Welcome to AI University! At Bospar, we understand that AI isn’t going anywhere. So instead of resisting, we are embracing AI and leveraging the power it holds to help us do our jobs better. Courtney Merolle and Kyle Ankney, our professors at AI University, join Eric Chemi to talk through how we thoughtfully leverage AI in our day-to-day operations and how we ensure every Bospartan is well-versed in the complexities AI can hold.
Click to read transcript
00:00:10.639 Today we are talking about AI of
00:00:13.040 course right AI is the only thing that
00:00:14.559 people seem to be talking about these
00:00:16.000 days and today I’m with Kyle and
00:00:17.840 Courtney Bospar employees who are in
00:00:19.840 charge of the Bospar AI initiatives AI
00:00:22.480 University AI programming everything AI
00:00:24.960 that’s happening here at Bospar so you
00:00:27.039 know Courtney I’ll start with you every
00:00:28.320 headline I read right now is that AI is
00:00:30.560 going to eliminate everybody’s jobs
00:00:32.479 including all the people that work in
00:00:34.399 places like this well what is your take
00:00:36.079 on that
00:00:37.920 I definitely don’t think it’s
00:00:39.120 eliminating our jobs but I am going to
00:00:42.239 sound like a broken record because all
00:00:43.840 the other CEOs are saying the same thing
00:00:45.920 it It’s going to make us more productive
00:00:47.840 it’s going to make us more efficient
00:00:49.600 there’s ways to create our outputs and
00:00:52.000 and think creatively and and do our
00:00:54.320 deliverables in a way that we’ve never
00:00:56.719 been able to do before and it doesn’t
00:00:59.840 eliminate our jobs we still very much
00:01:02.079 need a human touch within communication
00:01:04.080 we speak about emotion um and end
00:01:06.640 product and and what that looks like and
00:01:08.880 the synergy that exists within it you
00:01:11.040 can’t remove us but we can be more
00:01:13.840 productive
00:01:15.439 you can’t remove us yet right right kyle
00:01:17.520 what you’re you’re on the cutting edge
00:01:19.200 what do you see because I feel like a
00:01:20.320 lot of AI now it can do emotion it can
00:01:23.040 do reasoning it can do it can do so many
00:01:26.159 things that we would have thought needed
00:01:29.119 a human to do no I think that’s true and
00:01:31.680 I was we were Courtney and I were having
00:01:33.600 this discussion earlier that yes with
00:01:36.159 each new model and each new progression
00:01:38.320 there’s always things that are happening
00:01:39.680 we’re like “Wow that’s different that’s
00:01:40.960 new that’s exciting.” But at the end of
00:01:42.640 the day you know especially in places
00:01:44.720 like a PR agency I think people are
00:01:47.280 paying for expertise and a lot of people
00:01:50.159 can experiment with you know different
00:01:52.479 models and different prompts and things
00:01:54.479 and get really good outputs but at the
00:01:56.720 end of the day the application and
00:01:58.960 figuring out how to really package it up
00:02:01.600 in a way that not only makes sense to
00:02:04.479 the end user but media there there’s a
00:02:07.520 skill that isn’t quite at the AI level
00:02:10.479 and I won’t say that we’ll never get
00:02:12.080 there but we’re not there yet so so what
00:02:15.520 are some of the initiatives that you two
00:02:16.959 are working on internally is it about
00:02:18.879 educating employees is it about you know
00:02:21.440 showing them the right tools is it some
00:02:23.040 a AI ethics what exactly is the role of
00:02:25.920 of leading these Bosespar AI initiatives
00:02:29.920 feel free to jump in you two can
00:02:31.440 disagree with each other and argue with
00:02:32.879 each other don’t wait for me to ask you
00:02:34.160 guys can just jump right in and chat
00:02:35.280 with each other I would say right now
00:02:37.680 we’re creating what I like to refer to
00:02:39.920 as the AI playground I think a lot of
00:02:42.800 people are very fearful about AI to a
00:02:45.440 point where they’re scared to even use
00:02:47.440 it dabble in it we hear about you know
00:02:49.760 data concerns and and we as a PR firm
00:02:52.239 have NDAs and you know there’s a lot
00:02:54.720 from the journalist side about
00:02:56.080 plagiarism and you know not giving
00:02:58.400 commentary that’s humanmade um so
00:03:01.599 there’s a lot of concerns and scary
00:03:03.040 things attached to AI and it’s created a
00:03:05.680 world where people are not adopting it
