Latest / The STANDOUT Agency Podcast / EPISODE 10: Why Your Agency Needs SOPs More Than It Needs AI
Transcript
- 0:04SPEAKER_00: Welcome to The STANDOUT Agency Podcast, for agency owners who want to grow, not just survive.
- 0:09SPEAKER_01: I'm Gareth Healey.
- 0:10SPEAKER_01: I'm an ex-agency owner and author of STANDOUT OR DIE.
- 0:14SPEAKER_01: I help agency founders see where they're still a bottleneck in their own business
- 0:17SPEAKER_01: and make sure they do something about it.
- 0:19SPEAKER_00: And I'm Callum Healey.
- 0:20SPEAKER_00: I find where agencies are losing hours to manual work and then build AI systems
- 0:24SPEAKER_00: that get the time back.
- 0:27SPEAKER_01: Each week, one of us interviews the other on a subject related to agency growth.
- 0:31SPEAKER_01: Everything we get into comes straight out of our own experience and client work.
- 0:34SPEAKER_01: Occasionally, we bring in some guests to add their expert perspective on a topic.
- 0:40SPEAKER_00: Welcome back to The STANDOUT Agency Podcast.
- 0:42SPEAKER_00: This is episode 10, why your agency needs SOPs before it needs AI.
- 0:48SPEAKER_00: Episode 10, so 10 episodes in, we're doing well so far.
- 0:52SPEAKER_00: We are.
- 0:52SPEAKER_00: We're halfway to the magic number, aren't we?
- 0:54SPEAKER_00: Well, nearly.
- 0:55SPEAKER_00: Yeah, we were saying that in episode one,
- 0:58SPEAKER_00: Doing 21 podcast episodes means you're in some of the top percentiles
- 1:03SPEAKER_00: of podcasters worldwide.
- 1:05SPEAKER_00: So we're on track.
- 1:05SPEAKER_00: I mean, we're on track to get about halfway nearly.
- 1:08SPEAKER_02: So let's keep going.
- 1:09SPEAKER_02: Yeah, this one, really interesting.
- 1:12SPEAKER_02: So why your agency needs SOPs more than it needs AI.
- 1:17SPEAKER_02: Callum, you've worked in AI for a while now.
- 1:22SPEAKER_02: You're obsessed with it.
- 1:24SPEAKER_02: Yeah.
- 1:25SPEAKER_02: You're also...
- 1:28SPEAKER_02: It's a bit of a chip off the old block in terms of suddenly becoming obsessed
- 1:31SPEAKER_02: with agencies as well.
- 1:33SPEAKER_01: But you've never actually worked in an agency.
- 1:35SPEAKER_01: You've met
- 1:37SPEAKER_02: with about three this week.
- 1:39SPEAKER_02: You've never actually worked with an agency,
- 1:41SPEAKER_02: but you're constantly talking to me about SLPs in agencies.
- 1:49SPEAKER_02: Why is that?
- 1:49SPEAKER_02: I thought you were obsessed with AI, but you seem to be equally obsessed with SLPs.
- 1:54SPEAKER_00: Yeah,
- 1:54SPEAKER_00: SOPs is an interesting one and a year ago I couldn't tell you exactly what they are
- 2:02SPEAKER_00: or what they were but it's a phrase that keeps coming up because AI and SOPs kind
- 2:09SPEAKER_00: of go hand in hand and it simply is a process,
- 2:13SPEAKER_00: a workflow and I've soon found that it isn't actually that common either.
- 2:19SPEAKER_00: So it's one that I keep saying to agencies because for AI to work,
- 2:24SPEAKER_00: it needs a very defined process to understand what exactly the task is
- 2:29SPEAKER_02: that you want it to do.
- 2:30SPEAKER_02: Totally, yeah.
- 2:31SPEAKER_02: I think SOP is an acronym of Standard Operating Procedure, so it certainly is not a common task.
- 2:39SPEAKER_02: Acronym or Standard Operating Procedure is not a common phrase with an agency.
- 2:43SPEAKER_02: It's processes.
- 2:45SPEAKER_02: We've got processes.
- 2:47SPEAKER_02: We have workflows probably even more common than that.
- 2:50SPEAKER_02: And they're maybe not exactly the same thing, but they're in the same ballpark.
- 2:54SPEAKER_02: All those things, aren't they?
- 2:55SPEAKER_02: It's having,
- 2:55SPEAKER_02: as you just said,
- 2:56SPEAKER_02: a standard, a clear, documented in some way, standard way of doing things.
- 3:04SPEAKER_02: That's repeated
- 3:05SPEAKER_00: and adhered to.
- 3:05SPEAKER_00: It's not...
- 3:08SPEAKER_00: As common in the sort of creative industry,
- 3:11SPEAKER_00: in the marketing industry, as it would be manufacturing, let's say.
- 3:17SPEAKER_00: I suppose with products you have to get everything measured correctly and there's
- 3:25SPEAKER_00: a slightly more risk involved in some of these businesses that if they don't follow
- 3:29SPEAKER_00: the operating procedure then things would go wrong and further down
- 3:32SPEAKER_00: the production line or whatever it might be.
- 3:34SPEAKER_00: But creative industries I've noticed that it's tacit knowledge,
- 3:40SPEAKER_00: the processes tend to be knowledge in people's heads.
- 3:44SPEAKER_00: They've accumulated a lot of skill in their certain role
- 3:49SPEAKER_00: and they know how work should get done and their leadership and the founders and
- 3:54SPEAKER_00: the person who hires team members trust them to do that task.
- 3:59SPEAKER_00: If anything,
- 4:00SPEAKER_00: I think it can be a benefit,
- 4:01SPEAKER_00: but before AI, it allowed for a bit of diversity in outputs.
- 4:07SPEAKER_00: You have three people doing the same task,
- 4:09SPEAKER_00: but these people all have different strengths, so you can get a bit of their flair.
- 4:15SPEAKER_00: Particularly,
- 4:16SPEAKER_00: as you said,
- 4:17SPEAKER_02: on the more creative agency side,
- 4:20SPEAKER_02: on the creative agency section of agencies,
- 4:24SPEAKER_02: if you like,
- 4:27SPEAKER_02: Starting with a blank sheet of paper,
- 4:28SPEAKER_02: starting with a clean slate for fresh ideas without being bound by
- 4:35SPEAKER_02: a process is potentially a good thing.
- 4:39SPEAKER_02: It's certainly seen by many as a good thing.
- 4:43SPEAKER_02: But it's not good for training people.
- 4:49SPEAKER_02: It's not good for establishing accountability,
- 4:52SPEAKER_02: who's accountable for doing the work, at what stage.
