Latest / The STANDOUT Agency Podcast / EPISODE 12: The Robots Are Coming - 3 Key Takeaways
Transcript
- 0:04SPEAKER_02: Welcome to The STANDOUT Agency Podcast, for agency owners who want to grow, not just survive.
- 0:09SPEAKER_00: I'm Gareth Healey.
- 0:10SPEAKER_00: I'm an ex-agency owner and author of STANDOUT OR DIE.
- 0:14SPEAKER_00: I help agency founders see where they're still a bottleneck in their own business
- 0:17SPEAKER_00: and make sure they do something about it.
- 0:19SPEAKER_02: And I'm Callum Healey.
- 0:20SPEAKER_02: I find where agencies are losing hours to manual work and then build AI systems
- 0:24SPEAKER_02: that get the time back.
- 0:27SPEAKER_00: Each week, one of us interviews the other on a subject related to agency growth.
- 0:31SPEAKER_00: Everything we get into comes straight out of our own experience and client work.
- 0:34SPEAKER_00: Occasionally we bring in some guests to add their expert perspective on a topic.
- 0:40SPEAKER_02: Welcome back to The STANDOUT Agency Podcast.
- 0:42SPEAKER_02: This is episode 12.
- 0:44SPEAKER_02: We went to an event around a week and a half
- 0:47SPEAKER_01: ago, two weeks now
- 0:48SPEAKER_02: maybe.
- 0:49SPEAKER_02: It was, yeah, on the 8th of October.
- 0:50SPEAKER_02: Yeah, which is The Robots Are Coming,
- 0:53SPEAKER_02: hosted by Agency Hackers,
- 0:55SPEAKER_02: and we thought that we'd take what we learn on the day and talk about some of
- 0:59SPEAKER_02: the key talking points.
- 1:01SPEAKER_02: We've deduced it to three, so this episode is the three takeaways, and...
- 1:06SPEAKER_02: The main ones that we discovered were AI's impact on the margin model and that
- 1:12SPEAKER_02: by default clients can actually get more of the reward.
- 1:15SPEAKER_02: There is the ownership problem about AI that a lot of times Processes
- 1:21SPEAKER_02: and initiatives fail when people aren't claiming ownership and taking
- 1:25SPEAKER_02: the initiative themselves.
- 1:27SPEAKER_02: And finally,
- 1:28SPEAKER_02: the idea of data assets and a few of the data assets that are available to people
- 1:34SPEAKER_02: that they aren't quite using to their full potential.
- 1:39SPEAKER_01: Totally, yeah, it's a good event, wasn't it?
- 1:41SPEAKER_01: It's the first Agency Hackers event I've been to, believe it or not.
- 1:44SPEAKER_01: Yeah, we hope to go again, definitely.
- 1:46SPEAKER_01: Definitely, yeah,
- 1:46SPEAKER_01: credit to Ian Harris and his team there for putting on yet another great event
- 1:52SPEAKER_01: so quickly after they've been to Ibiza and all over the world as well.
- 1:55SPEAKER_01: So, yeah, it's great, definitely go again.
- 1:59SPEAKER_01: So, shall we take the first one first, the three?
- 2:01SPEAKER_01: Let's do it, yeah.
- 2:03SPEAKER_02: Before we begin, it was a passionate community.
- 2:06SPEAKER_02: It was one that I was pleasantly surprised by the amount
- 2:10SPEAKER_02: of enthusiasm people were showing for AI implementation.
- 2:14SPEAKER_02: We're now discussing the main three kind of talking points
- 2:17SPEAKER_02: that might be holding people back or...
- 2:20SPEAKER_01: It could indeed accelerate their transformation.
- 2:23SPEAKER_01: I don't know how many people were in that room, I guess there was tables of 10, maybe 150, 200.
- 2:28SPEAKER_01: It was difficult to know what the overall vibe
- 2:35SPEAKER_01: Naturally, the people who are
- 2:58SPEAKER_02: Paying and traveling to an AI event are going to be engaged and positive
- 3:05SPEAKER_02: about AI implementation.
- 3:06SPEAKER_02: I do sense a quiet desperation that although there's positivity,
- 3:11SPEAKER_02: it's because people are realizing how powerful AI is and so they want
- 3:15SPEAKER_02: to make sure they stay on top of everything and keep their business healthy and look
- 3:19SPEAKER_02: at how they might be able to grow the business.
- 3:23SPEAKER_02: Margins,
- 3:23SPEAKER_02: of course,
- 3:24SPEAKER_02: is the first topic and that's a big one because ultimately AI,
- 3:28SPEAKER_02: I'll say it over and over again,
- 3:29SPEAKER_02: like a broken record, it's speed and why it's so valuable.