00:03:07.519 as fast as you would expect considering
00:03:09.680 it does have incredible power in
00:03:11.920 speeding things up and we’re in a world
00:03:13.680 where you know we’re incredibly busy and
00:03:15.920 and would love to be able to power
00:03:17.519 through some of our workload smarter but
00:03:20.159 faster um so AI University is a lot
00:03:23.840 about creating guard rails and doing
00:03:27.120 some handholding but more so empowering
00:03:30.319 employees to find ways for them to
00:03:33.760 create a better workday for themselves
00:03:35.920 with the AI tool that they prefer oh
00:03:39.599 it’s interesting and I’ll let Kyle jump in
00:03:41.200 too like so you’re telling me actually
00:03:42.959 people aren’t using it enough i would
00:03:44.640 have thought you might have said the
00:03:45.519 other the opposite which is people are
00:03:47.200 using it too much you’re actually saying
00:03:48.560 no no they need to use it more
00:03:51.040 yeah we actually deployed a survey
00:03:53.200 internally throughout the agency just to
00:03:55.120 confirm you know what exactly is going
00:03:57.120 on with AI is it a world where people
00:03:59.760 are playing with a lot of different
00:04:01.280 tools are they using one and sticking to
00:04:03.680 it are they not using it at all um and
00:04:06.400 we do find a couple of trends did show
00:04:08.400 up one of which was people do find the
00:04:11.040 one tool that they like and they kind of
00:04:12.720 stick with it they don’t give themselves
00:04:14.560 to play with other options out there um
00:04:17.440 or use new LLM systems to see what the
00:04:19.759 results look like uh but yes
00:04:22.000 particularly from younger employees I
00:04:24.000 think there is still a level of
00:04:26.720 questions as to how to use it
00:04:28.560 responsibly to a point where they’re
00:04:30.800 slower to use it
00:04:33.440 and I want to just add to that I think
00:04:35.040 another component to this too is the
00:04:36.560 client side of things as Courtney was
00:04:38.080 saying um not only are we trying trying
00:04:40.479 to create guard rails for what works for
00:04:42.960 us internally but also letting clients
00:04:45.440 know we don’t want to say we’re not
00:04:46.639 using it and then get caught and say
00:04:48.000 like “Oh yes you did because this is
00:04:49.520 why.” We want to be very transparent and
00:04:51.840 and and intentional in saying “Yes we do
00:04:54.560 use it in these ways for this reason
00:04:56.800 this is how this is how your information
00:04:58.560 is safe.” and also empower our clients
00:05:01.840 to use it in ways that are smart and
00:05:06.720 practical for the media because the
00:05:08.240 media isn’t quite caught up with the AI
00:05:10.880 jargon or how to implement you know AI
00:05:13.840 written content we’re still figuring all
00:05:15.680 that out but to be able to utilize it in
00:05:18.160 a way that they’re getting the most PR
00:05:20.400 impact for their money and their
00:05:22.240 contract length they want to see how can
00:05:23.919 I utilize the tool as well to enhance
00:05:28.240 that experience so it really is like a
00:05:30.320 two-fold situation that we’re trying to
00:05:32.800 discover and unfold together so I see
00:05:36.160 the issue with clients
00:05:38.199 isn’t we’re concerned that you’re using
00:05:40.560 AI because we want you to be more
00:05:42.080 creative it’s we’re concerned you’re
00:05:43.759 using AI because we don’t want our
00:05:45.360 information now being put into a public
00:05:47.440 system is that really what the big issue
00:05:49.120 is I think so I think that’s more so
00:05:52.240 what I hear from clients than other than
00:05:54.960 it’s not necessarily a creative
00:05:56.639 challenge more so than a security and
00:05:59.080 safety concern so so what are the let’s
00:06:02.479 say basic let’s say a new employee
00:06:04.639 starts tomorrow what are you telling
00:06:06.400 them hey here’s the three things that
00:06:08.000 you definitely should be using AI for
00:06:09.840 and what are the three things that we
00:06:11.520 don’t want you to use it for
00:06:14.319 we don’t want you to use AI in the sense
00:06:17.280 of inputting any documentation from a
00:06:20.400 client that is not for external
00:06:23.319 consumption especially within our
00:06:25.199 industry there’s a lot of internalized
00:06:27.120 decks there’s a lot of company
00:06:29.520 information that’s not yet public
