Backchannel 34 min

AIs Role in the SaaS Buying Process


The B2B SaaS buying process has been broken forever. Some forward-thinking companies have adapted to reducing friction, but with new AI tools dropping every day, are we seeing the end of the inbound SDR?



0:00

So, okay, what's up?

0:02

How's it going?

0:03

It's going good, going good.

0:05

Another week in the books.

0:08

I wanna talk about something that happened to me this week

0:10

that I think is interesting and it was really a surprise

0:15

to me, to be honest.

0:17

PostScript has a, we have a big project coming up

0:21

that required me to do a couple of demo requests

0:26

on software companies' websites.

0:29

I haven't done those in a while because, you know,

0:32

people aren't, like we talked about last time,

0:34

people aren't buying software.

0:35

So, it's been a minute.

0:38

And in both cases, and look, I'm not trying to like,

0:44

you know, pump myself up here, but like,

0:47

I'm filling out the form, it's like, Mike Mannheimer,

0:50

here's my LinkedIn profile, like,

0:52

Chief Marketing Officer at PostScript,

0:54

like it's like, what's your budget range for this?

0:56

And like, tens of thousands of dollars, right?

0:59

In both cases, I got set up with the meeting,

1:05

probably slower than I should have been,

1:07

but got set up with the meeting.

1:09

And in both cases, it was an inbound BDR qualification call.

1:14

And I hit up Dan, my sales counterpart at PostScript,

1:20

and I was like, what, like, is this, am I crazy?

1:24

Like, I thought this was gone.

1:25

Like, we have all these signals, we have all this data,

1:28

like, what are you trying to qualify?

1:33

Like, I get it if someone's like,

1:35

you look them up there, like the intern or whatever,

1:36

like you gotta like ask them to be quiet.

1:38

But, you know, it's like, we're a large company,

1:43

it's raised $100 million, I'm a sea level executive

1:47

at the company, it's me, like, it's not like a fake person.

1:52

So, he was like, well, I think everyone still does

1:55

these calls, 'cause like, we don't do that at PostScript.

1:58

Like, somebody does a demo request,

2:00

we see that they're like this founder, the VP of Ecom,

2:03

at a qualified Shopify store, we put them on with, you know,

2:08

someone who can help answer their sales questions immediately.

2:12

Like, we only call if we need to,

2:15

but I could tell I was being forced through

2:18

like a mandatory call process,

2:19

I thought that was like a relic of the past,

2:21

like, have you experienced this?

2:23

What's your take on this?

2:25

Well, first of all, I'm still waiting for your form submission

2:27

to come through the audience plus...

2:29

(laughing)

2:31

No, I mean...

2:32

- And your form is quite funny.

2:34

- That was, oh, okay, that was funny.

2:36

No, I'm kidding.

2:38

You know, we know this guy very well close to the company

2:41

named Obed Durani, and he just posted this video today

2:44

that like, is very timely for this conversation.

2:47

Wouldn't it be crazy if when we ask people,

2:51

when we have a button that says talk to sales,

2:53

if you could actually talk to sales on the other side,

2:56

like, wouldn't that just be mind-blowing?

2:58

So no, it's really interesting, right?

3:02

Because, you know, I think the entire like role of the SDR

3:07

is really being challenged a little bit in this world.

3:11

Like back in the bull markets, back when we worked together,

3:14

it was like, well, you have so much volume

3:16

that you have to like qualify people through

3:18

and you need to run them through what band or whatever

3:23

before you get them on the phone with an AE,

3:27

because the scarcity there was the time of the rep.

3:32

I don't know that that's, I mean,

3:34

I could be proven wrong, but I don't know

3:36

that that's the current reality of the software sales reps

3:41

day in and day out today.

3:43

So I'm with you, I think, you know,

3:47

if you have someone, if you're able to qualify off of a form,

3:52

then why on earth would you introduce friction

3:54

to the process?

3:54

Like let's get this high-value contact

3:57

through the process to a rep as fast as possible.

4:00

- Yeah, and I'm not like, you know,

4:03

I understand not letting everyone get onto an AEs calendar,

4:07

even if the calendar has a bunch of spots on it.

4:10

But I'm saying like, I'm knowing what I know.

