livestream header

So What? Will AI take my job?

So What? Marketing Analytics and Insights Live

airs every Thursday at 1 pm EST.

You can watch on YouTube Live. Be sure to subscribe and follow so you never miss an episode!


In this week’s episode of So What? we break down the tasks in different digital marketing jobs. We walk through the tasks that are ripe for automation and AI, and which tasks will need human intervention. Catch the replay here:

So What? Will AI take my job?


In this episode you’ll learn: 

  • the 2×2 matrix of AI versus human task
  • how to unpac your role
  • what to do to protect your job

Upcoming Episodes:

  • TBD

Have a question or topic you’d like to see us cover? Reach out here:

AI-Generated Transcript:

Katie Robbert 0:24
Well, hey everyone, happy Thursday, happy Cinco Demayo. If you celebrate that, welcome to so worth the marketing analytics and insights live show, I’m joined by Chris and John. On today’s show, we are talking about will AI take my job. Now this is a topic that gets covered a lot. And we also covered a lot, but there always seems to be some sort of a new spin, especially as we learn more about the capabilities of AI as we’re seeing job functions adapt to the technology coming and one of the things that we did last week. And you can see that in the newsletter, if you want to subscribe to our newsletter Trust Insights or AI slash newsletter, is we actually created a two by two matrix that helped explain jobs that were likely or tasks or other that were likely to be taken over by AI, and tasks that were likely to stay with humans. And what we want to do on today’s episode is really walk through that two by two matrix of is it more AI? Or is it more human? We want you we want to help you think about how to unpack your own job role. So how to break down your job function into those individual tasks, and then give you some tips and advice on what you can do to protect your job. So John’s gonna play Vantage today. But before we get into that, Chris, I want to ask you, because this is a this is a question, I think you specifically get asked a lot. You know, where do you stand on will AI take my job?

Christopher Penn 2:00
Well, as with everything, it depends, right? depends on the level of complexity of your job. And as we pointed out in the in the two by two matrix, the amount of repetition in a given task determines how easily AI can take that task now, where I think we don’t think enough about this is understanding that AI will not take your job in total, right. But what it will do, particularly at companies where there’s a lot of very similar roles is that if you can take, say one task away from a role that maybe takes an hour, and you pull that one task out of eight people, essentially, you freed up eight hours. Now at more progressive companies, those companies will say, Okay, now we need to help these employees do something more productive, more beneficial with that time, at more, say, cost focused companies, they’ll say, Great, here’s eight hours that we don’t need, let’s fire one person, because those eight hours we’re paying for those eight hours, we don’t need them, and shrink the number of seats that we have, you know, butts in the seats. Because we’ve taken away that task. And so will AI take your job is going to be dependent on the tasks that are getting automated away the extra value that you can find to provide for your company, and your company’s perspective on whether it it feels like has a growth focus where you want to take time saved, and uplevel your people, or cost focus where say, take time saved and reduce headcount to cut costs. That’s, that’s gonna be the determinate.

Katie Robbert 3:52
So John, you were thinking we were talking about this, and you were thinking about it in terms of applying the will AI take my job matrix to the traction model? So can you walk through a little bit of what traction is?

John Wall 4:07
Yeah, sure, sure. So this is a, it’s actually now our model. I mean, this was based on the work of Justin mares and Gabriel Weinberg, their book traction was the first iteration of this. But this idea that you can take the majority of marketing functions and break it down into these 21 different categories. And so their big argument is that, you know, as you start your business out, you pick three of these, you test them, and then it’s just continue to repeat the cycle. And the idea is that as an organization grows, ultimately you will be doing all 21 You know, because you’ll always want to be testing where to go. But it did seem like a neat idea to take. Now let’s take everything on this board here and throw it over on the grid so we can get a picture of which of the most likely positions to be eliminated and which ones are considered rock solid. And it’s funny, I was even thinking this morning that we can actually change the focus of this to it doesn’t have to be as a coming to take Your job it can be, which job should I be getting better at to AI prove myself, because that’s the same thing here, we can give you this list of okay, here’s the ones where, if you spend your time on this, you’ll still be doing this two years from now, as opposed for ones that, you know, maybe you shouldn’t be signing up for that $2,000 course on, you know, ad copy testing for ads, that might not be the best move you can make right now. So yeah, we can run down and start throwing them in, I actually put together the matrix that we’ve got. And I’ve got my big board here of the actual topics. So we can start cherry picking these and throw them in, if you want me to pick one off the board. Yeah, so