00:06:31.360 anything of that nature is an absolute
00:06:32.960 red flag we do not want that within an
00:06:34.720 AI model funding valuations things of
00:06:37.199 that nature we just can’t risk it uh so
00:06:39.440 that’s a pretty blanketed no um but
00:06:42.720 things that they should be using it for
00:06:45.759 I like to guide people to using AI as an
00:06:49.560 assistant to allow you to
00:06:52.199 create more unique deliverables or think
00:06:56.160 outside of the box um I find ChatGPT,
00:07:01.000 Claude they are really practical not
00:07:03.759 just in speeding things up while that’s
00:07:05.680 a great thing it almost creates a little
00:07:08.639 buddy especially within a remote world
00:07:10.800 someone who can almost think a little
00:07:12.720 bit different from you so that you can
00:07:15.280 be a little bit more on your toes about
00:07:17.280 how you’re creating things what the end
00:07:19.360 product looks like so it’s not so cookie
00:07:21.360 cutter um we have to be really
00:07:23.840 innovative and and different when we
00:07:25.759 approach media and I do I think ChatGPT is
00:07:28.080 a really great way to get you thinking
00:07:29.360 outside of your standard thought
00:07:30.720 processes
00:07:32.319 yeah I know and I will say one of my
00:07:33.919 favorite additional prompts once I’ve
00:07:35.840 already got a piece of work that I like
00:07:37.840 or a pitch that I prefer or an approach
00:07:40.800 with a journalist that I think will work
00:07:42.639 I always like to flip it around and then
00:07:44.240 and say “Okay well now I want you to
00:07:46.000 think like this journalist at this
00:07:48.000 outlet who covers this topic and poke
00:07:51.599 holes in this entire situation make it
00:07:55.280 all question everything and more often
00:07:58.639 than not I’ll get really strong results
00:08:00.479 of hey here’s where the pain points are
00:08:02.800 here’s what you’re not answering or
00:08:04.160 here’s where it’s being overlabored and
00:08:06.240 it just helps me streamline from a
00:08:08.319 perspective that I’m aware of but am not
00:08:10.879 in day-to-day
00:08:12.639 I like that I like that approach hey be
00:08:16.000 this reporter be this journalist and and
00:08:19.440 basically give me the rejection now
00:08:21.919 before I actually send the email pitch
00:08:24.160 later and officially get the rejection
00:08:26.080 you might as well let’s practice it now
00:08:27.599 so I can avoid those pitfalls and and
00:08:29.520 try to either answer them in advance or
00:08:31.840 maybe just don’t even pitch that person
00:08:33.279 in the first place like I I like that
00:08:35.039 way of of taking on these
00:08:38.919 roles how much how much are we spending
00:08:41.440 on AI internally is everyone just doing
00:08:43.760 like the $20 a month you know little
00:08:46.000 ChatGPT+ or are we getting some
00:08:48.320 corporate like what what is the the
00:08:49.920 situation right now that’s such an
00:08:51.760 interesting question Eric
00:08:53.920 um so currently there’s no there’s no
00:08:58.720 internal system in place where the
00:09:00.839 company is funding everyone to have an
00:09:05.200 AI assistant or buddy or however we want
00:09:07.519 we want to frame that however we are
00:09:10.320 currently actively looking for and
00:09:12.440 researching AI specific tools
00:09:15.360 particularly for teams that are geared
00:09:18.000 for the PR space um so harder to find
00:09:21.680 but they are starting to trickle out um
00:09:24.480 and there are many benefits to those and
00:09:27.279 the goal I think for the agency I don’t
00:09:29.279 want to speak for people above me but I
00:09:31.360 the goal is to hopefully find something
00:09:33.440 that works for everyone but allows us to
00:09:35.680 be collaborative with our AI use as a
00:09:38.160 team so that we’re not using our
00:09:39.680 individual ChatGPTs and Claudes in a way
00:09:43.440 that is siloed from everything else the
00:09:46.880 the company is doing oh so I see so like
00:09:49.519 if you if the three of us were each
00:09:51.519 doing something it should at least know
00:09:53.120 about it like in some in some corporate
00:09:55.200 system where it’s aggregating all of our
00:09:57.200 individual work and and that smartness
00:09:59.519 improves over time is that is that the
00:10:01.360 idea yes one thing we get asked about
00:10:04.320 quite frequently is the concept of a
00:10:06.080 prompt library you know ChatGPT or any
00:10:09.920 other system it it only gives you as