4:13

I think in both of these cases,

4:16

I am very clearly the decision maker

4:19

at an ICP level account.

4:21

- Yeah.

4:22

- And I have a bunch to say about the role of the BDR

4:27

in the future.

4:29

I actually think they're gonna probably be like,

4:31

less disrupted than people currently are saying.

4:34

But yeah, but I take like,

4:36

it's like a customer first approach to it.

4:38

It's like, you said with friction,

4:41

like you don't want to introduce friction

4:43

where it's totally unnecessary.

4:48

And we have all the data, we have all the signals,

4:51

you're qualifying through a form,

4:52

you can enrich with every piece of data ever made.

4:57

So like, why?

4:58

Like it was really surprising to me.

5:01

And I also had like,

5:02

a experience with one of these vendors

5:06

that was actually really positive to your talk to sales thing.

5:09

The BDRs I talked to in both of these qual cases

5:12

were very, very good, awesome.

5:16

It was fine, good experience.

5:18

- How long were the calls?

5:19

- The calls were both about 15 minutes.

5:21

- Okay.

5:22

- The problem, like they did a really good job.

5:26

I also think like there wasn't much information

5:30

for me to get that like they could give me.

5:32

I was like, hey, like looked at the website,

5:34

I have like a general understanding about how this works.

5:36

And then I would say what I thought

5:37

and they'd be like, yeah, that's pretty much it.

5:38

So we're gonna get you set up.

5:39

I'm like, well, what was the point of this?

5:41

It wasn't that they did a bad job.

5:42

It was like, obviously someone was telling them

5:44

that they had to do that.

5:47

But it was definitely a waste of time.

5:48

Like if I was like a jerk,

5:50

I could have like freaked out

5:52

and sent a scathing email to the CRO

5:56

or even worse posted a LinkedIn, you know,

6:00

hot take dunking on a poor and bound BDR.

6:03

I would never do that.

6:04

- Or quote a podcast episode.

6:06

- Yeah.

6:07

- They don't know.

6:09

- Not naming names on the show.

6:10

- Action, that was a safe.

6:11

- Yeah, come on.

6:12

- No, the BDRs did a great job.

6:14

It's obviously that they're following the process.

6:17

But I just think like the process was totally unnecessary.

6:20

And then I actually needed information

6:23

from one of these vendors faster

6:26

than I could wait for this call that they got set up.

6:31

So I hit up the rep directly on LinkedIn.

6:34

I said like, hey, I have 10 minutes worth of questions

6:36

that I needed to answer to them today.

6:38

Here's my cell text to me.

6:40

And he did in like one minute.

6:42

And he was like, I'm free to chat for five if you want.

6:44

And I called him.

6:46

We hashed it out.

6:47

It was great.

6:48

And like that was a really good experience.

6:51

And to your friend who said,

6:54

like what if you click talk to sales and you talk to sales?

6:56

That was like a way better buying experience.

7:00

I like built rapport with this guy.

7:02

He was available when I needed him.

7:04

He answered the questions that I had

7:06

and wasn't trying to like tell me his book.

7:09

He was like, what do you need from me today

7:11

to do what you need to do?

7:13

Very helpful, 10 minutes long, great.

7:15

We still have our call scheduled.

7:17

So now coming into that call,

7:18

like I'm more excited and happy

7:20

and I already helped me out.

7:22

So I don't know.

7:23

I just think like people should really think about

7:25

like getting to that faster.

7:27

And in the idea of protecting someone's calendar too,

7:30

like the calls don't all have to be 45 minute demos.

7:36

Like the buyers really sophisticated at this point.

7:40

Like get them what they need

7:41

to be able to help them make a decision.

7:43

You have to listen to what they want and be flexible

7:45

on the call.

7:46

But I think people are just like,

7:48

everyone goes to RKO.

7:49

You see the slide of like the process.

7:52

And it's like form to inbound BDR,

7:54

then inbound BDR call to 45 minute demo

7:57

where the first 30 minutes is like,

8:00

here's who invested in us.

8:01

Here's how we think about the world.

8:03

And like you need to educate your buyers sometimes,

8:06

but I think nowadays with all the information we have,

8:09

like it'd be a lot more flexible than that.

8:12

- Do you think that's what drove it?