- 4:56SPEAKER_02: Improving the process.
- 4:57SPEAKER_02: Improving the process.
- 4:58SPEAKER_02: It's certainly not good for efficiency, necessarily, because we don't know whether
- 5:05SPEAKER_02: Whether we are being efficient or not, whether we are spending too much time.
- 5:09SPEAKER_02: Consistency
- 5:10SPEAKER_02: as well, training new people,
- 5:13SPEAKER_02: people doing a process in the same way, following the same steps.
- 5:18SPEAKER_02: We are not for a second here saying, are we?
- 5:23SPEAKER_02: The creative processes should be so regimented that it just delivers
- 5:28SPEAKER_02: the same results, of course.
- 5:30SPEAKER_00: Of course not, no,
- 5:31SPEAKER_00: but for AI and when you're using Artificial Intelligence with your processes,
- 5:39SPEAKER_00: it doesn't know what they are.
- 5:40SPEAKER_00: It doesn't have that same nuanced understanding of the creative
- 5:44SPEAKER_00: and marketing industry that your team members do.
- 5:47SPEAKER_00: And so it needs every process in,
- 5:50SPEAKER_00: frankly,
- 5:51SPEAKER_00: painful detail,
- 5:52SPEAKER_00: like really holding its hand because it is somewhat of an alien in the form of,
- 5:57SPEAKER_00: it isn't a human intelligence, it's a machine intelligence.
- 6:03SPEAKER_00: It needs everything described to it very plainly so that it can help you out.
- 6:06SPEAKER_00: And the argument that we make with agencies that we work with is even before AI,
- 6:12SPEAKER_00: SOPs are very useful for a lot of the reasons that we just said there.
- 6:16SPEAKER_00: And there's a few angles that we've prepared to discuss on this and I'm excited
- 6:21SPEAKER_00: to get into them.
- 6:22SPEAKER_02: So we've established,
- 6:24SPEAKER_02: and I know this for having worked in the industry for a long time,
- 6:27SPEAKER_02: Processes, workflows, SOPs, not the usual, not common within agencies.
- 6:34SPEAKER_02: It's not the strongest suit of an agency generally to do that.
- 6:38SPEAKER_02: It was actually in my agency that I ran, a very strong suit of ours actually.
- 6:45SPEAKER_02: I've put strong processes down as one of the key factors in the success of my agency
- 6:52SPEAKER_02: at that time, actually.
- 6:54SPEAKER_02: Maybe that's why I'm such a big advocate.
- 6:55SPEAKER_01: Maybe, maybe it's just in the blood.
- 6:58SPEAKER_02: But I don't say that to blow me on Trump.
- 7:00SPEAKER_02: In fact, that was well before common use of AI, of course, but it's not everybody's bag.
- 7:07SPEAKER_02: There's clearly a rush here to embrace AI, to adopt AI by agencies.
- 7:13SPEAKER_02: Correct me if I'm wrong, Callum,
- 7:15SPEAKER_02: but you're Clearly you're an advocate of that, you're trying to help people do that.
- 7:19SPEAKER_02: But there's a stage prior to this, isn't it?
- 7:21SPEAKER_02: The title of this podcast is Your Agency Needs SOPs More Than It
- 7:26SPEAKER_01: Needs AI.
- 7:27SPEAKER_02: We're not saying it doesn't need AI, but it needs more.
- 7:29SPEAKER_00: I think it's because SOPs precede AI.
- 7:34SPEAKER_00: They need to come first.
- 7:37SPEAKER_00: When you were talking there,
- 7:39SPEAKER_00: actually,
- 7:39SPEAKER_00: one of the first steps,
- 7:41SPEAKER_00: and I do a lot of adoption work and auditing AI use and where it should help,
- 7:46SPEAKER_00: you need processes to actually know which ones are failing.
- 7:50SPEAKER_00: You need to know where the pain is first.
- 7:52SPEAKER_00: Before you even construct a new process with AI,
- 7:55SPEAKER_00: it's important to address,
- 7:57SPEAKER_00: okay, what is actually the problem in your agency right now?
- 8:00SPEAKER_00: And let's use this new technology to alleviate that pain.
- 8:05SPEAKER_00: And I think ultimately it's about removing error and establishing consistency.
- 8:14SPEAKER_00: Standardizing.
- 8:14SPEAKER_00: Standardizing,
- 8:15SPEAKER_00: that's why it's called standard operating procedure because it's standardized and
- 8:20SPEAKER_00: before AI and I'm sure a big success of the agency that you ran was the fact
- 8:26SPEAKER_00: that it removes a lot of human error.
- 8:29SPEAKER_00: People know how work gets done.
- 8:32SPEAKER_00: They replicate the same process each time.
- 8:34SPEAKER_00: That actually means that they get better at replicating the process
- 8:38SPEAKER_00: because they're used to it, they do it more.
- 8:41SPEAKER_00: Perhaps that process gets improved over time as well.
- 8:45SPEAKER_00: In this case with AI,
- 8:46SPEAKER_00: it's removing There's a lot about prompt libraries,
- 8:48SPEAKER_00: that was something that early on in generative AI,
- 9:11SPEAKER_00: It was perhaps even a big job role of the future where people were thinking, oh,
- 9:16SPEAKER_00: we should really study how to prompt AI correctly and you're going
- 9:21SPEAKER_00: to get paid loads.
- 9:22SPEAKER_00: This is going to be a sought after job role that's since died out.
- 9:26SPEAKER_00: I don't know if anyone listening has seen those online.
- 9:30SPEAKER_00: I can't see them anymore because simply,
- 9:33SPEAKER_00: Workflows are a step above that and they're better
- 9:37SPEAKER_00: because they're packaged prompt libraries.
- 9:41SPEAKER_00: It's a standardized process, but AI knows each step.
- 9:45SPEAKER_00: So instead of you telling it what to do for a prompt,
- 9:48SPEAKER_00: you standardize what it should do and then it repeats that process each time.
- 9:52SPEAKER_00: So this is great,
- 9:53SPEAKER_00: I think,
- 9:54SPEAKER_00: because instead of sharing massive books of prompt libraries
- 9:58SPEAKER_00: and Delivering them throughout your team and trying to communicate and keep track of
- 10:03SPEAKER_00: which prompt do I use for this process.
- 10:06SPEAKER_00: You configure that.
- 10:07SPEAKER_00: A lot of the systems we both have,
- 10:10SPEAKER_00: we just say, off you go, because it knows the process already.
- 10:15SPEAKER_00: It removes that human error.
- 10:17SPEAKER_00: It's a big part of it.
- 10:19SPEAKER_02: Cool.