00:10:12.160 good as what you can question it and so
00:10:14.880 putting together just a really short
00:10:17.120 sentence that’s missing a lot of details
00:10:19.040 is is not going to get you the output
00:10:20.880 you’re looking for because there is some
00:10:23.440 level of repetition to what we’re doing
00:10:25.360 on the day-to-day it would behoove us to
00:10:28.399 eventually have a prompt library if you
00:10:31.279 will where you know you could just kind
00:10:33.360 of swap in and out some details but
00:10:35.519 ultimately it gives you a really good
00:10:37.920 in-depth thoughtful response when it
00:10:39.680 comes to survey questions or pitch
00:10:42.640 writing or you know award research
00:10:45.279 things of that nature things that are
00:10:46.560 really duplicative for us that’s a
00:10:48.560 that’s a good point but then you know
00:10:49.839 it’s funny a lot of times you can use
00:10:51.360 the AI to give you the prompts right now
00:10:53.920 it’s like hey I don’t have prompts make
00:10:56.240 me a prompt library and then boom it
00:10:58.079 does that for you so yeah I love kind of
00:11:01.560 flipping
00:11:03.160 AI and so there have been times where I
00:11:06.160 put it into a position where it’s asking
00:11:08.000 me questions rather than me asking it
00:11:10.160 questions oh really how do you So how do
00:11:12.000 you do that what are you saying to to
00:11:13.600 set that up so I will similar to the
00:11:16.880 role playing concept if I have a a piece
00:11:19.440 of um let’s say a press release or a
00:11:23.040 pitch I can put that into my ChatGPT and
00:11:26.320 say you know if I were doing a live
00:11:28.399 interview with such and such reporter
00:11:30.720 what are they going to be asking me and
00:11:32.560 then it’ll you know spit out a couple of
00:11:34.079 questions I would give it a response as
00:11:35.760 if I’m the the key spokes person and I
00:11:38.240 really just get a better sense of what’s
00:11:39.760 the forward thinking conversation at
00:11:41.920 play here that way we can get ahead of
00:11:43.839 those conversations and create talking
00:11:45.600 points create a briefing document that’s
00:11:47.680 more specific and cater to the you know
00:11:50.880 conversation that we’re going to have
00:11:52.240 things of that nature but it’s it’s a
00:11:54.320 slight twist on AI that people don’t
00:11:56.399 necessarily think of I like that I like
00:11:59.120 that yeah make them the one asking the
00:12:00.959 questions as opposed as opposed to us
00:12:03.279 you mentioned different tools you
00:12:04.839 mentioned how some people they’ll get
00:12:07.200 stuck on one tool and they won’t try the
00:12:08.959 others what tools are the two of you
00:12:12.000 experimenting with right now have do you
00:12:13.839 feel like you’ve tried all of them I
00:12:15.360 mean I guess no one has tried all
00:12:16.720 there’s so many different niche ones but
00:12:18.079 but what are the tools that you feel
00:12:19.279 like you’ve spent the most time with
00:12:21.120 what are the ones that you feel like you
00:12:22.800 still need to get to and then and then
00:12:24.240 I’m curious how would you rank them
00:12:26.560 right like hey this was really good at
00:12:28.000 something but this was really bad I’m
00:12:29.920 curious your adventures in this space oh
00:12:32.320 Kyle I haven’t heard from him in a while
00:12:34.560 okay i will say I have too um I like
00:12:37.519 ChatGPT because it’s kind of where I
00:12:39.360 started it’s where a lot of people
00:12:40.560 started but also I’m a huge Perplexity
00:12:43.200 fan um so what I find myself doing I
00:12:45.920 think Perplexity is much better at data
00:12:48.320 and data points and stronger research
00:12:51.120 information so typically I will if I’m
00:12:54.600 researching start with perplexity take
00:12:57.600 those data points or those links or
00:13:00.000 whatever it feeds me that I find to be
00:13:02.399 relevant and actually plug that back
00:13:05.279 into ChatGPT and say here’s the data
00:13:07.920 that I have or that I want to use please
00:13:11.360 like let’s make this more creative
00:13:13.519 depending on the task whatever it is the
00:13:15.519 prompt library but I use perplexity for
00:13:18.000 data mining and then ChatGPT for more of
00:13:21.360 the creative flow if you will
00:13:26.399 that’s actually my exact answer really
00:13:29.120 and I I think a lot of people you know
00:13:31.440 when we did kind of first iterations of