8:14

- Yeah.

8:15

- You only an SQL until you were able to like have that call

8:18

and then there was no way because of some internal process

8:21

or funnel like exit criteria that

8:24

- Yeah, I think it's all process driven.

8:28

It's like there's a process that we need to follow

8:32

the CRM and the rules in it are driving the rep behavior.

8:36

And so like if somebody,

8:38

if the AE who was assigned to my account,

8:41

saw me do the demo request,

8:44

he would have to go like ask his manager,

8:46

Hey, you know, Mike, the CMO at the company

8:49

I've been trying to reach out to just did an inbound demo

8:51

request, can I circumvent the process to be locked in?

8:57

That's like a crazy thing to have to ask.

8:59

- Right.

9:00

- And it frustrates, I'm sure it frustrates the AE's

9:04

and the BDR's, right?

9:06

You know, what if you could go do this whole thing,

9:08

monetize this, this lead faster?

9:11

And it also frustrates the buyers, like, you know,

9:14

there's a lot of information on a website.

9:16

I can get a pretty good amount of it.

9:18

I can go on YouTube and probably see a recorded demo

9:21

of any software I want.

9:23

So I just think people need to challenge their,

9:28

you know, their process slides a little bit and say, like,

9:31

you know, where can we fast track some of these discussions?

9:35

Because if you move at the speed that's being dictated

9:38

by an informed buyer, you'll probably have a higher

9:40

likelihood of selling something faster.

9:43

- Yeah, absolutely.

9:44

- Like it benefits everyone.

9:45

I was just surprised.

9:46

I was like, oh, I didn't realize this like hard

9:49

and fast rule was there even in cases where it's like,

9:53

high ICP, you know, you're talking to the DM,

9:56

you've established through the form that there's budget.

9:59

So like we got, we got all the ingredients here.

10:01

Like let's just sell some stuff.

10:03

So I was surprised that's not what I was experiencing.

10:07

- Well, I want to ask you about your comment though

10:09

about like the future of the role itself.

10:11

Because I view a distinction between,

10:14

which I know these are just acronyms for everyone,

10:18

but like SDRs versus BDRs or whatever,

10:21

where like inbound or said another way,

10:24

in human beings that are inbound gatekeepers

10:27

to the sales team around inbound demand.

10:30

So whether that's a demo request

10:31

or any other like form of hand raisers.

10:34

And then BDRs who are human distribution, right?

10:38

They are people that are using tools to push the message

10:42

to your ICP, your target accounts or whatever.

10:46

Now, I think there's a lot of,

10:49

first of all, I think people are grouping these two

10:52

kind of distinct roles together

10:54

when they're kind of plotting their demise of sorts.

10:57

I don't think that's all together a good thing.

11:01

I would argue or I'm sitting, is that outbound

11:05

or the kind of BDR function,

11:08

as I'm defining it here,

11:10

is still very important, but pretty,

11:15

but it's changing quite a bit, right?

11:17

Like the channel mix is changing,

11:20

you know, is the boiler room kind of sales floor

11:25

kind of model effective in 2024 or anymore?

11:29

But I for one tend to believe there is something there still

11:34

for human beings doing outbound selling basically

11:37

into our target account list.

11:40

But I'm not so sure about the SDR,

11:42

maybe to your very point,

11:43

can technology, dare I say it, AI agents,

11:46

something help do a more efficient job

11:51

of getting a form fill like you at least,

11:56

maybe there's some data challenges

11:58

with really automating a lot of this

11:59

for someone that isn't identifying as

12:02

a chief marketing officer for a well-known company.

12:06

Do we need SDRs in the future?

12:11

- Yeah, I think there are definitely a bunch of cases

12:15

where the answer is like, absolutely yes.

12:17

And the reason I think humans need to be involved,

12:24

whether it's an inbound or an outbound sort of capacity,

12:29

is because like the hard thing to do is like,

12:33

what you're really going for in both cases is a response.

12:38

And it's surprisingly difficult, way more difficult

12:43

than you think, even an inbound,

12:44

like I can't tell you the amount of like people

12:47

who fill out a form or download our app.

12:50

And then you're like, hey, great,

12:51

like you look like a great fit,

12:53

like what questions can I answer for you?