- 3:34SPEAKER_02: Speed is only useful if it then translates to profitability.
- 3:39SPEAKER_02: The agency model is sort of grounded,
- 3:42SPEAKER_02: you know a lot more about this than me,
- 3:43SPEAKER_02: but it's timesheets and it's hourly rates and the guys at teamwork.com were also at
- 3:49SPEAKER_02: the event and were realising that margins is a topic,
- 3:54SPEAKER_02: a big sort of critical area for discussion as AI evolves.
- 3:59SPEAKER_01: So the first point was AI is changing and will certainly change the agency business model.
- 4:05SPEAKER_01: And as you mentioned, I guess this was our top takeaway really, wasn't it?
- 4:11SPEAKER_01: From the event of many.
- 4:14SPEAKER_01: Off the back of a presentation,
- 4:16SPEAKER_01: as you mentioned, Callan,
- 4:17SPEAKER_01: Teamwork.com,
- 4:18SPEAKER_01: Mossy O'Mahoney from Teamwork did a great presentation in general, actually.
- 4:23SPEAKER_01: Very engaging.
- 4:24SPEAKER_01: Very good, lots of stats in there.
- 4:26SPEAKER_01: But one of the things that stood out for me was he said in his opinion,
- 4:29SPEAKER_01: and it was just an opinion,
- 4:30SPEAKER_01: it wasn't based on stats,
- 4:31SPEAKER_01: that people are now in agencies resourcing,
- 4:38SPEAKER_01: planning resource, putting agents, AI agents in alongside people.
- 4:44SPEAKER_01: True, yeah, that was fascinating to me as well.
- 4:46SPEAKER_01: And that made sense.
- 4:47SPEAKER_01: I've seen people start to do that, but he's seen a lot of that.
- 4:50SPEAKER_01: Yeah.
- 4:51SPEAKER_01: And so actually when you're thinking,
- 4:53SPEAKER_01: rather than the traditional agency business model,
- 4:55SPEAKER_01: and I've had these conversations with agencies this week,
- 4:59SPEAKER_01: as we grow, well,
- 5:00SPEAKER_01: we need to add this person and that person and that role and this role,
- 5:03SPEAKER_01: and it's all about people in the future and in the present.
- 5:06SPEAKER_01: It is now people starting to think,
- 5:07SPEAKER_01: hang on a minute, do I need a person or do I need, can I utilise an AI agent?
- 5:13SPEAKER_02: Yeah,
- 5:13SPEAKER_02: it's interesting and I like the idea of that and people are of course experimenting,
- 5:19SPEAKER_02: it's the phase where AI is very new so it's a challenge to determine the best way
- 5:25SPEAKER_02: of tracking their utility ultimately.
- 5:30SPEAKER_02: Because human intelligence and AI intelligence is very different,
- 5:36SPEAKER_02: I do like the idea of them being segregated in a way because We were saying in
- 5:42SPEAKER_02: a car ride home yesterday,
- 5:44SPEAKER_02: that a big change in the industry is Software used to be hard to change
- 5:51SPEAKER_02: and you'd get used to the software and it wouldn't be constantly upgraded
- 5:56SPEAKER_02: but AI is actually subject to getting sort of new versions,
- 6:01SPEAKER_02: new models,
- 6:02SPEAKER_02: you know, V2,
- 6:03SPEAKER_02: V3 and getting quicker and it's actually important to evaluate agents in this case
- 6:10SPEAKER_02: and see how they improve over time.
- 6:13SPEAKER_02: I'd argue it's easier to track An agent's performance against the previous versions
- 6:18SPEAKER_02: of the agent's performance, rather than comparing it to a person.
- 6:24SPEAKER_02: I wonder what you thought about that divide.
- 6:27SPEAKER_01: Oh, definitely.
- 6:27SPEAKER_01: Yeah, I suppose you're right.
- 6:29SPEAKER_01: It is potentially easier to sort of track that because it's got the data.
- 6:32SPEAKER_01: The agent's always going to do its timesheets, isn't it?
- 6:34SPEAKER_01: If you ask it to.
- 6:35SPEAKER_01: Yeah, that's a good advantage.
- 6:37SPEAKER_01: That is one advantage.
- 6:39SPEAKER_01: One of the things that Mossy,
- 6:41SPEAKER_01: when he was doing his presentation,
- 6:43SPEAKER_01: he talked about that sort of resource planning piece,
- 6:45SPEAKER_01: but he also said, again,
- 6:47SPEAKER_01: it was his opinion that as things stand,
- 6:50SPEAKER_01: he reckoned that 30% of junior work within agencies In the future is probably going
- 6:58SPEAKER_01: to be done by AI.