00:13:33.200 AI University throughout the agency many
00:13:35.440 did uncover that Perplexity was
00:13:38.800 significantly better with the research
00:13:40.760 function um people also liked the voice
00:13:44.000 function within Perplexity uh but
00:13:47.240 ChatGPT it stole everyone’s heart it was
00:13:49.920 the the first one it was the app that
00:13:52.240 people downloaded it was the interface
00:13:54.079 people just got really used to so I find
00:13:57.120 it’s just kind of what people gravitate
00:13:59.279 towards you you find yourself pulling
00:14:01.440 your pulling up ChatGPT more than
00:14:03.199 anything else um but I do think ChatGPT
00:14:06.079 is very creative
00:14:08.199 um really good with more of the one
00:14:11.120 pagers and the strategy documents and
00:14:13.600 also the more you use it the more it
00:14:15.600 understands you the more it understands
00:14:17.120 what you’re looking for and so you know
00:14:19.360 it’s a self-fulfilling prophecy you know
00:14:21.680 because I’m using ChatGPT more than
00:14:24.079 anything else it’s it’s over time
00:14:27.360 becoming the best version for me right
00:14:29.920 right what about anyone using Grock
00:14:33.680 haven’t touched it won’t touch it but
00:14:36.399 that’s that’s a different cover won’t
00:14:37.920 touch it I find you know for some
00:14:40.000 certain research projects I find it’s
00:14:41.920 its thinking can be better than ChatGPT
00:14:44.560 there were some scenarios because I
00:14:45.760 would do both um what about what about
00:14:49.199 Gemini from Google
00:14:52.240 i’ve dabbled in Gemini but I’ve never
00:14:55.440 loved it it was it was a short short
00:14:58.880 fling with I will agree with that but I
00:15:00.959 will say if you live in Google Drive
00:15:03.920 like we do they’re starting to really
00:15:06.480 implement Gemini in Google Drive so
00:15:09.440 depending on how that relationship
00:15:12.079 builds itself out to with my day-to-day
00:15:14.880 activities and behavior I could see
00:15:16.399 myself using it more if through Google
00:15:19.600 Drive and Sheets and whatever else it
00:15:21.839 becomes more intuitive with the data
00:15:23.920 that we already are using then who knows
00:15:26.720 but currently it’s not my favorite it it
00:15:29.839 um it’s not as great I think as the
00:15:31.760 other ones but I find it if you go into
00:15:33.680 NotebookLM Yeah and you can create a
00:15:36.480 podcast which I find fascinating
00:15:39.120 that’s crazy and you can you know you
00:15:40.959 can start to um you can start to ask
00:15:43.279 questions of those podcast hosts they’ll
00:15:45.680 hear your question stop what they’re
00:15:47.360 talking about and then start interacting
00:15:48.639 with you so and I’m sure you know it’s
00:15:51.519 not exactly Gemini but it’ll all get
00:15:53.199 related at some point you know on some
00:15:54.800 back end it’s similar like you said
00:15:56.480 Google Drive Microsoft Copilot I find
00:15:58.800 that that starts to pop up everywhere if
00:16:00.560 I’m in an Edge browser I’m doing Office
00:16:02.560 that starts to pop up that can be
00:16:04.399 helpful sometimes I don’t know if if you
00:16:06.160 two have used much of that
00:16:09.279 no no okay and then Claude I feel like a
00:16:12.720 lot of people love Claude a lot of
00:16:15.120 people love Claude that’s the the tech
00:16:16.959 darling for sure um I personally don’t
00:16:19.839 know anyone at the agency who really
00:16:21.519 stuck with Claude in any way but I do
00:16:24.399 think I think Curtis is a huge Claude
00:16:28.800 yeah
00:16:30.240 but one one thing I would like to point
00:16:32.000 out which I think is interesting and and
00:16:34.160 Kyle can probably elaborate more but one
00:16:36.160 of the products we were test running to
00:16:37.759 me is kind of more so the future of AI
00:16:41.279 which is one system that gives you
00:16:43.440 access to all of these LLMs so you feed
00:16:46.720 in your one question and then you can
00:16:48.800 toggle between what the cloud response
00:16:50.720 is what the perplexity response is what
00:16:52.880 the ChatGPT response is so that you have
00:16:55.839 more of an easy navigation to you know
00:16:58.399 all the variables that exist which I
00:17:01.199 think we’ll see a lot more often as new
00:17:03.920 products come out there’s a company I
00:17:05.520 just heard about yesterday it’s called
00:17:07.119 Profound and their idea is getting your