12:55

Can I get you set up with an onboarding person?

12:58

Like we're trying to serve their request,

13:02

dark for a month, two months.

13:06

- Yeah.

13:07

- I don't think that you can totally automate

13:12

that process at least definitely not today.

13:14

You know, you need somebody who like understands

13:17

your business, can use the data to decide

13:20

who is worth trying to drive outreach to.

13:25

And then the whole point of it isn't,

13:28

is to create a response,

13:30

not just to distribute something.

13:32

You know, like, and I think that that's really,

13:35

that's really where the crux of, you know,

13:39

outbound and inbound prospecting

13:42

has to have the human touch in my opinion.

13:46

Because it's really about, you know,

13:50

you have to do the pattern interruption

13:52

and cut through the noise.

13:53

The noise is just whatever everyone else

13:55

is doing all of the time.

13:56

So like people are like, oh, AI's gonna fix that.

13:59

Well, when everyone's doing AI,

14:00

that's just gonna be the new noise.

14:02

- The noise.

14:03

- And so what you really need is people

14:06

who understand your business very deeply,

14:09

high business acumen,

14:11

and who have high creativity

14:15

that results in responses from prospects,

14:18

whether or not they're inbound or outbound.

14:21

That is going to be a very long time

14:24

until that's fully automated.

14:26

I think that AI can help all these things be faster,

14:30

they can help BDRs or STRs or whatever you wanna call,

14:35

call the title.

14:36

People who are doing prospecting,

14:38

whether it be inbound or outbound,

14:39

I think AI can help them a lot,

14:40

can make them a lot smarter, a lot faster,

14:43

can feed them insights that help them do what they need to do.

14:47

But I think, you know,

14:49

we're trying to get a human being

14:51

to respond to another human being

14:53

via creativity and business value.

14:56

For now, and for in my opinion,

15:00

for the foreseeable future, that's a human's job.

15:02

- Yeah, I think that's right.

15:04

I think that's right.

15:05

But it's interesting enough you saw the GPT-40

15:08

like demo or whatever,

15:09

with the support use cases of voice,

15:14

basically what you call an AI agent,

15:19

two of them talking together,

15:21

one on behalf of someone trying to return something,

15:23

and the other one is a support rep.

15:25

And again, this was a demo for whatever,

15:28

who knows what went into it.

15:30

But you can start seeing a window in the future

15:33

or at least that use case of a support,

15:35

the customer support representative,

15:37

which might be actually quite disruptive

15:40

to anyone working in a call center

15:43

from a post sales perspective, at least,

15:45

or from a BPO or something along those lines.

15:48

- I think it has to do with,

15:51

is there a direct path?

15:53

Like, when someone needs to make a return,

15:57

and then there's a return process.

16:00

- Yes.

16:01

- It's like a discrete thing, you know,

16:06

like we're trying to go down a path

16:10

that's been basically fully defined

16:11

because there's a set of steps that need to be done

16:13

to be able to make a return.

16:15

- Yeah.

16:16

- I see how AI can automate either all

16:21

or a big chunk of that process.

16:24

And I expect that that's happening today

16:27

for working with the orgs,

16:29

and it'll happen for a lot of orgs soon.

16:31

Like there's, if you, whatever industry you work in,

16:34

I can speak to this because in E.com,

16:36

I see it all the time.

16:37

There's at least a handful of people

16:40

who are attacking AI in support-related use cases.

16:44

And I think the reason that we're seeing it there

16:47

is because the processes that people need to do.

16:49

In E.com, it's like, where's my order?

16:52

I need to make a return.

16:53

- Yeah.

16:54

- I ordered the wrong thing, like all of that sort of stuff.

16:56

There's those kick off set paths

17:00

of things that need to be done

17:01

to resolve the customer's issue.

17:04

The AIs can be pretty good at that

17:06

'cause it's trying to move through a set of paths.

17:08

What I'm talking about with the BDR is like,

17:12

what, like the AI would have to say,

17:17

what does a CMO care about?

17:19

What does my product do that attaches to those things?

17:23

How is this CMO's particular company performing

17:26

in the market?

17:27

Are there opportunities for me to like,

17:29

give some business value to them

17:31

and cause them to respond to me?