- 10:20SPEAKER_02: I know this because you've talked to me about it a lot,
- 10:22SPEAKER_02: but you've talked to a number of agencies and And come and spoke to me afterwards
- 10:30SPEAKER_02: and say,
- 10:30SPEAKER_02: you know,
- 10:30SPEAKER_02: I'm trying to,
- 10:31SPEAKER_02: they want,
- 10:32SPEAKER_02: they're desperate,
- 10:32SPEAKER_02: of course, or keen to deploy AI or embrace AI more in their agency.
- 10:37SPEAKER_02: Of course they are.
- 10:38SPEAKER_02: But I've talked, you've talked to them about SOPs and they don't have them.
- 10:46SPEAKER_02: They're asking either, what's an SOP?
- 10:49SPEAKER_02: Or even if they know the term, well, we don't really have it.
- 10:52SPEAKER_02: It's in people, it's in John's head or Jane's.
- 10:57SPEAKER_00: I've had a founder explicitly say to me that isn't terminology we use
- 11:02SPEAKER_00: in our business and when you're auditing a business and you're asking
- 11:06SPEAKER_00: about these processes because that's what is needed for me to help them embed those
- 11:11SPEAKER_00: into AI systems because that's what AI needs,
- 11:14SPEAKER_00: it kind of leaves you speechless a little bit
- 11:17SPEAKER_00: where it's like okay well let's Figure out how to find those.
- 11:22SPEAKER_02: Yeah, which is perfectly, that was exactly the question I was going to ask.
- 11:26SPEAKER_02: So if that happens, we've seen it happen.
- 11:28SPEAKER_02: And we understand it, particularly, as we've said, in the more creative side of the sector.
- 11:35SPEAKER_02: And when that's the case,
- 11:37SPEAKER_02: you know what,
- 11:37SPEAKER_02: we don't really have,
- 11:38SPEAKER_02: we have processes,
- 11:39SPEAKER_02: but they're just in people's heads or we do it sort of by,
- 11:43SPEAKER_02: it just gets done by culturally,
- 11:47SPEAKER_02: We talk about it in the business and it's not particularly written down anywhere,
- 11:50SPEAKER_02: but the work gets done and we do a good job, of course.
- 11:53SPEAKER_02: But when that happens, we don't have any processes
- 11:56SPEAKER_00: to show you, Callum.
- 11:57SPEAKER_00: What do we do?
- 11:58SPEAKER_00: Yeah, like I said, it's a bit of a speechless moment.
- 12:03SPEAKER_00: Actually tailored my own offers to this after learning that it wasn't a massive,
- 12:09SPEAKER_00: they weren't as popular as I thought they'd be SOPs.
- 12:12SPEAKER_00: And I have a part of my kind of, one of my services that I do is called a sprint.
- 12:20SPEAKER_00: And before that sprint,
- 12:22SPEAKER_00: which is simply we construct a proprietary workflow together An agency comes
- 12:28SPEAKER_00: to me saying they want this system.
- 12:30SPEAKER_00: We do a blueprint session and it's absolutely mandatory because we need
- 12:36SPEAKER_00: to actually define what this build should be and what the process is and a lot
- 12:41SPEAKER_00: of times it's undefined and it's quite satisfying.
- 12:48SPEAKER_00: It's a nice part of my job.
- 12:50SPEAKER_00: I enjoy kind of getting into processes and mapping them out.
- 12:55SPEAKER_00: It's part of the brief actually,
- 12:57SPEAKER_02: isn't it?
- 12:57SPEAKER_02: Agency's clients should provide the agency a brief as to what they want to do,
- 13:03SPEAKER_02: not AI necessarily here.
- 13:06SPEAKER_02: And agencies often lament the fact that they don't get a brief or when they get
- 13:10SPEAKER_02: a brief it's unclear.
- 13:13SPEAKER_02: So it's that same principle, isn't it?
- 13:15SPEAKER_02: What are we trying to achieve here?
- 13:18SPEAKER_02: What pain point,
- 13:19SPEAKER_02: what issue,
- 13:20SPEAKER_02: what opportunity are we trying to We need to know what the process is first.
- 13:24SPEAKER_02: If you're not clear, you need to get it right.
- 13:34SPEAKER_00: I've had briefs that are fantastic actually,
- 13:37SPEAKER_00: developers briefs even,
- 13:39SPEAKER_00: where it's multiple pages and in great detail and those are usually the agencies
- 13:44SPEAKER_00: who are already using AI well because they're able to generate a massive document
- 13:50SPEAKER_00: at such speed and it's very clear.
- 13:52SPEAKER_00: But if the process doesn't exist, I look towards the Most knowledgeable person on that process.
- 14:00SPEAKER_00: So we define what sector,
- 14:03SPEAKER_00: what area of the business this process relates to,
- 14:06SPEAKER_00: where is the pain point and then we tend to look at the department leader
- 14:12SPEAKER_00: and usually they're the more senior team member and they've done this process
- 14:18SPEAKER_00: for years.
- 14:19SPEAKER_00: A lot of the time they don't actually do it anymore,
- 14:22SPEAKER_00: hands on, they're the one that's teaching others.
- 14:25SPEAKER_00: Get them to teach me in these blueprint sessions.
- 14:28SPEAKER_00: I sit down and we really start small and go, what is the complete first step?
- 14:36SPEAKER_00: And then we go, right, what is it exactly that you do after that?
- 14:40SPEAKER_00: And then we simply sit down for an hour, hour and a half.
- 14:43SPEAKER_00: And by the end of it, there's five to ten steps.
- 14:47SPEAKER_00: We know exactly the parts that are being slowed down and the parts that AI should interject.
- 14:54SPEAKER_00: I love the phrase that you should learn through AI rather than from it.
- 14:59SPEAKER_00: AI's data sets are completely massive,
- 15:03SPEAKER_00: you know, you can't even comprehend how much data they ingest.
- 15:08SPEAKER_00: And that's true for agency data as well.
- 15:11SPEAKER_00: There is great data that it learns from and there's also quite rubbish data in terms
- 15:16SPEAKER_00: of its theory and how agency work should get done.
- 15:19SPEAKER_00: And so you'll likely get very average outputs if you try to get AI to build you a process.
- 15:25SPEAKER_00: But you should tell it how to build your process because
- 15:28SPEAKER_00: if you want your agency's kind of quality standard,
- 15:33SPEAKER_00: you need to give it your process and your data.
- 15:35SPEAKER_00: And AI is simply a...
- 15:38SPEAKER_00: A replication machine,
- 15:39SPEAKER_00: a plagiarism machine, if you will, where it regurgitates the facts that you give it.