Katie Robbert 5:39
can you go back to the two by two matrix. So I just want to walk through a little bit about where that came about. So when we were talking about it, a couple of weeks ago, I was trying to find a different way to visualize, will AI take my job. And so what I did in the newsletter was I took a sample of some of the tasks that I personally do, and I put them to this matrix. So a standard two by two matrix just kind of helps you organize things from high to low. And so in this example, we’ve taken tasks that are not repetitive to highly competitive, and tasks that are not creative, to very creative. And so we’re going to map those, as will AI take my job. And so tasks that are not not repetitive, and highly creative are the tasks that AI is going to struggle to do the most. Whereas tasks that are very repetitive, and not a lot of creativity goes into it or the tasks that AI will do the best with. So anything that falls into the top left is where AI is going to step in. And anything that falls into the bottom right are going to be the safest task, it’s and so for example, you know, I have said, you know, interviewing people for an account manager position, that’s not necessarily something that AI can do, because every conversation is going to be different. There are elements of that task that I can, that I can use AI for such as screening and questionnaires and running through resumes, with keywords, looking for those kinds of things. But the actual conversations are not something AI will be able to take over. So that’s how we were thinking about plotting these, these tabs to figure out what’s safe and what you’d be concerned about. Chris, anything to add to that,

Christopher Penn 7:32
just the caution that, again, it’s not the job that will go it was the bundles of tasks. And so as we go through this, and we think about, say, search engine optimization, there’s four or five subcategories of that, like technical SEO, link building, etc. And then there’s tasks within each of those disciplines. So just keep in mind that wherever we net out on this chart, it’s not an ironclad, this is, you know, for sure what’s going to happen, this is going to be more there, there’s potential for more automation in certain tasks, and these roles have more of those tasks.

Katie Robbert 8:05
Alright, so John, where would you like to start?

John Wall 8:07
Yeah, well, you know, we can throw SEO up on there, that’d be a first one to jump through. So Okay, grab that.

Katie Robbert 8:14
So. So as we’re talking about SEO, so Chris, you had mentioned that SEO has a lot of different components? Can you talk through the four major pillars of SEO.

Christopher Penn 8:24
So the four pillars are technical SEO, which is infrastructure level work, for example, making sure that your server is fast, making sure that your network connections faster your DNS setup properly, all those very, very technical infrastructure level tasks. Then you have on site SEO, which I would group into a word called technical onset. So use the use of schema, for example, making sure that pages don’t fall for your redirects are functioning properly. So that’s, that’s all the optimizations you can do on a website. Your third bucket is content, right? So what content Have you written, you know, your blog posts, your videos, all that stuff that is going to try to appeal to an audience and be found by search. And the fourth is off site, where you’re looking at things like link building and brand building, which are almost synonymous these days getting placed in publications were featured on someone’s YouTube channel, whatever you can do to get links and reputational activities pointed towards your site. So those are the four big buckets and as you heard, there’s a lot of tasks within each one of them.

Katie Robbert 9:32
So if we start with technical SEO, for example, so making sure your servers are quick, making sure that your site is optimized and working, how much of that I mean, just the word technical by itself, to me implies that there is some kind of automation that is probably available, some kind of AI that can be programmed to look for issues and fix them. So, you know, within technical SEO, you know, what is your opinion of repetitive and creative.

Christopher Penn 10:03
If you do things right, it should not be very repetitive at all should be very, it should be, hopefully one and done, or, you know, once than an annual checkup to make sure that things have not drifted out of compliance, for example, your DNS entries, how are your CDN works, etc. Maybe Maybe you go to a quarterly webinar from your CDN to see the new features being introduced to Cloudflare or Akamai or whatever. And this is where we get into into challenges, because in some aspects, some parts of technical SEO are very non creative, for example, there are some very specific ways to set up your Apache or NGINX web server, right, and there are best practices and you really shouldn’t deviate from those unless you have a really good reason and you know, what you’re doing. But then there are other aspects which are highly creative, which is, you know, the strategic part of making decisions, where what kind of infrastructure are you going to have, right? How are you going to do your routing? How are you, you know, how are you doing your cat, you know, versus caching and proxy and stuff, those are decisions that you can’t really automate, because they’re highly constrained by budget, by your own skills, by knowing what’s possible, and by what vendors and platforms you work with. So it’s not repetitive, I would, it’s not creative, per se, or the strategy part is creative. But the implementation is not creative. But it’s not something that lends itself well to automation at all.