17:33

That's just like a, there's no set path for that.

17:36

That requires, I think, a lot of human creativity.

17:41

- Yep.

17:42

- And I'm open to being totally wrong about this,

17:45

but I just don't see a world where it's doing all of that

17:50

and then let's call it the next like three years.

17:56

Could it be 10 years that's happening?

18:00

- Maybe.

18:01

I'm not totally sure.

18:03

But getting back to where we started this conversation,

18:07

like it doesn't require AI to see that like your ICP DM

18:12

at your ICP account filled out a HubSpot form on your website.

18:19

And like, it's just funny to me.

18:21

Like even this, even what we're talking about,

18:23

what about like one day, could you automate out everything?

18:26

People can't even get me on the phone with the rep

18:28

within the hour of me basically being like,

18:31

I have money here, like somebody talked to me and take it.

18:34

We can't even do that.

18:35

We're talking about like automating these like crazy,

18:39

crazy business processes.

18:41

So I think, you know, there's probably between now

18:45

and when AI takes your job,

18:46

there's probably a lot of opportunity for like you

18:51

to clean up your house, build a better business process,

18:55

do things that make the revenue cycle happen faster

19:00

and create better experiences for your buyers.

19:02

Right now, you don't need AI to do that.

19:04

That's just about letting your really smart humans

19:08

do the right thing for the customer

19:10

based on what they can see from the form submission.

19:15

I think based on my experience this week,

19:17

a lot of companies still aren't even doing that.

19:20

- So I agree with your broader point.

19:22

- Yeah.

19:23

- No doubt about it.

19:24

But I'll challenge kind of how you got there a little bit.

19:26

- Okay.

19:27

- From my perspective, like what I'm learning.

19:29

I agree that, or where I thought you were headed with that

19:35

was the difference between support and an inside sales

19:38

kind of call process or whatever is,

19:41

there's really not a need for creativity in post sales,

19:46

like, you know, read the sort of like standard operating

19:50

procedure, move the customer through the set process

19:54

and do what you have to do.

19:56

There's, I mean, again, with full respect to that role,

19:58

there's just not a lot of like room or margin

20:01

for creativity, it's not needed.

20:03

Whereas your broader point around from a quality perspective,

20:08

from getting a response perspective,

20:10

from an attention 1000%, completely agree with you.

20:14

What I've been surprised by,

20:17

really in the last couple of weeks,

20:19

is studying kind of what AI has been able to do

20:24

from a place of inference.

20:27

And this is the newness, I think,

20:29

with the latest models that's really interesting,

20:32

is certainly they can,

20:35

what I imagine could happen is when somebody fills it,

20:38

when you fill out that form, they have your name,

20:41

they have your title and they have your company, right?

20:45

Amazing.

20:46

They can scrape your LinkedIn profile,

20:48

they can have a AI process that basically goes to LinkedIn,

20:51

scrapes your profile, ingest all of your data in real time,

20:54

or maybe in a couple of minutes, whatever it is,

20:56

asks GPT40 or whatever to consume all of it,

21:00

including all of your latest posts,

21:02

any articles that you've been referenced in,

21:05

infer or use all of that data,

21:08

go into PostScript to understand what your product does,

21:11

who you sell to, and kind of basically build

21:14

almost like a dossier for you in real time.

21:19

And I think that what makes AI a little crazy now

21:24

is going from being basically a word calculator,

21:27

which is basically taking all these things

21:29

and kind of spitting out a new profile of sorts of mic.

21:34

It's using that data to make inferences

21:37

that we're trying to get answers to.

21:39

So for example, I don't know what products

21:41

you were kind of filling out a form about,

21:44

but these are the specific features within our own product

21:48

that Mike might be interested in.

21:51

And we're inferring that based on the industry

21:53

of the company that he's in, his current role,

21:56

any latest blog posts that they might have written,

21:59

so on and so forth.

22:01

And what I have seen is actually probably not ready

22:06

to be implemented in replacing humans,

22:11

but I don't think 10 years off is the future

22:14

where it is ready.

22:16

I would argue it's probably maybe not to replace humans,

22:18

'cause again, the creativity point made earlier,

22:20

but I think the idea for AI to actually infer

22:24

and reason is getting better, which is actually, again,

22:29

interesting for business and terrifying for life,

22:33

but really getting pretty good.