- 15:44SPEAKER_00: So if we get the gold standard process from that senior team member who knows
- 15:49SPEAKER_00: the process the best,
- 15:51SPEAKER_00: it's very helpful because AI can simply help people learn that process.
- 15:59SPEAKER_00: Senior team leaders or team members' brain has been downloaded into the system
- 16:05SPEAKER_00: and people can actually ask AI, but it's like speaking to the leader in a way.
- 16:10SPEAKER_00: They learn that process and AI says it the same every time.
- 16:14SPEAKER_00: It learns the way it should get done.
- 16:16SPEAKER_02: But it's documented and by nature of AI then...
- 16:23SPEAKER_02: I've seen this many times,
- 16:25SPEAKER_02: I must have done it myself as well,
- 16:26SPEAKER_02: many times,
- 16:28SPEAKER_02: where the senior person in an area,
- 16:31SPEAKER_02: the department head,
- 16:32SPEAKER_02: the leader,
- 16:33SPEAKER_02: somebody that is experienced and does this thing often has done it many times
- 16:39SPEAKER_02: before.
- 16:41SPEAKER_02: Again, setting AI aside for one second, they train people.
- 16:46SPEAKER_02: When new people come into the business,
- 16:48SPEAKER_02: they will probably show people how to do something and do that maybe once or twice
- 16:54SPEAKER_02: a year, maybe 10 times a month.
- 16:57SPEAKER_02: The problem with that,
- 17:00SPEAKER_02: if it's not written down and documented,
- 17:02SPEAKER_02: as we know, human nature, you're a psychologist.
- 17:05SPEAKER_02: If it's not written down and documented,
- 17:06SPEAKER_02: people listen and understand about the process and hopefully get it and start
- 17:10SPEAKER_02: to use it, but they will add their own particular
- 17:19SPEAKER_02: They will do things in a slightly different way.
- 17:22SPEAKER_02: It could be cutting corners.
- 17:23SPEAKER_00: It could actually be making certain steps longer than they need to be,
- 17:27SPEAKER_00: over-servicing and things like that.
- 17:28SPEAKER_02: Yeah, absolutely.
- 17:30SPEAKER_02: And therefore,
- 17:31SPEAKER_02: if you've got 3, 4, 5,
- 17:32SPEAKER_02: 10 people,
- 17:33SPEAKER_02: 30 people all doing that,
- 17:35SPEAKER_02: suddenly you're Maybe the core process is there,
- 17:39SPEAKER_02: but it's got 10 different flavors, slightly nuanced flavors.
- 17:44SPEAKER_02: What we're trying to do is standardizing,
- 17:46SPEAKER_02: standard operating procedures so that AI can deploy it and use it in the same way.
- 17:51SPEAKER_00: Yeah, I've noticed that when I do speak to these experts on the process,
- 17:58SPEAKER_00: Unsurprisingly,
- 17:59SPEAKER_00: they know it very well and they can just talk me through it and take me through
- 18:03SPEAKER_00: the process.
- 18:03SPEAKER_00: There's no problem with defining the steps.
- 18:08SPEAKER_00: The fact that they even need to talk me through it is an issue
- 18:12SPEAKER_00: because it's not been standardised.
- 18:15SPEAKER_00: Because they can't show it to you.
- 18:17SPEAKER_00: Exactly.
- 18:18SPEAKER_00: And they could have it written down and documented and where it could be held on
- 18:23SPEAKER_00: a wall of the office and people can look at it and go,
- 18:25SPEAKER_00: right, oh yeah, of course, that's how it gets done.
- 18:28SPEAKER_00: But frankly, a lot of the time it isn't.
- 18:32SPEAKER_00: It's great to have it standardised before AI because you can train team members
- 18:36SPEAKER_00: and people who are new to processes,
- 18:39SPEAKER_00: but now AI is here,
- 18:40SPEAKER_00: it can act as that voice of reason and making people accountable and aware
- 18:46SPEAKER_00: and adhering to the process that has been documented and AI can be
- 18:51SPEAKER_00: the knowledge base for that.
- 18:54SPEAKER_02: When I ran my business,
- 18:57SPEAKER_02: I said earlier that we were a very good process,
- 18:59SPEAKER_02: which is true,
- 19:00SPEAKER_02: but also it's interesting you said about how things can get added on
- 19:04SPEAKER_02: that are unnecessary.
- 19:06SPEAKER_02: I did experience that a number of times.
- 19:07SPEAKER_02: We had great processes,
- 19:08SPEAKER_02: but there was many occasions where somebody new came into the business,
- 19:12SPEAKER_02: for example, and asked a question, why do we do that?
- 19:15SPEAKER_02: Why do we do step A, B and C when we could just do A and C?
- 19:20SPEAKER_02: We don't really need B.
- 19:22SPEAKER_02: There's a number of times where we all step back and say,
- 19:25SPEAKER_02: you know what,
- 19:26SPEAKER_02: we used to do B because of this reason,
- 19:28SPEAKER_02: because the client wanted it for this reason, but now they don't.
- 19:32SPEAKER_02: So why do we still continue to do efficiency?
- 19:36SPEAKER_02: That's a really interesting take.
- 19:38SPEAKER_02: Processes can always be refined, can't they?
- 19:41SPEAKER_02: But if an agency is struggling,
- 19:44SPEAKER_02: sitting down for an agency person generally,
- 19:47SPEAKER_02: particularly if they're more creative,
- 19:48SPEAKER_02: but even more analytical people, They're busy, they've got client demands.
- 19:57SPEAKER_02: If you say to them, right,
- 19:58SPEAKER_02: what we need to do is we need to sit down and write out your processes,
- 20:02SPEAKER_02: I guess not many people are jumping at the opportunity to do that, are they?
- 20:06SPEAKER_00: No, no.
- 20:07SPEAKER_00: They just want that pain point removed and AI helping them out before,
- 20:11SPEAKER_00: but that's why this title is so important because it's a prerequisite.
- 20:17SPEAKER_02: So how do you go about helping them
- 20:19SPEAKER_00: or encouraging them?
- 20:20SPEAKER_00: Yeah, so the method,
- 20:23SPEAKER_00: as we've said before,
- 20:24SPEAKER_00: is starting from the very beginning and instead of focusing on the process first,
- 20:31SPEAKER_00: we go, where is the pain?
- 20:33SPEAKER_00: What is it that you need AI to come and help you out with?
- 20:38SPEAKER_00: It's never the same answer for everyone.
- 20:41SPEAKER_00: Every agency has a different pain point.
- 20:45SPEAKER_00: Weirdly, timesheets is one that's very common.
- 20:48SPEAKER_00: There's a few similarities in pain points.
- 20:51SPEAKER_00: Timesheets are definitely painful for
- 20:53SPEAKER_02: everybody.