Katie Robbert 11:41
So if we start to pull that apart a little bit, so I feel like decision making, you know, strategy, setting, planning, all of that is highly creative, and probably not repetitive, because it’s something you might revisit. But as you have more information, the conversation changes. So I would put, you know, technical SEO planning and strategy in the bottom right hand corner of creative and not repetitive, then I would put the implementation in, not creative and not repetitive, because you already have the decisions, you’re already executing it. So I would put that in the bottom left, what I would put in the very repetitive and not creative is the annual check. So like you already have what the settings should be. So you could theoretically just automate that to say, make sure it’s hitting this threshold, or make sure that this thing is happening. And so while it’s not repetitive in something that you do every day, it’s repetitive and how you audit to look for errors.

Christopher Penn 12:50
The one other factor I think is worth considering is that machine learning and AI really focused on large volumes of data, when we’re talking about configuration of something that we’re really talking about very basic script driven automation, right there. You don’t need to train a machine learning model, you know, ideal memory size for an Apache web server, just implement a very, you know, something very basic scripting. So in that aspect. For technical SEO, there’s almost nothing that you would using machine learning at best would be overkill, and at worst to be a complete waste of your time when it should be automated for sure. But it should, but AI is the wrong application just like you need to stir your coffee and you’re like, let’s use the biggest power blender we have no no. Spoon.

Katie Robbert 13:39
Well, and I guess maybe the question is, Will automation take my tasks? You know, and so we can sort of think about it in that respect. So John, you know, I feel like we have three distinct buckets of things that we can plot on the matrix. We have technical SEO planning. We have technical SEO, implementation and then we have technical SEO audits and so the sorry I’ll vana catch up. I’ll be it’s, it’s one of those things. If you’ve never done it, try, like you can be a very quick and efficient typist, but when you’re trying to do it in a live setting with other people watching you, suddenly all the skills just go out the door. At least that’s what happens to me. Because it’s like the pressure is on for people watching in type.

John Wall 14:33
What do you make? Is that about where we’re mapping them?

Katie Robbert 14:36
Yeah. Can you change the color so it’s a little bit easier to see maybe just make them all black?

Christopher Penn 14:43
Makes me do this, we should use the Google whiteboard thing.

Katie Robbert 14:47
Well, I was trying to create a so what I learned is that you can make quadrant charts in Excel. And so I was trying to prepare one for today but because I’m not the technical person on the team, there was just a step that I couldn’t master. So I couldn’t make it work. Okay, so this looks great. So this, right now, if you are someone in your organization who is responsible for technical SEO, then you probably don’t need to worry too much about AI, machine learning or automation, taking your job away from you, because there’s not a lot that the machine or the computer would necessarily do better than you can. So if you’re part of the technical SEO, planning and strategy, perfect, you’re totally safe. If you’re part of the technical SEO implementation, then you’re probably still good there. Because as Chris mentioned, that’s kind of a one and done. And you want to make sure that it’s being done to the specifications, and it’s probably more efficient, to just do it yourself versus programming a machine to do it. And then with the audits, those happen less frequently, but you’re also just sort of following what was implemented to make sure if this has changed you this sort of thing. So there’s not really a lot of room for any sort of AI or machine learning or automation in that particular role.

Christopher Penn 16:11
For a brand, the exception would be if you are a relatively large SEO agency, and you have a canned audit process for say, you know, that you want to run on 200 clients, then there’s some automation opportunities there for sure.

Katie Robbert 16:26
What do you think, John?