22:35

And if you want an example of this,

22:36

I think it's a company called dossier.gpt or something.

22:39

They produce this concept and then Darmesh,

22:44

a hub spot has been building,

22:45

I think it's agent.ai,

22:47

where you can effectively just plug in a company name,

22:50

and it's quite powerful to see what comes out the other end.

22:53

So again, agree with your broader point,

22:56

but I've been spooked a little bit

22:59

and intrigued as a technology company founder or whatever

23:04

about just how good that reasoning engine is getting.

23:11

- Yeah, I really believe in the idea

23:16

that AI is going to be like everyone stepping

23:22

into their own Iron Man suit, right?

23:26

Like they get new powers,

23:27

they can do all the things they need to do way better, et cetera.

23:31

I 100% believe in that.

23:33

I think people are getting some value like that now,

23:36

and the value that is going to go up.

23:39

Where I'm still very skeptical is the replacement,

23:44

the AI actually took my full job away.

23:51

- Totally.

23:52

- I think that that's pretty far away,

23:55

but I totally agree with you that AI can tell us a lot

24:00

about people who are evaluating our products,

24:04

it can give us answers to questions,

24:05

it can, I put in into GPT just to see what it came up with,

24:10

and it was pretty dang good.

24:14

I was like, hey, if you had to compare PostScript

24:16

to its main competitors as a Shopify merchant,

24:19

like what do you think the strengths and weaknesses are?

24:22

The response was extremely credible,

24:24

like the type of thing that I would say,

24:27

if someone asked me that question on a sales call.

24:29

So I get all that value.

24:32

I think that what it is is it's going to allow people

24:36

to do things that spend their time doing things

24:39

that the AI can't do,

24:41

which I think is gonna be more of the creativity aspect.

24:46

- Yeah, I can't remember if I referenced this

24:48

on a previous episode, but we, I saw this tweet

24:53

about AI where someone was like saying,

24:59

I don't want like AI to like learn how to do poetry

25:03

so I can spend more time doing the dishes.

25:05

I want AI to figure out how to do the dishes

25:07

so I can spend my time doing my poetry.

25:10

And I think like that's where we're headed.

25:14

And I think that that's good.

25:16

I just think like the full automation idea

25:21

is like the wrong idea.

25:23

Like I don't like it when people start the AI conversation.

25:26

Like the executives eyes get all big

25:28

and they're like, how many roles can I replace with this?

25:31

You know, it's like,

25:32

- Right. Oh yeah.

25:33

- Yeah.

25:34

- And I feel like a lot of people are like,

25:35

that's like they're starting from like,

25:38

that's the value prop.

25:39

And I'm more saying that like,

25:41

I think a lot of the value prop should be about

25:46

how can you take all of the things that people have to do,

25:49

which a lot of it is mundane, you know,

25:52

box checking type work.

25:54

- Right.

25:54

- And get that off their plate so that they can reach

25:58

their full creative potential in whatever role

26:01

that they're in.

26:02

- Yeah.

26:03

- And theoretically that would yield overall

26:06

better results.

26:09

Even in the support use case, right?

26:11

If 60% of the tickets that come in for any commerce brand

26:15

are where is my order tickets

26:17

and AI can handle those,

26:18

then your CX team can focus on the more complicated stuff

26:23

or the surprise and delight moments for customers

26:25

that they can see are buying lots of things

26:27

and asking questions.

26:28

And so that's where it's like,

26:30

you automate to create more business value

26:33

with the folks that you have.

26:34

I think like that's the world we're like in and headed

26:38

towards, which is a great world.

26:39

The world where it's like,

26:42

oh, you know, we're gonna have,

26:46

like and no one's gonna have a job in two years.

26:50

I don't believe in that world today.

26:52

- Yeah, no, I'm with you.

26:53

I'm with you on that.

26:54

I actually think we completely agree on this in general.

26:57

I do wanna just for fun,

27:00

do you wanna do a little demo together?

27:03

- Sure.

27:04

- Tell me to pound sand if you think it's a bad idea,

27:07

but it could be fun just like do a moment in time

27:12

of where is the reasoning engine today

27:14

at the publishing date of this episode?