- 20:54SPEAKER_02: In different ways, maybe.
- 20:56SPEAKER_00: But with AI as well,
- 20:57SPEAKER_00: some of these softwares,
- 20:58SPEAKER_00: even though everyone's using a different one, they all seem to be suboptimal.
- 21:02SPEAKER_00: But anyway, we find the pain point and start there.
- 21:07SPEAKER_00: Can AI help people...
- 21:10SPEAKER_02: Right, create the process, or document the process, I guess it can, can't it?
- 21:17SPEAKER_00: Yeah, I have a kind of step-by-step method that I use in Blueprint sessions,
- 21:23SPEAKER_00: and there's two types of data, primarily.
- 21:26SPEAKER_00: There's quantitative,
- 21:27SPEAKER_00: and it's everything that's written down, it's explicit, it's files, it's frameworks.
- 21:33SPEAKER_02: Of course you can.
- 21:34SPEAKER_00: SOPs would be a quantitative measure of data,
- 21:38SPEAKER_00: but even if you don't have the process standardized,
- 21:42SPEAKER_00: you will have files that relate to the process.
- 21:44SPEAKER_00: It might be spreadsheets, tables, or steps of it.
- 21:48SPEAKER_00: And then there's qualitative,
- 21:50SPEAKER_00: and that's the type of data that's usually harder to derive
- 21:54SPEAKER_00: because it's nuanced data, it's It's intuitive.
- 21:59SPEAKER_00: It's open to interpretation.
- 22:02SPEAKER_00: It ultimately lives in people's heads.
- 22:04SPEAKER_00: That's what we're trying to get out.
- 22:07SPEAKER_00: A great way of establishing your process,
- 22:10SPEAKER_00: if you haven't already,
- 22:12SPEAKER_00: is the end goal is to give it to AI, but AI can help you get there.
- 22:17SPEAKER_00: And if you give it all of the quantitative data that you have and then go, okay, AI,
- 22:21SPEAKER_00: I've given you some background information,
- 22:24SPEAKER_00: maybe there's a few prompts along the way saying, oh, we do X, Y, and Z.
- 22:28SPEAKER_00: What is it that you're still unclear of?
- 22:30SPEAKER_00: You can ask AI directly because the way that it works is by pulling in the context
- 22:36SPEAKER_00: that you give it and then making judgements based on that context and
- 22:39SPEAKER_00: so it will tell you where the gaps are.
- 22:41SPEAKER_00: So using the interview prompt maybe to say interview me on what you'll do.
- 22:49SPEAKER_00: You know,
- 22:50SPEAKER_00: in AI audits and blueprint sessions,
- 22:52SPEAKER_00: I'll be the one asking the questions,
- 22:54SPEAKER_00: but a lot of times I'll tell an agency to go away and do an interview
- 23:00SPEAKER_00: with AI itself, you know, if there's anything that's left unclear.
- 23:03SPEAKER_00: Meeting transcribing is something that is very popular and for good reason.
- 23:08SPEAKER_00: I think that everyone should be doing it in most situations as long
- 23:13SPEAKER_00: as your clients are aware that you're doing it and
- 23:15SPEAKER_00: that everything's contractually sound there.
- 23:18SPEAKER_00: But it's so important to get that data.
- 23:21SPEAKER_02: So if a department leader,
- 23:24SPEAKER_02: as we talked about earlier,
- 23:26SPEAKER_02: has got a new person in the business and they're Onboarding them,
- 23:30SPEAKER_01: training them on a particular process.
- 23:32SPEAKER_02: Yeah.
- 23:33SPEAKER_02: Obvious point, I guess.
- 23:34SPEAKER_02: But just to highlight it, we could record that using AI, that meeting.
- 23:42SPEAKER_02: It's a good idea.
- 23:43SPEAKER_02: Exactly, yeah.
- 23:44SPEAKER_02: And you've got a live transcript,
- 23:47SPEAKER_02: a record of what you've actually just briefed your new person or person that's new
- 23:52SPEAKER_02: to doing this job.
- 23:53SPEAKER_02: Yeah.
- 23:55SPEAKER_02: You can use that then, can't you?
- 23:57SPEAKER_02: You probably don't have to repeat it as much.
- 23:59SPEAKER_02: You can use it as documentation,
- 24:02SPEAKER_02: give that person some sort of link to that to say,
- 24:07SPEAKER_02: if you forget or you're not clear on anything I've said, there's documentation.
- 24:12SPEAKER_00: Exactly.
- 24:13SPEAKER_00: Even if it wasn't the primary objective to track down the workflow,
- 24:19SPEAKER_00: if you're teaching someone anyway,
- 24:20SPEAKER_00: if you just have a meeting transcription tool in the background,
- 24:24SPEAKER_00: You are practically doing this process of giving AI that context.
- 24:29SPEAKER_00: And the interviewing prompt,
- 24:32SPEAKER_00: it depends how much time you have really,
- 24:34SPEAKER_00: the more questions the better,
- 24:35SPEAKER_00: but it's good to get the quantitative data,
- 24:38SPEAKER_00: the ones that already exist and are written down first,
- 24:41SPEAKER_00: and then AI can plug in the gaps and you can say something as simple as,
- 24:46SPEAKER_00: What else would you need for me to tell you to know the process?
- 24:50SPEAKER_00: And it could say, you know, give 50 questions.
- 24:53SPEAKER_00: You could sit down.
- 24:53SPEAKER_00: It's not even one person in the department.
- 24:56SPEAKER_00: It could be three people who know the process best.
- 24:59SPEAKER_00: And each of them might have a nuanced take.
- 25:01SPEAKER_00: You know,
- 25:01SPEAKER_00: like we said at the beginning,
- 25:02SPEAKER_00: people have their own flair and their own way of doing things.
- 25:05SPEAKER_00: Person A might cut corners in one step.
- 25:08SPEAKER_00: Person C might add more steps.
- 25:11SPEAKER_00: And it's good to get all of that information down.
- 25:14SPEAKER_02: Get AI to record the three and synthesize,
- 25:19SPEAKER_02: if I can say that,
- 25:22SPEAKER_02: the three into the one standard process that maybe all three people can sign up to,
- 25:27SPEAKER_02: or at least discuss,
- 25:29SPEAKER_02: and that becomes the one That's not to say you have to stick to
- 25:32SPEAKER_02: that process forever.
- 25:32SPEAKER_02: If you decide that process needs improving or even
- 25:49SPEAKER_00: So you want to be doing this process and ironically this is
- 25:57SPEAKER_00: a standard operating procedure in itself,
- 25:59SPEAKER_00: the actual deriving the information and the But you want to be doing it in
- 26:05SPEAKER_00: a project, in any AI tool that you're on, or a chat window.