John Wall 16:27
Yeah, and it’s interesting to me, the idea that, you know, not only is SEO a black box, but it’s also a moving target. You know, you don’t, nobody knows for certain what the algorithms are. So it’s constantly having to be tested and prodded. And then the fact that it does change every time, you know, the search engines can continue to adjust. So that just steal so much of the repetitive nature out of this, like Chris said, there’s definitely if you’re an SEO firm, then you may have some stuff that you can stack up and automate. But for an individual organization, there’s just, there’s just way too much testing and responding to come up with anything that you can easily automate or throw to AI. It says, It seems like this is a solid career path. If you’re doing something in SEO, you shouldn’t be losing too much sleep right now.

Katie Robbert 17:14
So let’s skip over to within SEO. Chris, you had mentioned content creation is the third of four pillars. I feel like this is a really good role to unpack down to tasks, because this is also a hot topic of conversation, because there’s a lot of software that will write content for you. And actually, I’ll be participating in the freelance chat next Thursday on Twitter talking about, you know, AI and content creation. And it’s a question that comes up a lot of should I just let the machines create the content for me? And so obviously, it’s a complicated answer, but we can at least start to unpack it. And John, I know that you found an article that basically explain Google is going to reject any content that they think is AI created. So let’s sort of get into the steps, the tasks that go into content creation,

Christopher Penn 18:17
no content creation always comes in. There’s the ideation of it, there is the creation of it, and then there’s refinement and then eventually publication. And then you have distribution and promotion. Where machines are being used today, is on the generation portion. So be having machines be able to generate with generative adversarial networks, things like images, or music, or with a larger language models, the ability for them to write words that are coherent, the challenge with a lot of those approaches is it depends on the level of quality that you’re after. The machines today can write extremely large volumes of mediocre content. Right. So if you, for example, a press release is a very structured language, there’s a lot of commonalities in every press release. And machines are really good at generating them, you know, start to finish. You have to have someone edit them to, to make sure that it didn’t make up something like invent a new executive company, but for the most part, they’re really good at that. Machines can write really, really high volume, mediocre blog posts that don’t really say anything. But you know, if they’re coherent, that makes sense. They are they’re kind of like repeating just general platitudes in whatever the topic is. But if you’re, if your requirements are that you have exceptional content, machines is going to struggle with that. And so it’s an interesting quandary and one night think that not a lot of people are thinking about from an AI perspective. When you train a machine learning model, on language, you have to use a lot of language, most languages mediocre, right? When you ingest all of Wikipedia, you know, just all of Reddit, you’re not ingesting the classics here, right? I mean, they’re in there. But they represent such a small portion of the corpus that you’re effectively training machines on mediocre middle of the road, Middle Level Education language. And that’s what’s going to generate. And so here’s the, the scale you have to balance on. If you if part of your SEO, which is the off site is contingent on getting people to link to you, and people linked to you, and naturally if your content is exceptional, but all you have behind the scenes are machines generating mediocre, you have a substantial mismatch in two parts of your SEO strategy, you can either go with a lot of mediocre content and hope that you pick up some traffic on the longtail because you’ve got 100 million web pages that all say, you know, tribe, or you scale back your automation on the generation side, you make it fully human, to create things that have never been done before. And that generates the uniqueness, the pitch worthiness of your content that then the pitching team can go out and get a ton of links for and get get you on Good Morning America. And also because you’ve got something so exceptional. So those two things are actually at odds. And it’s I think it’s kind of an important point for people who are on the generation side to remember is your machine’s not going to generate exceptional content. And as a result, it’s gonna be harder to pitch it.

Katie Robbert 21:43
So it really comes down to your standards. And if you just need large volumes of content to sort of get start to get that awareness and to have that presence, and then simultaneously creating that really high quality, unique content. So

Christopher Penn 22:00
your stamp that that standards or mark is a super important, Katie is super important. Because let’s say the machines can create this level of mediocrity, right? If all you’ve got is Bob, the surly intern, who’s creating content down here, right at I hate my job, and I hate you level, you know, I’m gonna face roll on my keyboard, because coming to work every day drunk, a machine is the better choice get rid of Bob, right? Bob the attorney go and go with go and get your yo by going with machines, you’ll often get to mediocre. So for organizations, we have to generate a lot of content, and you have employees who on say, the left hand side of the bell curve in writing quality, it might make sense to take away all those tasks from all those employees and say, Okay, we’re gonna have machines generate the mediocre stuff, to least get us to mediocre. Now, if you’re a company like Trust Insights, hopefully, our standards Yeah, you know, what, AI is not the right choice.