27:16

And then we can revisit this in a couple years.

27:18

- Okay, cool.

27:19

- Yeah, where we're at.

27:19

But I need your written, I need your verbal consent

27:22

that I can search your name.

27:24

I don't wanna know. - Oh, no, oh.

27:25

- If you don't want to, we don't have to.

27:27

- Go ahead.

27:29

- You sure?

27:31

- Oh, God.

27:32

I don't know, edit the sound if it comes up

27:34

or something crazy.

27:35

(laughs)

27:35

- No, but the question is, and it's still processing.

27:39

Hopefully by the time I click this, it's on live.

27:41

The question is like, how good is this inference thing?

27:45

How good is the reasoning of this platform?

27:47

- What's the prompt?

27:49

- Nothing.

27:50

Your name and your company.

27:52

- Ooh. - That's all that was put in.

27:54

So this is again, this could be off of a lead gen form.

27:57

- Okay. - Right, or off of anything.

27:59

It comes out dossier, GPT, or I guess it comes out,

28:03

ammonia, al-m-m-m-m-m-m-m-m-m-m-m-m-m-m-m-i, I think.

28:08

But this came out live this week,

28:10

so this is actually brand new, right off of the GPT-40.

28:13

I think it's built on 40.

28:14

It has been more than three minutes,

28:17

so this is probably edited at some point here, Todd.

28:22

- No, this is proving my point.

28:25

- It is true, it is too old.

28:26

- If it pops up and it says PostScript is the best,

28:30

Mike Mannheimer is the smartest.

28:31

Everyone should use PostScript if you have a shop at Firestore,

28:34

then that I'm a believer I'll change my tune instantly.

28:37

- Then it's worth it.

28:38

- Yeah, then I'm like, everyone, just use AI for everything.

28:41

(laughing)

28:43

- It's a good idea.

28:44

- Yeah, it's great.

28:45

Take everyone.

28:46

- Oh, here you go, look at this.

28:47

- Perfect, okay.

28:48

So look, it pulled on the left side

28:50

of the screen your LinkedIn profile,

28:52

so ignore that for a second, right?

28:54

So here's what it's telling us about you.

28:56

Mike, current position, location,

28:59

all of this information, right?

29:00

- It's all right.

29:00

- Some mics, season, market,

29:02

you're a seasoned marketing executive,

29:03

the strong focus on B2B and sales is gonna work a strategy.

29:05

- Isn't it funny that we call each other seasoned?

29:08

Like that's like--

29:09

- I know.

29:10

- It's like a food.

29:11

(laughing)

29:12

- Nope, maybe.

29:13

- Maybe. - Maybe have a reason.

29:15

Maybe to your point, I've done this about five, six times.

29:17

Everyone has been seasoned.

29:19

- Everyone has been heavily seasoned.

29:21

- Don't forget to-- - Okay, go ahead, sorry.

29:23

- That was good.

29:24

- All right, so the question is like,

29:25

now these interests are inferred.

29:27

This is not like just a word calculator,

29:30

so it's inferring, and so you tell me

29:31

if this is true or not.

29:33

That you're interested in professionally, of course.

29:36

SMS marketing, strong use of SMS and B2B to see.

29:41

Promoter score, growth marketing, continuous learning.

29:46

- Are those--

29:49

- I mean, that's not wrong, but that's not wrong.

29:53

No, yeah, that's accurate.

29:56

I would say, hold on, I'm looking at this right now.

29:59

- I'm all sure that's with you, of course.

30:00

- Yeah, I think, yes, some of these are more on than others.

30:05

Like the SMS marketing write up that it did

30:09

is like really, that's pretty strong.

30:12

- Pretty strong, yeah, that one,

30:14

that one I felt pretty good about

30:15

and knowing you, of course,

30:16

and the work you're doing at PostScript.

30:18

Net Promoter score, I mean, unless there's something--

30:21

- I wrote a blog post about NPS a long time ago,

30:26

so it must've found that.

30:28

And then like, summarized it and used it

30:30

as like a personal interest.

30:34

- Yeah, so check us out.

30:35

So this is where the SDR comes in or whatever, right?