00:17:10.799 brand featured in the AI responses so
00:17:14.000 you know before it was like Google SEO
00:17:15.839 so now it’s like oh when someone asks a
00:17:17.760 question in ChatGPT or Perplexity let’s
00:17:20.240 make sure your brand is the one that
00:17:21.760 gets mentioned and I think that’s going
00:17:23.919 to be a fascinating world of well how do
00:17:26.559 they even do that how do you how do you
00:17:28.960 infiltrate the back end to pop up you
00:17:31.200 know Courtney’s Bakery or whatever it is
00:17:34.320 well I was
00:17:35.559 thinking I was thinking about it within
00:17:37.679 our own industry because we always talk
00:17:39.760 about data and we always really go after
00:17:42.000 the quirky data and I was curious one
00:17:44.160 day and I I put in the prompt you know I
00:17:46.559 wanted to see if one of our clients
00:17:47.919 would pull as the response because we
00:17:50.240 had some of the questions that no one
00:17:52.000 else had and so I put in a specific
00:17:55.360 question and it did in fact pull our
00:17:57.360 previous client and the data that we had
00:17:59.760 created on their behalf so for us
00:18:02.480 specifically not on the marketing side
00:18:04.080 but the PR side it does infiltrate how
00:18:07.120 we should be thinking about you know the
00:18:09.039 messaging the talking points the data
00:18:10.880 that we’re creating because exactly to
00:18:13.520 your point if you do have someone with a
00:18:15.520 very specific question in an AI platform
00:18:18.240 and it pulls your client’s data that’s
00:18:20.640 gold in the future
00:18:23.120 yeah I was just going to say I was on my
00:18:24.799 first client call today actually and it
00:18:26.799 was the first time I heard that a new uh
00:18:29.280 survey was released and it pulled the
00:18:32.160 client survey data and it wasn’t like
00:18:34.160 they didn’t ask for it to pull that
00:18:35.520 specific data but it organically came up
00:18:38.320 and they were like that felt
00:18:39.840 gamechanging because it felt like this
00:18:42.240 is where people are going to not search
00:18:45.039 but get organic information in a
00:18:47.840 conversational way which is very
00:18:50.160 different than Google SEO has ever been
00:18:52.400 in the past
00:18:54.640 break that down again so so tell me so
00:18:56.559 what so what was the difference just for
00:18:58.160 people who are listening that like they
00:18:59.440 want to clearly get like so what was new
00:19:01.360 what what yeah sure so I’m not exactly
00:19:04.160 sure what the client data was but
00:19:06.400 essentially um we were the the client
00:19:09.840 was asking questions not related to what
00:19:12.799 they found in their own data but just an
00:19:14.880 industry type question and it started
00:19:17.760 feeding their latest report which I
00:19:20.000 think was less than 24 hours old um back
00:19:23.360 to them as conversational
00:19:25.799 information with cited sources and it
00:19:28.880 happens to be their own data um so when
00:19:32.559 you’re going in and you’re having these
00:19:34.080 conversations about how do I and this is
00:19:36.240 oversimplified but how do I bake a
00:19:38.400 brownie and it pulls something that’s
00:19:41.360 from you know a top chef for example
00:19:44.880 that’s great because that’s what people
00:19:46.240 are going to look for because they’re
00:19:47.440 not going to go in and Google that
00:19:48.880 they’re going to have a conversation
00:19:50.480 with ChatGPT first and say “What do I
00:19:53.120 need to do to make this happen” and
00:19:55.360 whether you know it or not you’re
00:19:56.720 getting data from somewhere like you’re
00:19:58.799 getting that information from somewhere
00:20:00.480 so I really think that’s going to be the
00:20:02.080 gold mine moving
00:20:04.120 forward I see I see yeah no I agree I’ve
00:20:07.200 been using it a lot more now for
00:20:08.960 questions I have like hey you know my
00:20:11.679 Sonos speakers at home they’re not
00:20:13.440 showing up on my app what do I do it’s
00:20:16.080 I’m not googling it i’m asking AI and
00:20:17.600 it’ll just say okay like go and do these
00:20:19.360 things like boom here’s your answer this
00:20:20.720 is how you’re going to fix it right or
00:20:22.240 just other there were just something
00:20:23.760 like okay there’s some tech question I
00:20:25.280 forget what it was like hey how do I fix
00:20:27.039 this other issue in my house and it’s
00:20:29.200 just like okay boom just go do these