Katie Robbert 23:01
Well, you know, and, John, I would like your take on this, but I feel like we’re sort of in that, you know, unique position is probably the wrong way to phrase it. But we’re sort of, we have the luxury of being picky about our content, because the business is really just three of us, we don’t work at an enterprise sized company where the volume of content, you know, is directly correlated to other things, and goals and KPIs and that kind of thing, we can take our time a little bit more. So I would say our standards for the content that we push out for ourselves is much higher, because we have the luxury of having complete control over all of the different aspects of the content and the business. So with all of that, John, what are your thoughts on content creation? Where does it fall on the matrix?

John Wall 23:49
Yeah, well, I’d say, you know, if content is part of your strategic direction as an organization, then it basically is not something he farmed AI. But this, it’s interesting that, you know, aligning content creation or with SEO, I could say yes, that’s an SEO function. You know, if you have 1200 McDonald’s, and you want a web page for each McDonald’s, well, then yeah, you could automate and have, you know, 1200 pages spun up automatically, you know, via machine for that, and that’ll be fine. But basically, any org where content creation is part of your strategy, yeah, AI generated content is not going to be good enough, you need to be proving your human worth. And that’s totally, you know, what, you’re where you’re at, I mean, I would definitely put this over. I mean, there is some repetitive stuff, but it’s definitely, you know, high creative. Iterative in generation, you know, it’s, it takes human touch all the way through to get it before you get it out the door.

Katie Robbert 24:46
Well, and I think that that, you know, you know, Chris, as you’ve sort of been describing, and John, you know, to your point about the multiple web pages for different locations. You know, I think that there’s we could break it down into those two categories as well. One is volume and one is, you know, uniqueness. And so, you know, I can see scenarios where you know, a larger company or just anyone who needs to create large volumes of content, they can turn to AI to create the content, but then the time is then taken to do the editing and massaging and sort of adding your own personal spin on it. Because I think a lot of what happens, at least this happens to me personally, is you start to get stuck in what you’re writing. But if you’re reacting to something, it can help, you know, move along the process of writing. So I can see a scenario where having AI generated content, at least a start can be helpful to getting to that higher value content.

John Wall 25:52
Grab something else off the bingo list here.

Katie Robbert 25:54
Yeah, Chris, what? I’ve picked the last couple, what do you want to tackle?

Christopher Penn 25:59
So let’s tackle public relations, traditional media.

Katie Robbert 26:03
All right. So do you want to start so I can, I can take a stab at sort of breaking down. So I’ll be honest, I probably have the three of us know the least about public relations. It’s not a discipline that I ever had a lot of access or insight into. And so my understanding is that public relations involves a lot of content generation. I think it can involve a lot of pitching the story, the headline, the brand, the person, two different outlets, whether they be online or offline. You know, it’s, it’s basically that brand and public awareness of something. But I know that I’ve probably missed a lot of steps in there. So Chris, why don’t you fill in the blanks for me,

Christopher Penn 26:55
no, really, public relations, at least for traditional media. And even you know, with social and influencer marketing is fundamentally a sales job. So your job is to sell your clients story to some form of outlet of distribution. And everybody in their cousins pitching those same outlets, you know, John, and I get horrendous pitches to marketing over coffee, a couple of times an hour of just, you know, check out the you want to interview are a person who did this thing that 150 Other companies have done better. But we’ve got a cooler name that has more vowels. We get a lot of that. So it is that sales job it is, is working with your client to some degree to figure out what it is that what idea whether it’s content with this interview, whether it’s you know, industry news, what idea you want to sell, and then you pick up that phone, and you start Island and it like the boiler room and Wolf of Wall Street, you got to just hit the phones and and see who’s going to pick up now. In that aspect, there is a lot of automation in the PR industry already. And it’s all uniformly terrible. The automation that is in the field is so bad that you’re better off not using it, it’s it’s dangerously bad, because what happens is you’ll go into a media database, you will identify, you know, you choose your list of targets, unless you’re like that one idiot we worked with at the last agency who just emailed the entire database of 55,000 publications.

Katie Robbert 28:34
Everyone’s entitled to a bad day. Yeah, wasn’t me.