- 26:09SPEAKER_00: But make sure at the end of it,
- 26:12SPEAKER_00: for most people,
- 26:13SPEAKER_00: unless you've got some knowledge base that you connect externally to AI,
- 26:18SPEAKER_00: the context dies within the chat.
- 26:20SPEAKER_00: So at the end of the process,
- 26:21SPEAKER_00: you want to make sure that you're saying,
- 26:24SPEAKER_00: save this, all of this context as an MD file, a PDF.
- 26:28SPEAKER_00: HTML, any kind of resource where it captures the context and remembers it
- 26:32SPEAKER_00: so you can redistribute.
- 26:35SPEAKER_00: And it's important,
- 26:36SPEAKER_00: this process,
- 26:36SPEAKER_00: because a lot of the times the pain point that you think is
- 26:40SPEAKER_00: the big friction is actually the wrong one.
- 26:45SPEAKER_00: Or even I've come into meetings with agencies thinking
- 26:50SPEAKER_00: that AI can really help them out in a certain area.
- 26:53SPEAKER_00: And I was wrong because we've gone through the process together
- 26:56SPEAKER_00: and we've established that they really need help with one part,
- 26:59SPEAKER_00: but the other part's covered.
- 27:00SPEAKER_00: AI doesn't need to interfere.
- 27:01SPEAKER_00: So it's about saying where...
- 27:04SPEAKER_00: Is the agency's skill and still completely human and we don't want AI to interfere
- 27:11SPEAKER_00: at all and where is it that we're really struggling?
- 27:13SPEAKER_00: So a brief example of this is I work with a media planning agency and And a lot of
- 27:19SPEAKER_00: that is research and it's analysis of research and giving insights to their clients
- 27:25SPEAKER_00: on how they should spend on marketing budgets.
- 27:29SPEAKER_00: And I was thinking, AI is incredible at research.
- 27:33SPEAKER_00: I'm going to be...
- 27:34SPEAKER_00: And this was kind of after the audit when I started to implement the workflows.
- 27:38SPEAKER_00: I was thinking, oh,
- 27:38SPEAKER_00: there's so many different tools and systems where you can get this information
- 27:43SPEAKER_00: and section it nicely and whatnot.
- 27:46SPEAKER_00: But me not having worked with a media planning agency before,
- 27:49SPEAKER_00: didn't realize that research is something that they're supplied with.
- 27:54SPEAKER_00: They pay media suppliers to give them the insights and send them the documents so
- 28:00SPEAKER_00: the research is already done.
- 28:02SPEAKER_00: The AI didn't need to help them there.
- 28:04SPEAKER_00: What was the massive pain was taking the research and then analysing it.
- 28:09SPEAKER_00: So a lot of the way that it was supplied, it was from different suppliers.
- 28:13SPEAKER_00: Perhaps even 20 in some cases,
- 28:15SPEAKER_00: all of these different formats,
- 28:17SPEAKER_00: so some of them were 80 page documents,
- 28:18SPEAKER_00: some of them were 60,000 row spreadsheets,
- 28:22SPEAKER_00: others were kind of nuanced,
- 28:25SPEAKER_00: the media suppliers' written take,
- 28:27SPEAKER_00: Which came with a kind of quantitative data take and sometimes they didn't match up.
- 28:34SPEAKER_00: A lot of inconsistencies here.
- 28:36SPEAKER_00: So it was the analysis and I was thinking, okay, well,
- 28:39SPEAKER_00: we know that the system should completely go into analysis rather than research.
- 28:43SPEAKER_00: So it's important.
- 28:44SPEAKER_00: It helps you establish exactly what the pain point is.
- 28:48SPEAKER_00: As you can see,
- 28:49SPEAKER_00: there's a lot that's done before AI even enters the picture,
- 28:53SPEAKER_00: before you even start to get these gains.
- 28:56SPEAKER_00: There's a lot of planning involved and that's really how you get the best outputs
- 29:00SPEAKER_00: and best systems put together.
- 29:04SPEAKER_02: I've used this a lot in my career as an agency advisor,
- 29:08SPEAKER_02: this little When talking about efficiency,
- 29:17SPEAKER_02: I've often said that clients often,
- 29:20SPEAKER_02: this is again without AI involved,
- 29:24SPEAKER_02: clients often want to travel by taxi,
- 29:26SPEAKER_02: they want to get in the agency and give them a destination and
- 29:34SPEAKER_02: But if on the course of that journey they want to go somewhere else or have a detour
- 29:37SPEAKER_02: or go back and do some amends or go faster or get there and actually change
- 29:45SPEAKER_02: the destination completely, that happens and the taxi meter clocks up.
- 29:51SPEAKER_02: And the taxi meter is the time sheets.
- 29:54SPEAKER_02: And time can scope creep and all that sort of stuff.
- 29:58SPEAKER_02: But to be efficient for both the agency and the client, you really need to run to a timetable.
- 30:03SPEAKER_02: You need to go by train, really.
- 30:06SPEAKER_02: You need to set the destination and the time.
- 30:08SPEAKER_02: I like the analogy, yeah, that's good.
- 30:11SPEAKER_02: So here's one for you.
- 30:12SPEAKER_02: So I was thinking, I've used that just to illustrate that point about efficiency.
- 30:19SPEAKER_02: And I know there are challenges with that.
- 30:20SPEAKER_02: We talked about it earlier.
- 30:21SPEAKER_02: Some people would say, well,
- 30:22SPEAKER_02: you can't go by a timetable because we're a creative business and we need
- 30:25SPEAKER_02: to go explore and all that sort of stuff.
- 30:27SPEAKER_02: But there's a principle there about setting a timetable and using a set route and
- 30:33SPEAKER_02: a set time rather than just meandering about.
- 30:37SPEAKER_02: I think,
- 30:38SPEAKER_02: as you've been talking,
- 30:38SPEAKER_02: I've been thinking,
- 30:39SPEAKER_02: and I've never used this before,
- 30:41SPEAKER_02: so I'm interested in what you think,
- 30:42SPEAKER_02: but if you're using that train analogy,
- 30:44SPEAKER_02: you could say the process piece is setting the tracks.
- 30:49SPEAKER_02: So you've got the tracks in place for the destination, then you put the train on top.
- 30:55SPEAKER_02: And the train used to be a set of agency people, but the train is now...
- 31:00SPEAKER_02: A high-speed AI-powered vehicle.
- 31:05SPEAKER_02: If you don't set the tracks properly, it's just going to fly off.
- 31:08SPEAKER_02: It's going to derail, isn't it?
