few fits here.

10:03

So we've got Uber, we've got a mechanic, we've got RevOps manager,

10:07

consultant fashion designers, pretty good fits.

10:10

And I'd say just an eyeball it here, retail.

10:12

Yeah, that's not as much a common fit for us.

10:15

So it's good to see that score at six.

10:17

Ten manufacturing, it's a pretty good fit.

10:20

Senior analyst, a pretty good fit.

10:22

So pretty good job by the formulas here to fit into an ICP for us here.

10:28

And again, salt powered on the back end by these GPT formulas.

10:32

Yeah, this is really cool.

10:35

Because I think what always gets me with ICP formulas,

10:38

if you talk about like MQL scores or anything like that is,

10:40

it's kind of you taking a stab in the dark.

10:43

Like it's you kind of guessing, you know, you're right,

10:45

but like you're kind of making the formula up.

10:47

So to your point of like, even if these GPT formulas can get you most of the

10:52

way there,

10:52

and like what you see is so much easier, like just scroll through this and eyeb

10:55

all them and say,

10:56

like, oh yeah, this is pretty right.

10:57

Like it's getting you so much farther a step in the right direction versus you

11:01

trying to think

11:01

of the right, not only formula, but like, okay, how many points do I get?

11:05

For title versus how many points do I get for country and like,

11:08

why wait them?

11:09

Like this is going to do most of that work for you.

11:11

And that is you just a ducking process.

11:14

So yeah, that ICP score is I can see where that would definitely come in handy.

11:18

Yeah.

11:18

And I think, you know, we're all we've gotten more use, I think, to using

11:22

things like chat

11:23

DPT in our day to day as well as our respective tools.

11:26

And there's always that human element in it.

11:28

And you're never going to get 100% of the way there with the pure AI formula

11:31

that you can get.

11:32

Cut out 95% of the manual work, at least that saves your time.

11:35

To really take this one step further, I wanted to showcase on the power of our

11:39

tool,

11:40

which is being able to export things back to Salesforce.

11:43

In this case, I have this ICP, I have the region, the country,

11:47

details, those three columns are probably already in Salesforce.

11:51

But if I want to take that ICP score and edit it in our Salesforce instance,

11:55

that either requires a typical tool like a data loader going into a manual in

11:59

Salesforce,

12:00

you can actually export directly to Salesforce from coefficient here.

12:04

So I've set it all up.

12:05

I've mapped the fields, contacted the ICP score.

12:08

And with one click, I can automatically update all my records in Salesforce

12:13

with that AI powered ICP score.

12:15

So again, so, so easy for RevOps or BDR manager or Sales Manager,

12:21

really anyone to get to a high conference value and then act on that in their

12:26

tool choice,

12:27

in this case, Salesforce.

12:29

So cool.

12:30

Yeah. And then final thing here, I kind of want to bring this full circle.

12:34

We've done a demo, we've pulled data with coefficient, we've exported it,

12:37

we've done AI formulas, we've done AI summaries, GPT SQL to English SQL.

12:43

I want to bring this full circle because we really see RevOps teams today,

12:48

trying to give their teams better usage of data in the tools that they're in

12:52

all day long.

12:53

For us, the CRM, it's also Slack, it's also some other tools.

12:56

But I wanted to set up a Slack alert here.

12:58

And you see how this automation is in my sidebar here.

13:02

And I set this automation up to look at the sheet.

13:05

And whenever we have a new ICP score above seven.

13:09

So again, whenever something matches to whoever might be great ICP for our

13:13

product,

13:13

I told it to send an alert to our Slack channel, actually have it up right here

13:18

And yeah, actually, it went just a few minutes ago when we updated that,

13:22

it sent the alert in here.

13:24

This is really customizable.

13:25

So I just have this showing all the rows here, but if you wanted to

13:28

highlight a lead or highlight, you know, Cynthia at Uber or Acme wherever she

13:33

was,

13:33

you could set that up.

13:34

And you could really automate this.

13:36

Our tool lets you pull in the data, automate that pull, automate the refreshes,

13:40

automate formulas.

13:41

Our goal is to really let you set it and forget it.

13:45

And again, ultimately get that data and all the systems that your team is using

13:49

and loving on a day-to-day basis.

13:52

So if you had a circuit demo, you wanted to go ahead and yeah.

13:54

Yeah, no, we use one of the things that we've really enjoyed using coefficient

14:00

for.

14:00

And I think just for any marketers that are listening to this episode and

14:03

giving it that,

14:04

again, that full circle moment is we pull a lot of our data out of Salesforce

14:08

into Google Sheets.

14:09

And then we use it to feel like we run a pipeline council every Friday.

14:13

It's like a big thing to us.

14:14

It's where our marketing and sales get together.

14:16

We talk about how pipeline is performing.

14:17

We look at all this data for the week, for the quarter, for the year.

14:21

And we use coefficient a lot by pulling the data out of Salesforce and then

14:24

building.

14:25

All of these tabs and building are like our charts and everything off of that

14:30

data.

14:30

And then having it linked into a Google slide that we then run pipeline council

14:34

off of,

14:35

talked about the automations.