Christopher Penn 28:42
The automation, just at best attempts of keyword matching, at worst, you know, has you know, spraying and praying. Again, one of the things that we know for sure, is horrendously done is when we John and I get pitches for coffee, like that’s not what our show is actually about our show is about. And while they’re entertaining, you know, there is a clear sign that the software, the automation hosed it. And so in that role of public relations, the automation opportunities, much less the AI opportunities are sporadic and very ad hoc. So for example, one of the things that we used to do a lot and we worked at the agency was use topic modeling, to ingest all the news stories about a specific topic and see what audio has been written. Right to Try and get a sense of, okay, here’s everything has been covered. So that, you know, when when the PR person goes into pitch, they know that they’ve got to find a different angle. If the client is saying we’ve got this flexible, scalable, integrated, fully integrated, turnkeys SaaS solution. Everybody has heard that. That’s not news. So those are sort of the kind of ad hoc opportunities where automation and AI can Make a difference. But on the management of relationships, and the ideation of you know, is this a noose words where the story and even identifying who to pitch that stuff that right now is actually better left to humans.

Katie Robbert 30:15
So it sounds like weeding out what’s already been done is good for AI. But the actual going back to that content creation, that still takes a lot of human intervention, and so that part of the role is still protected. As long as you have those standards that we don’t want to just generate the same turn key, whatever news story that everybody else is generating, to be quite honest, I used to work in academia, and even, you know, my products, part that were part of clinical trials were not protected from terms like turnkey, like it’s find a different word, just just find a different one, open and find something else.

Christopher Penn 31:04
And the pitch in part two is really a part that should not be automated right now. crafting the pitch, dust I eat, like we just did a round of pitches, encouraging people to partner with us on some of our reporting. And the three of us spent a lot of time thinking about how do we want to approach this? Who do we want to approach? What’s the message going to be? I would not have left any part of that process is to a machine.

Katie Robbert 31:32
Right? Well, and you know, it’s, it’s funny, because when I was thinking about this matrix last week, and it was breaking down, my like a small sample of my tasks, I had put that, sure, I could let AI generate the sales pitch based on some key words, and, you know, some direction. But to your point, Chris, like, that’s not what our standards are, if we were generating 10 sales pitches a day, that might be different might be like, Okay, let’s just get something out the door, see if we can throw a bunch of stuff against the wall and see what sticks. But that’s not the kind of standards that we’re trying to hold for the company. So John, it looks like you’re putting traditional media PR traditional media in the highly creative, but also very repetitive. Yeah, final answer,

John Wall 32:24
not super repetitive, but some repetitive.

Christopher Penn 32:26
No, there’s a lot that’s repetitive processes

John Wall 32:29
more endlessly, not repetitive.

Katie Robbert 32:31
But there’s a lot of creativity to it. So yes, if you’re good at your job, there’s, and I think that that’s an interesting thing to just sort of note for a second is, the process itself might be very repetitive. So think about a writing framework, that’s a repetitive process, you can structure your content the same way every time. What’s unique about it, what’s creative about it, is actually the point that you’re trying to make and your personal point of view into how you’re making it. And the same is true of traditional media, the PR, of the process of pitching might be identical every single time you have the thing, you pitch the thing you close the thing, like whatever that process is, clearly I’m not, again, not a PR expert. But the creativity that goes into getting that pitch scene, getting that content created in such a way that it’s unique, is that’s where the creative piece falls in. Exactly.

John Wall 33:30
So this gives us an easy hit to so now we can I can just take direct sales, which, as we just talked about, traditional media is just a sales job. So that just goes right side by side with Alan,

Katie Robbert 33:43
you’re Are you just trying to protect your job, John?

John Wall 33:46
No, no, I, I know no machine can deal with the level of insanity. I see. There’s nothing repetitive and there’s nothing that doesn’t require creativity to get out of.

Katie Robbert 34:01
So I think similar to PR, I think there are opportunities for automation in direct sales, such as those initial cold pitches, so getting people to respond to you to you know, become aware of you. So you could create a sales campaign in your CRM or your marketing automation system, and send that out in mass. And so that part you can automate and sort of you can create those nurture campaigns, you know, if they take this action, do this, you can build all of that to automate. But John, to your point, once you have someone who’s interested, once they’re engaged, you probably want to take the automation and the AI out of it, because that’s when you need that human intervention.