- 7:02SPEAKER_01: At the time in the presentation on the day,
- 7:06SPEAKER_01: I thought in my head that was probably a little bit low actually.
- 7:10SPEAKER_01: And I've thrown that stat out to a couple of my clients just in meetings in passing,
- 7:15SPEAKER_01: saying that that was part of that presentation.
- 7:17SPEAKER_01: And generally people have sort of said, yeah, that feels as it could be even more than that.
- 7:24SPEAKER_01: It could than 30%.
- 7:25SPEAKER_01: It could be half the junior work or more that is done by AI going forward.
- 7:31SPEAKER_02: Yeah, and that's almost a topic for a full episode,
- 7:35SPEAKER_02: the kind of cultural shift in what happens to the future generation of workers in
- 7:42SPEAKER_02: the agency space when they're not able to develop the reps that their senior sort
- 7:49SPEAKER_02: of counterparts have done because in the past they were in the engine room
- 7:52SPEAKER_02: and picking up the kind of the skills that You can only learn through going through
- 7:59SPEAKER_02: that brunt work in the engine room.
- 8:03SPEAKER_01: So the future for the generation of human workers is another episode,
- 8:08SPEAKER_01: another debate all in itself.
- 8:09SPEAKER_01: But coming back to your point of speed,
- 8:12SPEAKER_01: Do you remember there was a chap that was from the agency that was Inspire KBB,
- 8:19SPEAKER_01: his niche was kitchens, bathrooms and bedrooms, which I love a niche agency.
- 8:26SPEAKER_01: It's a great niche.
- 8:30SPEAKER_01: So he did a great presentation,
- 8:32SPEAKER_01: that chap as well,
- 8:33SPEAKER_01: and I remember him,
- 8:35SPEAKER_01: the point on this change,
- 8:37SPEAKER_01: he's using AI,
- 8:39SPEAKER_01: and I remember him saying,
- 8:39SPEAKER_01: what used to take an hour per client,
- 8:42SPEAKER_01: I think it was some sort of report, now it takes about 15 minutes.
- 8:47SPEAKER_01: The AI must take three minutes to produce it,
- 8:56SPEAKER_01: but it takes 12 minutes or whatever just to check it by a human and send it through.
- 9:02SPEAKER_01: So the takeaway number one is the model of resourcing agencies will change and
- 9:14SPEAKER_01: the point on margin is where is that margin going to go?
- 9:17SPEAKER_01: Another point that Mosse made I thought was a great one was to say we've got to make
- 9:21SPEAKER_01: a conscious decision as to if you're saving time
- 9:25SPEAKER_01: and therefore you're Youare saving cost ultimately and youare speeding
- 9:31SPEAKER_01: the process up,
- 9:32SPEAKER_01: therefore your margin as an opportunity,
- 9:35SPEAKER_01: the profit margin,
- 9:36SPEAKER_01: you need to decide are you going to take those improvements and take it as
- 9:43SPEAKER_01: a margin improvement for yourself as an agency?
- 9:47SPEAKER_01: Or are you going to give that back to your client?
- 9:49SPEAKER_02: There's more optionality now,
- 9:50SPEAKER_02: which is intriguing because in five years time we might see that there was
- 9:55SPEAKER_02: a clear winner of what is this pricing strategy that you go down,
- 10:01SPEAKER_02: is it going to be the case that I think that there was a point,
- 10:07SPEAKER_02: an idea from the day that you could just bank the margin and not change anything
- 10:13SPEAKER_02: and just see it as a good thing that processes are becoming more efficient
- 10:17SPEAKER_02: or you could realise that the process takes less time and therefore you can win work
- 10:23SPEAKER_02: on price because a client is more likely to engage with you if the
- 10:30SPEAKER_01: So you could
- 10:30SPEAKER_01: then fulfil more of their requests.
- 10:31SPEAKER_01: Scope creep maybe used to be a terrible
- 10:44SPEAKER_02: It was a terrible thing,
- 10:45SPEAKER_02: but now with AI, you can at least, you can both agree on- Controlled scope creep.
- 10:50SPEAKER_02: Controlled scope creep,
- 10:51SPEAKER_02: where you're saying,
- 10:51SPEAKER_02: we now can do this process a lot quicker, so we're willing to add these benefits on.