14:36

It's so easy for a RevOps team where they just have it one time that data is

14:43

automatically updated

14:44

every single morning.

14:44

And it's automatically updating this chart.

14:47

So there's very little like, it used to be such a manual process.

14:49

I used to we were much better.

14:50

I was variable, not manual process every week of trying to like,

14:54

look at that data and how do we get it out of Salesforce into a Google slide?

14:57

And how do we run that really effective pipeline council?

14:59

And having this integration with co-session, I know it's just made it such a

15:03

much more painless process for our team.

15:07

And it's a little bit easier than to visualize it.

15:09

So yeah, that's a use case that we've definitely found a lot of value from.

15:14

Yeah.

15:14

And I'll just hop out of this tab here in a moment, but you could easily have a

15:17

chart here.

15:18

This could refresh from coefficient.

15:19

It could be late to your sheet and click refresh hours back in your week.

15:23

You know, that's what we're trying to do really just to help you,

15:26

help you get that time back.

15:26

That's my demo.

15:28

Happy to show anyone more if they reach out, but I could showcase some of the

15:31

powers co-efficient.

15:32

Well, thank you so much for taking us through the demo.

15:35

Then if you are ready, I would love to jump into our Q&A section.

15:39

First question is how long have you been building AI into co-efficient?

15:43

Yeah, about a year.

15:46

About a year it's April, 2024.

15:48

So about a year, I think what it's yet to come out of the scene about a year

15:51

and a half ago.

15:51

It feels like it's been a year a while, but it's pretty new.

15:55

But yeah, we started shortly thereafter.

15:57

We've gone incrementally.

15:58

We initially launched an AI powered chart builder for things like charts and

16:03

pivots and all that sort of stuff.

16:04

And we've steadily launched a new thing since then, but we've been working on

16:08

it for about a year.

16:09

Amazing.

16:11

And then what you showed today is all of that available to co-efficient

16:14

customers.

16:14

Yeah, it is.

16:16

And if you're not a co-efficient customer, you can sign up, get a free pro

16:19

trial.

16:19

And when it signs up, it gets a trial of our pro plan with future all those

16:22

features.

16:23

All those AI features are available to everyone.

16:26

We don't get them.

16:26

They're not an enterprise plan.

16:28

They're available to everyone.

16:29

Amazing.

16:31

And then who are some of the current customers that are benefiting from co-

16:35

efficient?

16:36

Yeah, we have an amazing set of customers that I feel so thankful to work with

16:40

every day

16:41

and we get great new customers on board all the time.

16:44

Again, all the way up from small, scrappy to five person startups.

16:48

We've been working with a lot of customers in the Gen AI space recently.

16:52

All the way up to the Fortune 100.

16:54

Some of our most notable customers, we have case studies with them on our

16:57

website.

16:58

Miro, their RevOps team, uses co-efficient to really power a lot of their hyper

17:03

growth.

17:04

They're just such a rocket ship of a company.

17:06

Clavia, we also have a great case study up on our website with them.

17:10

They used us all the way from early days pre-IPO, but they're not a public

17:14

company,

17:14

asked success.

17:15

Twilio, Unity, Docker, Spotify, those are former logos that we listen to our

17:20

website there.

17:20

But again, it spans the gamut from small startups all the way up to large

17:24

sophisticated companies.

17:25

Amazing.

17:27

And then my last question for you is what's next on co-efficient AI roadmap?

17:31

Yeah, I think I could share most everything.

17:33

I think that really just going back to the core of what we want to do.

17:38

Ultimately, we want to get data in key business systems into a really usable

17:42

format.

17:43

For us, that it's typically been a spreadsheet interface, but also I showed you

17:47

all, it's also

17:48

slack and it's also being able to edit that data back in your business systems.

17:52

And ultimately, AI is a really empowering tool for democratizing access to that

17:56

to the data

17:57

because it lets people again continue on our theme, which is enable business

18:00

users to

18:01

make better use of that data without highly technical skillsets or go-out-the-

18:05

data team.

18:06

So you could see us really investing in a lot more spaces in this theme.

18:10

I think something that would be really great to do is being able to naturally

18:14

query,

18:14

kind of like I did with Snowflake there, but what if you could query all of

18:17

your tools?

18:18

And so if I said, what are my top 10 accounts and also show me all our

18:21

gone recordings and also give me the number of outreach emails or something?

18:25

And if we can magically stitch together your HubSpot instance with all your

18:29

gone calls,

18:29

with all your outreach sequences and get that one picture, you might be able to

18:34

do that if

18:34

you went to your data, it seemed that it would require SQL, it would require a

18:37

lot of work.

18:38

If you could just query it out with plain English, how powerful would that be

18:41

for sales, for

18:41

webops, for marketing, for everyone? So that's one area we'd like to invest.

18:45

Probably lots more as well, but I think that's top of mind for us at the moment

18:50

Amazing. Well, Ben, thank you so much for joining us on the show today. It was

18:54

great to have you

18:55

that was such a fun demo, so thank you so much for joining us.

19:04

(drum music)