John Wall 34:50
Yeah, it’s funny as going through this process has made me realize that all of this is dependent on ability to scale. You know, a small organization is doing a lot of work on the press. To Front and the customers are new. So there’s just so many of the things that could be automated are not in the mix yet. And it’s when an organization matures, and the buying cycle is solid, and it’s a large enough company, then suddenly, more and more opportunities for automation and AI, come to light, I think it kind of starts forcing everything on the chart over into the more repetitive quadrant. And so then you have more opportunity to automate.

Christopher Penn 35:26
The other aspect that’s not on here, but is implicit in some ways in which way people think about it, but it should not be is on analysis and analysis and analytics. So for example, in sales, the process of lead scoring is something that is exceptionally qualified for, for AI. Because, and not just automation, but true AI because you have so much data, and you can’t see the patterns in the data. But a machine could say yes, we’re going to surface these five leads as though these are the five leads, you should call today, because they’re of all the indicators. And behind the scenes. These are the ones showing the sort of the green light in SEO, taking a series of headlines and, and measuring them and scoring that would be one example. On the other end, any kind of attribution analysis, all that is 100% stuff that AI should be doing because, again, people are really bad at it. So you know, is your SEO working? Is your content, marketing, working and working on something based on last week’s show about exploratory data analysis in content marketing, that is all AI, that should all be AI? Because you don’t want to try and figure this stuff out by hand, it’s just not going to go? Well,

Katie Robbert 36:45
as someone who had to QA, logarithms by hands I can wholeheartedly attest to you should be leaving that to machines.

John Wall 36:56
Yeah, jobs humans don’t want to have to do.

Katie Robbert 37:01
And, you know, I think that the analysis is true of any of these roles. I think that that’s a really good point, Chris. And I think that that’s where we always start to break down. What can a I do? What can AI not do? And so AI can do all of the analysis, prepare the data and hand you the bundle and say, here’s what I found. It’s then incumbent upon you, the human to draw those insights to understand context, and nuance and timing, and the audience. And all of those different factors that Sure, you could spend a lot of time programming that into the AI. But it’s forever changing, especially as you know, conversations change and opinions change, and people change. All of those factors will change the kinds of insights that you’re going to draw from the data that you’re looking at. And that is uniquely human.

Christopher Penn 37:53
The thing that we usually tell people to do, particularly when they’re just trying to get their wrap their brains around AI is say, substitute the word AI for spreadsheets. Right? So what things in this job could use spreadsheets, right? We’re working at spreadsheets make a difference? How can we? How can we use spreadsheets to increase our revenue? When you think about that aspect? Because that’s really what AI is just math? It’s just a lot of math. lead scoring on the basics here, it makes total sense. Yes, that is basically just spreadsheets run by a machine. Right? Content Creation, not, not a lot of that part fits into a spreadsheet, some technical SEO planning does, but it, it’s one of my favorite ways to demystify AISC. Just substitute one word spreadsheets, and then suddenly, either question becomes obvious, or you sound like an idiot, like, how can I? How can you use spreadsheets to re envision my company? No.

Katie Robbert 38:47
You have to write a lot of things down in a spreadsheet, and then a spreadsheet is the wrong tool.

Christopher Penn 38:52
Exactly. But saying how can I use spreadsheets to improve my ROI? Yeah, that’s a place where there’s a lot of numbers, and an AI will eventually be part of that.

Katie Robbert 39:06
Yeah. How can I use spreadsheets to do my lead scoring? Absolutely. How can I use spreadsheets? To do my technical SEO planning? It’s a tougher one. Because that, you know, in I think that that’s where, you know, again, that human intervention of Sure, you can use the spreadsheet to collect the information. But there’s still so much human intervention that goes into actually putting the information into the spreadsheet.