00:20:30.799 things so I’m not googling anything
00:20:32.159 right so you do wonder how how will
00:20:33.840 people make money in that situation hey
00:20:35.840 there’s no ad to click on what’s that
00:20:37.520 going to look like um I’m I’m curious
00:20:40.240 though outside of let’s say these large
00:20:41.760 language models a lot of the the tech
00:20:43.280 stuff is anyone playing with video
00:20:46.480 content creation audio content creation
00:20:49.200 like you know editing tools a lot of
00:20:50.880 this stuff
00:20:51.919 A lot of our job is to make sure things
00:20:53.679 become good video good audio and at some
00:20:56.240 level a podcast like this AI could do
00:20:58.480 that podcast right AI could do a news
00:21:00.159 segment has anyone messed with any of
00:21:02.080 those kinds of things i would say it
00:21:04.960 doesn’t necessarily come internally but
00:21:06.960 we feel these questions a lot with
00:21:08.799 clients since they’re the ones usually
00:21:10.880 delivering you know B-roll social media
00:21:13.520 content things of that nature there’s
00:21:15.760 questions about is this allowed um would
00:21:19.280 a reporter or a news anchor think less
00:21:23.440 of a potential segment with us if we’re
00:21:25.600 giving AI generated content questions
00:21:28.159 like that are bubbling up quite often
00:21:30.080 very specifically we used to work with
00:21:33.360 um a client and a video vendor on you
00:21:36.799 know standard you know video content
00:21:38.960 creation and quotes would be anywhere
00:21:40.880 between $5,000 and $10,000 uh and of course
00:21:44.159 one very smart CEO came around within
00:21:47.039 about 24 hours and said you know we got
00:21:50.320 off that brainstorm and I saw the price
00:21:52.480 quote and I decided to play with AI and
00:21:55.039 this was the end result and I think it
00:21:57.520 looks pretty good and honestly we agreed
00:22:01.280 so there is definitely a world where
00:22:03.520 that is happening you will have
00:22:05.600 reporters who do you know think less of
00:22:08.720 AI generated photos video content they
00:22:11.360 might go after another story because it
00:22:13.280 is you know more traditional assets but
00:22:17.039 with the cost savings that we’re seeing
00:22:19.200 clients are moving towards that level of
00:22:21.679 content creation more and more are are
00:22:24.240 clients saying hey why do I even need a
00:22:25.840 PR agency I can just ask you know ChatGPT
00:22:28.480 for a PR strategy and reporters to
00:22:31.039 pitch and write the pitch for me right
00:22:33.120 like I’m sure people think that
00:22:36.720 people definitely think that but
00:22:38.280 surprisingly it hasn’t come up in any of
00:22:41.280 our conversations um I think people
00:22:44.320 still realize one you know AI just can’t
00:22:46.799 be trusted yet with the outputs uh I’ve
00:22:49.679 seen a lot of errors and and things
00:22:51.760 going wrong within searches um we’ve
00:22:54.880 also seen a lot of horror stories hit
00:22:56.960 media and LinkedIn about you know
00:22:59.440 pitches gone wrong they strike the wrong
00:23:02.080 tone they’re really aggressive a whole
00:23:04.559 agency get gets blacklisted because of a
00:23:07.120 a pitch that went south um so I think
00:23:10.400 even clients they still see that right
00:23:13.360 now AI is not ready um and of course if
00:23:16.400 you ask me obviously I have skin in the
00:23:18.240 game but reporters are human and they’re
00:23:20.960 always going to want to deal with humans
00:23:22.720 on the other side as well can we
00:23:24.640 optimize with AI to make it a lot more
00:23:26.559 efficient definitely I want to word this
00:23:29.200 very carefully all right let’s let’s
00:23:31.360 give you time let’s get you make sure
00:23:32.720 you got your careful words here I want
00:23:34.159 to word this very carefully but I will
00:23:35.679 say I think it’s no secret to all of us
00:23:38.960 that some people just don’t fully
00:23:41.600 understand PR even if they are engaged
00:23:43.679 in a PR contract they’re not really sure
00:23:45.760 what we do or how we do it or why it’s
00:23:48.159 important and there are other clients on
00:23:50.880 the flip side that fully understand the
00:23:53.039 power of PR and what we do and I think
00:23:56.559 something that cannot be underscored and
00:23:59.039 this is an I don’t remember where I
00:24:00.960 think maybe it was the New York Times
00:24:02.720 but it was a podcast I was listening to