- 31:10SPEAKER_02: There's going to be an accident because you've not set the The track's in
- 31:15SPEAKER_02: the right order, in the right place, or you're missing a better track.
- 31:18SPEAKER_02: If you put this suddenly AI-powered engine on top of the tracks, it's not going to work.
- 31:26SPEAKER_02: Did you come up with that on the spot?
- 31:27SPEAKER_02: I've just come up with that there.
- 31:29SPEAKER_00: I like that analogy because I'm even thinking that there's bullet trains in Japan,
- 31:34SPEAKER_00: so there's speeds of trains or steam engine trains.
- 31:40SPEAKER_02: So as people, we're in a steam engine.
- 31:42SPEAKER_02: We're good, we get to the destination, but we're quite slow and quite laboured in some cases.
- 31:47SPEAKER_02: But no, the AI is the bullet train.
- 31:50SPEAKER_02: Don't put a bullet train on steam engine tracks.
- 31:52SPEAKER_00: Yeah, or don't put a bullet train on no tracks because it'll just collapse
- 31:56SPEAKER_00: and it won't go anywhere.
- 31:59SPEAKER_00: It needs the ability to actually move.
- 32:02SPEAKER_00: We're talking about trains here, but AI needs the process to speed up.
- 32:08SPEAKER_00: If it doesn't have that scaffolding, it's just going to deviate.
- 32:15SPEAKER_00: Clients of agencies want the flexibility.
- 32:18SPEAKER_00: I've built a system for an agency and it's a timesheet reconciliation
- 32:23SPEAKER_00: and it compares project prices mid-month and month-end because in two weeks
- 32:28SPEAKER_00: the price of a project can change because a client comes back and they want
- 32:33SPEAKER_00: to Maybe remove a certain deliverable and they want to add something else.
- 32:39SPEAKER_00: They want something changed.
- 32:41SPEAKER_00: If you're locked into it, that can change in weeks.
- 32:45SPEAKER_00: It's important to keep track of that.
- 32:47SPEAKER_00: But if you have a standard operating procedure, it could feel a bit rigid.
- 32:50SPEAKER_00: People don't want to...
- 32:53SPEAKER_00: It doesn't give you exactly the ability.
- 32:55SPEAKER_00: If people go, oh, we have to deviate from what's standardized, what do we do?
- 33:00SPEAKER_00: I'd argue that having it written down allows you to deviate.
- 33:04SPEAKER_00: It's having the tracks that we said.
- 33:07SPEAKER_02: Absolutely, I agree.
- 33:08SPEAKER_02: I think that's my philosophy.
- 33:10SPEAKER_02: Again, even regardless of AI in my working career,
- 33:13SPEAKER_02: I've always had that philosophy of you put the There are processes in place,
- 33:18SPEAKER_02: but they shouldn't stifle creativity or flexibility.
- 33:22SPEAKER_02: They should help it because you know what the core work is,
- 33:25SPEAKER_02: how you're going to do the core process,
- 33:28SPEAKER_02: but then you also have the ability to go outside of those lines if you need to,
- 33:32SPEAKER_02: and it gives you a structure rather than What could be chaos,
- 33:37SPEAKER_02: it gives you a structure to follow,
- 33:38SPEAKER_02: but also the scope to be able to say,
- 33:40SPEAKER_02: in this instance,
- 33:41SPEAKER_02: we're going to step outside the process and do something slightly differently
- 33:45SPEAKER_02: because either the clients ask for it or because we've decided that that's the right
- 33:49SPEAKER_02: in this particular individual instance.
- 33:51SPEAKER_00: Yeah, it might be an unpopular opinion,
- 33:53SPEAKER_00: but I believe that creativity is just an amalgamation of what already exists.
- 33:58SPEAKER_00: So it's someone's ability to join together certain principles and ideas
- 34:05SPEAKER_00: and formulate something new.
- 34:06SPEAKER_00: So if you have an SOP,
- 34:09SPEAKER_00: Everything is nice and organised and visible in front of you.
- 34:14SPEAKER_00: And so you can then think about how deviating from what's standardised would improve
- 34:21SPEAKER_00: the process.
- 34:22SPEAKER_00: So it's a massive part of how to make your agency more efficient.
- 34:27SPEAKER_00: If you have processes written down,
- 34:29SPEAKER_00: so you can analyse it and go, how do we make this process more efficient?
- 34:32SPEAKER_00: And then you've got 10 other SOPs and you can...
- 34:36SPEAKER_00: Deal with them in kind of packaged segments.
- 34:41SPEAKER_00: One thing that is very apparent to me with the,
- 34:44SPEAKER_00: we talked about timesheets and we brought that up as is such a big factor and
- 34:50SPEAKER_00: a big area of agencies and unfortunately for many people,
- 34:55SPEAKER_00: but it's the way that, it's the culture.
- 34:59SPEAKER_00: Every time AI speeds up a process of yours,
- 35:04SPEAKER_00: You therefore spend less time on it and your charge out rates
- 35:09SPEAKER_00: then include less hours, Which is the agency business model, isn't it?
- 35:16SPEAKER_02: People listening to this will be aware that the time-based business model
- 35:20SPEAKER_02: under threat through AI,
- 35:21SPEAKER_02: as good as AI is,
- 35:22SPEAKER_02: because time isn't going to be a currency as much anymore,
- 35:26SPEAKER_00: because AI can do it faster.
- 35:28SPEAKER_00: Exactly.
- 35:28SPEAKER_00: Lots of things faster, isn't it?
- 35:30SPEAKER_00: Yeah, so the time model was based on human speed, but now AI speed's here.
- 35:35SPEAKER_00: Bullet train speed.
- 35:36SPEAKER_00: Bullet train speed.
- 35:37SPEAKER_00: Yeah, timesheets were steam engine speed, now it's bullet train speed.
- 35:42SPEAKER_00: SOPs are essential for combating this.
- 35:47SPEAKER_00: So instead of actually earning less for being better,
- 35:51SPEAKER_00: and I say better because I think that quicker,
- 35:54SPEAKER_00: as long as it's the same quality, is better.
- 35:58SPEAKER_00: You have to have something to measure.
- 36:01SPEAKER_00: And having an SOP down,
- 36:03SPEAKER_00: especially an AI SOP,
- 36:05SPEAKER_00: and if it was manual to begin with,
- 36:07SPEAKER_00: then you add AI, you can measure, okay, how much time did it take?
- 36:12SPEAKER_00: How much time does it take now?
- 36:14SPEAKER_00: And you can be really transparent with your clients about certain projects.
- 36:19SPEAKER_02: So benchmarks, how did we create it and do it?
- 36:26SPEAKER_02: And in what time before, now using AI?