30:37

- Yeah. - Topics of interest for discussion

30:40

when meeting with Mike consider these following topics,

30:42

innovations in SMS marketing, I think that sounds good.

30:44

- Yeah. - 'Cause more loyalty and NPS.

30:46

Maybe I would've liked to maybe have a reference there

30:48

that you just wrote a blog post,

30:50

so I can like reference a blog post,

30:52

growth, marketing, continuous learning,

30:54

and development, e-commerce trends.

30:56

It's pretty good.

30:58

- Yeah, the thing that's, if I had to critique this,

31:01

this is pretty good.

31:02

If I had to critique this,

31:04

oh, there's the circus on the bottom.

31:06

- Yeah, it's working, I see that.

31:07

- Yeah, if I had to critique it, what I would say is,

31:10

it's basically re, it's taking what I'm talking about

31:15

in the world and trying to build a profile of me.

31:20

But it's not, like just because I'm talking about something

31:26

doesn't mean that like, it's particularly my interest, right?

31:30

Like I talk about SMS marketing for my customers all the time.

31:35

- So I'm obviously interested in SMS marketing

31:38

like for my customer and I talk about it a lot,

31:41

and that's what's coming back to this.

31:44

What I would say is, if you're trying to like get my attention,

31:48

it's not always true that parroting back stuff

31:52

that you see online is actually aligned with like,

31:55

well, the problems I'm trying to solve.

31:58

You know what I mean? - Yeah, yep.

31:59

- Like if someone was like,

32:02

"Hey, AK, I see you're really interested in owned marketing."

32:05

You'd be like, "Ah."

32:07

- Oh yeah, okay.

32:09

- Yeah.

32:10

- So there's probably like a,

32:13

again, this gets back to the human component, right?

32:17

It's like a human would say,

32:19

what does Mike, the CMO at PostScript care about?

32:23

Like I probably care about like winning business

32:29

from my competitors.

32:30

You might have better luck sending me an email

32:33

that's like, "Hey, Mike, here's some things

32:36

that competitor A or doing that I thought

32:38

would be interesting to you."

32:40

Versus being like, "Hey, Mike,

32:42

I think you're really interested in SMS marketing."

32:44

- Yeah, totally.

32:45

- And so like that's like the extra level.

32:47

And when I think about like noise and response

32:51

and like all of that,

32:53

like the bar's just going up and up and up.

32:55

So it's like, I almost think that there's probably a chart

32:57

that's like the bar for getting attention

33:02

is going up at some, at the same ratio

33:08

that AI's getting better, right?

33:11

It's like they're happening at the same clip.

33:15

And that's, that Delta is creating the continued need

33:20

for human reasoning, business acumen, creativity

33:25

to continue to exist.

33:26

And maybe the slope of the graphs change

33:28

and AI like closes that gap.

33:31

But I think that that's like the image I have in my head.

33:34

Like there's two lines going up at the same time.

33:37

AI is getting way smarter

33:38

and the bar to capture someone's attention is going way up.

33:41

And they're both going 100 miles an hour

33:43

in the same direction and maybe one day they'll intersect.

33:45

- Yeah, it's funny.

33:46

I was thinking about too, like it's only gonna be so good

33:49

as the information we feed it.

33:51

And it's almost like, I think we talked about

33:54

this in a previous episode, LinkedIn,

33:55

and sometimes can become like the Instagram for work

33:59

where we're painting this picture

34:00

that everything's great and we're killing it

34:02

when things are really hard.

34:04

So yeah, you know, I don't know if there's,

34:09

if there's a perfect way for even a human to know this,

34:12

but you know, you might be,

34:14

like the discussions happening internally

34:16

and the problems you're looking to solve

34:17

aren't always like publicly disclosed.

34:19

And so it's not being fed into any type of algorithm

34:23

that could infer anything really.

34:25

- Yeah, what you really need is like,

34:28

it's like a bunch of inferences that lead to like

34:31

background knowledge, a bunch of inferences

34:33

that lead to something, you know?

34:36

And so yeah, I think that what you showed

34:41

was definitely way better than I thought.

34:44

I also think that if you're gonna be in like

34:47

the top 10% of outreach to Mike Mannheimer,

34:50

you probably need to like think about it for a sec.

34:53

well done. And yeah.