Christopher Penn 39:36
Exactly. And your easiest implementations of machine learning are all with rectangular data and rectangular data is fancy for spreadsheet. So when you think about is AI going to take my job, look at any individual task and say, how much how many times do I use a spreadsheet in this particular task or something like that, or there’s something that’s so repetitive and if you’re just like, at our old PR firm There’s this one job at the bottom of the totem pole. The account coordinator copying and pasting results from Google into a spreadsheet. That was their job eight hours a day, how are they to claw their own eyeballs? I don’t know. But that you look at that go, wow, that is 100% a job, you don’t even need AI to do that, that is just straight write a 10 line program in PHP, and then call it a day because that’s, that’s not even a job. And there was a spreadsheet at the heart of that. So that was one where Yeah, machine learning should take that job because it’s a miserable job. Well, and

Katie Robbert 40:36
you just said something, Chris, that, you know, strikes me as the the people who can introduce AI and the people who can’t introduce AI. And so if I was the person who was responsible for copying and pasting the numbers into a spreadsheet all day long, and you know, came along and Segal and said, just write a 10 line, you know, PHP piece of code, I would look at you like you have six sides and be like, I don’t know how to do that. Does anyone here know how to do that? And the answer is probably no. And so the learning curve for being able to write and create and implement AI is still pretty steep. It’s not, you know, there’s not a lot of great like, here’s the out of the box solution. We’re getting there in terms of the technology. But that’s also a factor of, do you even have the skill sets to implement, execute and maintain the AI? So while it may be so lead scoring, for example, while it may be a task that AI should do, it doesn’t mean you have the means to do it.

Christopher Penn 41:44
We’re working on it. But yeah,

Katie Robbert 41:47
no, I know. Yeah, I mean, that’s the thing. I know we’re working on it. But like in that example, the solution is to write some PHP, I don’t know how to do that. And so I would be like, uh, I guess I’m gonna keep doing it the same way, because I don’t know how to implement a different solution.

Christopher Penn 42:04
And it’s interesting, because when you think about the matrix, then write the not create a very repetitive tasks are going to be typically tasks lower down on the org chart where theoretically on reasonably, the creative, non repetitive tasks like strategy and stuff are going to be higher up in the org chart and an organization. So there may be a lot of value. And that can be unlocked by artificial intelligence and machine learning, or even just plain old automation, at the lower levels in the organization. But the stakeholders to your point, the people who have the authority to authorize it may not even understand what’s happening in the lower levels of the organization. And therefore, they don’t know that there’s massive cost savings to be had, or massive productivity gains to be had. Because they don’t they have no visibility into the actual work being done. And as we talked about, in recent podcast, when you have organizational silos that also include very strict hierarchies, the junior person whose job it is to copy paste eight hours a day, can’t talk to somebody and say, Hey, I think this is a waste of my time.

Katie Robbert 43:06
Yeah, that’s a whole other issue.

Christopher Penn 43:08
That’s a new shot.

Katie Robbert 43:10
Yeah, that’s I you’re trying to get me up on that soapbox, Chris. And I’m just not taking the bait not today.

Christopher Penn 43:17
So well, we should probably start wrapping up anyway.

Katie Robbert 43:20
Yeah. So John, is there any other things you want to note before we close down the show on this particular matrix? What are your thoughts?

John Wall 43:31
The big one that this has opened up for me is that AI is not the Boogeyman. It’s automation is still your problem. You know, I mean, AI is there but it’s really you need to be more worried about stuff getting automated out of your job. It’s AI is, is on the horizon is very specific application. But don’t worry, there’s other ways you can lose your paycheck.

Katie Robbert 43:53
Plenty of other opportunities for that. So on that note, on that note, Chris, any final thoughts?

Christopher Penn 44:02
No, I think John, John said it all. We’ll see you next week, everybody. Thanks for watching today. Be sure to subscribe to our show wherever you’re watching it. For more resources. And to learn more, check out the Trust Insights podcast at trust AI podcast, and a weekly email newsletter at trust Got questions about what you saw in today’s episode. Join our free analytics for markers slack group at trust for marketers, see you next time.

Transcribed by

Need help with your marketing data and analytics?

You might also enjoy:

Get unique data, analysis, and perspectives on analytics, insights, machine learning, marketing, and AI in the weekly Trust Insights newsletter, INBOX INSIGHTS. Subscribe now for free; new issues every Wednesday!

Click here to subscribe now »

Want to learn more about data, analytics, and insights? Subscribe to In-Ear Insights, the Trust Insights podcast, with new 10-minute or less episodes every week.

One thought on “So What? Will AI take my job?

Leave a Reply

Your email address will not be published. Required fields are marked *

Pin It on Pinterest

Share This