- 36:30SPEAKER_02: What does that take now in the future using AI as we've improved our briefs to AI,
- 36:36SPEAKER_02: our prompts, our way of using it, it could improve even further, couldn't it?
- 36:40SPEAKER_00: Yeah, and it's been client asks from me before even starting any work.
- 36:47SPEAKER_00: We really want to be transparent with our clients and procurement, things like this.
- 36:53SPEAKER_00: Just negotiations on project fees.
- 36:56SPEAKER_00: If we're speeding processes up, we don't want to be punished for that.
- 37:00SPEAKER_00: If anything,
- 37:01SPEAKER_00: we'd like to be transparent so that we can give some of the AI gains
- 37:04SPEAKER_00: to our clients out of goodwill,
- 37:06SPEAKER_00: but we keep a lot of them because it's our internal efficiency
- 37:10SPEAKER_00: and it's our innovation that's made the processes quicker.
- 37:14SPEAKER_00: The only way you can measure something and capacity model,
- 37:17SPEAKER_00: if we're talking about pricing, As if you have something to model, really.
- 37:23SPEAKER_00: You have the process to kind
- 37:25SPEAKER_02: of measure.
- 37:26SPEAKER_02: You've got the foundation, you've got something to judge it against.
- 37:29SPEAKER_02: So as we've said, it's not just in people's heads, it's written down.
- 37:34SPEAKER_01: Great.
- 37:35SPEAKER_02: I suppose we should bring this episode to a close, as much as I've enjoyed it.
- 37:39SPEAKER_02: We always end with a...
- 37:41SPEAKER_02: So, Callum, if an agency owner is...
- 37:47SPEAKER_02: Yeah, well, I think...
- 38:01SPEAKER_00: Number one for everyone is getting on a team plan or an enterprise plan.
- 38:05SPEAKER_00: It's the AI model.
- 38:06SPEAKER_00: As much as I don't like to prioritize the tools,
- 38:11SPEAKER_00: everyone wants to be in the same environment so that they can use these SOPs and
- 38:17SPEAKER_00: the ones that you embed into AI, everyone has access to.
- 38:21SPEAKER_00: So that's important to start out.
- 38:24SPEAKER_00: Hopefully,
- 38:25SPEAKER_00: if you're listening and you've gotten this far,
- 38:28SPEAKER_00: We might be flattering you, I guess, because you might have a lot of SOPs already.
- 38:32SPEAKER_00: So the ones that you do have in this business AI plan,
- 38:39SPEAKER_00: use the method that we talked about to embed them.
- 38:41SPEAKER_00: So if they're written down, perfect.
- 38:42SPEAKER_00: You've got loads of quantitative knowledge.
- 38:45SPEAKER_00: And then I think no matter how detailed it is,
- 38:48SPEAKER_00: do the step that you're saying to AI, what else would you need to know?
- 38:52SPEAKER_00: What are you still unclear?
- 38:54SPEAKER_00: So you have the full picture.
- 38:56SPEAKER_00: Do that interview process.
- 38:58SPEAKER_00: Either you can dictate your thoughts in an AI chat with something like Whisperflow
- 39:03SPEAKER_00: so you don't have to type it all out.
- 39:05SPEAKER_00: Or you could have 50 questions in front of you that you have a meeting with
- 39:09SPEAKER_00: the team members who know the process best and answer them together,
- 39:13SPEAKER_00: record your insights, feed that to AI as well.
- 39:16SPEAKER_00: So it's got quantitative, qualitative information.
- 39:19SPEAKER_00: Then make either,
- 39:21SPEAKER_00: you know,
- 39:21SPEAKER_00: there's called projects or called skills,
- 39:24SPEAKER_00: custom GPTs, Gemini Gems, very simple ways of packaging these SOPs.
- 39:30SPEAKER_00: Ask AI to help you make one once it's got the context.
- 39:35SPEAKER_00: And then for the ones that you don't really have written down,
- 39:39SPEAKER_00: if you've not got any SOPs, start with the pain first.
- 39:43SPEAKER_00: Think about which department is slow, where efficiency is needed.
- 39:47SPEAKER_00: And before you establish the SOP and get it into AI,
- 39:52SPEAKER_00: you need to establish where AI can help and map out each step of the process.
- 39:57SPEAKER_02: Let me think about what of either you or one of your team is constantly...
- 40:02SPEAKER_02: Regurgitating,
- 40:03SPEAKER_02: constantly telling people how to do something,
- 40:06SPEAKER_02: either the same person or different people, you know, that sort of onboarding piece.
- 40:13SPEAKER_02: Somebody needs to join the business,
- 40:14SPEAKER_02: we've got to sit down with them and tell them again how to...
- 40:17SPEAKER_00: I love that question, yeah, that's really good.
- 40:19SPEAKER_00: I use that a lot as well because I build And that is the core question
- 40:24SPEAKER_00: when I'm doing Blueprint sessions for those.
- 40:38SPEAKER_00: If you're in the agency today, what would you tell them?
- 40:41SPEAKER_00: How would you teach them who you are,
- 40:42SPEAKER_00: what you do,
- 40:43SPEAKER_00: your values, missions, your voice, your brand design, all of this stuff?
- 40:47SPEAKER_00: But in the context of SOPs, if someone was entering a new department, what would you tell them?
- 40:53SPEAKER_00: How would you teach them how work gets done?
- 40:55SPEAKER_00: And do that with AI.
- 40:56SPEAKER_00: So you train it and it needs more help.
- 40:59SPEAKER_00: It's very clever and, well, it seems to be, anyway, it appears to be very clever.
- 41:05SPEAKER_00: But it needs your help to get your processes.
- 41:07SPEAKER_00: It needs the tracks for the bullet train.
- 41:11SPEAKER_02: Great episode.
- 41:14SPEAKER_02: See you the next time in episode 11.
- 41:17SPEAKER_02: Are we going to do episode 11 or maybe episode 10 and a half to be halfway?
- 41:21SPEAKER_00: Yeah, true actually.
- 41:22SPEAKER_00: To 21.
- 41:22SPEAKER_00: We'll see.
- 41:24SPEAKER_00: We'll figure that out.
- 41:25SPEAKER_00: Yeah, definitely.
- 41:26SPEAKER_00: I'm looking forward to it.
- 41:29SPEAKER_00: You've been listening to Gareth Healey and Callum Healey.
- 41:31SPEAKER_00: Together, we are Agents of Change, which is also our URL, weareagentsofchange.com.
- 41:37SPEAKER_01: If you want support with growing your agency,
- 41:39SPEAKER_01: including how AI can help you future-proof your business,
- 41:43SPEAKER_01: visit us online to learn more about what we do and find out how to connect with us.