00:24:04.400 where it says because information is so
00:24:07.840 readily available now the importance of
00:24:11.760 relationships and true connection based
00:24:15.200 on recommendation or word of mouth is
00:24:19.279 actually going to come become full
00:24:21.039 circle and become more powerful because
00:24:25.240 everyone you know broadly speaking has
00:24:28.720 the same power at their fingertips if
00:24:31.039 you know how to use it but the
00:24:32.960 relationships the the one-on-one
00:24:34.960 connections the the power of the years
00:24:38.720 of experience that we have collectively
00:24:41.120 and how to get certain things placed or
00:24:44.080 pitched or what have you that that’s
00:24:47.039 never going to be an output that AI can
00:24:50.320 can comprehend or deliver to you because
00:24:54.320 that personalized touch will never be
00:24:56.880 there so I think it’s understanding how
00:25:00.559 PR works but also understanding there
00:25:03.120 are just certain things that AI will not
00:25:05.279 be able to deliver that people inside an
00:25:08.159 agency will I like that yeah it’s almost
00:25:11.120 like there’s so much information that
00:25:13.279 having information doesn’t matter
00:25:14.720 anymore it’s it’s it’s about the human
00:25:17.840 touch or like you know like they can’t
00:25:19.360 replicate your connections your network
00:25:21.679 your relationships that kind of thing
00:25:24.960 not yet not yet it’s been fascinating to
00:25:27.200 watch it’s interesting to watch how more
00:25:28.880 companies are are taking on initiatives
00:25:30.400 like what you two are doing there was a
00:25:32.320 friend of mine works at another agency
00:25:34.400 and and they got the sense from
00:25:36.320 management there basically that is we’re
00:25:38.240 not hiring any entry-level people
00:25:39.600 anymore like AI is the entry level
00:25:41.360 person so we’re not going to hire
00:25:42.960 someone so that they can learn when all
00:25:45.039 they’re going to do is the stuff that AI
00:25:46.480 can do so it’s this weird thing now well
00:25:47.840 how are you going to get experience you
00:25:49.679 can’t get started
00:25:51.600 you know they only want veteran people
00:25:52.960 now because veteran people have
00:25:54.320 relationships have a network have
00:25:55.760 connections have all that but just to
00:25:57.440 come in and say I’ll help you with
00:25:58.720 information that’s not a job at least
00:26:00.720 right now
00:26:02.799 there’s also a lot of conversations
00:26:04.159 right now about the death of middle
00:26:05.840 management yeah and I asked the same
00:26:07.600 question I said okay if you look at the
00:26:09.120 corporate ladder and you remove middle
00:26:11.080 management how do we get people to be
00:26:14.320 above middle management and and no one
00:26:16.559 has an answer I’ve asked many very
00:26:18.559 intelligent thought thoughtful people
00:26:21.279 and no one’s been able to give me
00:26:22.960 something smart as a response well when
00:26:25.440 you find out come back we’ll do the
00:26:27.600 podcast again like Courtney found out
00:26:29.279 the information she figured out how
00:26:30.880 we’re going to get up the career ladder
00:26:32.240 here by jumping from entry level to CEO
00:26:35.279 I don’t do middle management I just
00:26:36.720 jumped ahead yeah apparently although
00:26:39.279 the data is currently showing GenZ
00:26:41.279 doesn’t want leadership or C-suite at
00:26:43.520 all so that’s another conversation yeah
00:26:46.320 I was going to say as a middle manager
00:26:48.240 like I would love to have that
00:26:49.360 conversation let’s figure it out yeah
00:26:51.799 exactly this is great Kyle, Courtney
00:26:54.640 awesome awesome time thank you so much
00:26:56.480 for sharing your experiences I need to
00:26:59.279 do more with perplexity that’s the one
00:27:00.880 actually I’ve I haven’t done enough with
00:27:02.240 so it’s interesting to hear what you
00:27:03.200 guys are saying about it so like okay
00:27:04.480 that’s my that’s my to-do list here when
00:27:06.320 we’re done definitely cross reference
00:27:08.720 and let us know what you define as your
00:27:11.360 favorite after you experiment for sure
00:27:13.600 for sure awesome thanks so much Kyle and Courtney
00:27:15.840 thank you
00:27:17.840 thank you to my guest and thanks for
00:27:19.520 listening subscribe to get the latest
00:27:21.360 episodes each week and we’ll see you
00:27:23.200 next time