- 10:56SPEAKER_02: You can discuss extra value,
- 10:58SPEAKER_02: so it's like,
- 10:59SPEAKER_02: do you keep If everything's the same,
- 11:01SPEAKER_02: I think that would be a good client development strategy to go,
- 11:05SPEAKER_02: we'll be transparent on our AI usage,
- 11:07SPEAKER_02: it's saving us a lot of time,
- 11:08SPEAKER_02: we're willing to then make our services to you better for the same price,
- 11:14SPEAKER_02: and everyone's happy.
- 11:19SPEAKER_01: Raised that though,
- 11:20SPEAKER_01: or the fact that he did raise it, because it is an issue and it is a choice.
- 11:25SPEAKER_01: My fear personally is that this is backed up in some early stage research
- 11:33SPEAKER_01: that has been done My fear is that using AI to speed up the agency's process,
- 11:36SPEAKER_01: to be able to do more things in less time,
- 11:51SPEAKER_01: My fear is that agencies will just naturally give that away to clients.
- 11:57SPEAKER_01: Agencies in massive generalization here,
- 12:00SPEAKER_01: warning,
- 12:01SPEAKER_01: so not necessarily true for everybody,
- 12:03SPEAKER_01: I apologize,
- 12:03SPEAKER_01: but I think agencies,
- 12:05SPEAKER_01: culturally in agencies,
- 12:07SPEAKER_01: in what I've seen over my many years,
- 12:10SPEAKER_01: We too tend to be people pleasers in agencies,
- 12:14SPEAKER_01: we tend to give away,
- 12:15SPEAKER_01: if we've done something or achieved something or we can do something faster,
- 12:18SPEAKER_01: we tend to try and shout to the client and make a point to the client
- 12:23SPEAKER_01: that we've done it,
- 12:24SPEAKER_01: but give it away rather than say, Rather than take that margin for ourselves.
- 12:29SPEAKER_01: So I thought it was a great point by Mossy to raise that.
- 12:32SPEAKER_02: Yeah, it's interesting that you say that and I do agree.
- 12:36SPEAKER_02: It's almost as if to make AI implementation gain your company more margin
- 12:44SPEAKER_02: and therefore increase profitability,
- 12:47SPEAKER_02: you're forced to engage in uncomfortable conversations as a prerequisite
- 12:52SPEAKER_02: to claiming margin.
- 12:54SPEAKER_02: It's not the case where you can just embed loads of agents,
- 13:00SPEAKER_02: focus on the processes,
- 13:01SPEAKER_02: internal operations,
- 13:02SPEAKER_02: all this stuff, and then happy days, you get the margin straight away.
- 13:06SPEAKER_02: You then have to confront that awkward conversation with a client to say,
- 13:11SPEAKER_02: we've made processes more efficient, but let's negotiate.
- 13:15SPEAKER_02: It's almost like another form of sales, isn't it?
- 13:19SPEAKER_02: You've worked with a client for years and years,
- 13:21SPEAKER_02: but then you have to now negotiate your relationship together.
- 13:25SPEAKER_02: Or renegotiate.
- 13:26SPEAKER_02: Or
- 13:29SPEAKER_01: just hide it.
- 13:30SPEAKER_01: Or hide it.
- 13:31SPEAKER_01: Don't have the uncomfortable conversation.
- 13:32SPEAKER_02: And that's a choice in itself, because if you do hide it and then your client finds out,
- 13:37SPEAKER_01: that could harm the relationship.
- 13:39SPEAKER_01: So a lot of different choices, quite frankly, we don't have the answer.
- 13:44SPEAKER_01: Although I've just raised it as a sort of,
- 13:45SPEAKER_01: it's like the tongue-in-cheek,
- 13:46SPEAKER_01: I think hiding it It's not my preference,
- 13:50SPEAKER_01: it's not my go-to solution,
- 13:51SPEAKER_01: because as you've just rightly pointed out there,
- 13:55SPEAKER_01: if and when a client either discovers this or you've been doing things a lot quicker
- 14:00SPEAKER_01: or easier or using AI or whatever,
- 14:02SPEAKER_01: if that's then uncovered at some point,
- 14:03SPEAKER_01: that could damage the relationship, wouldn't it?
- 14:08SPEAKER_01: If not damage the relationship in terms of commercially they believe that
- 14:17SPEAKER_01: the fees they've been paying are too high or they might believe that,
- 14:19SPEAKER_01: but it would potentially damage the relationship in the future, wouldn't it?
- 14:23SPEAKER_01: They would maybe reconsider,
- 14:25SPEAKER_01: it would damage the personal relationship they have with them, the trust.
- 14:30SPEAKER_02: So there are many different options and it's going to be intriguing to see
- 14:33SPEAKER_02: which one is the most effective.
- 14:36SPEAKER_02: I will say that there's a clear answer that the best way you can prepare yourself is
- 14:40SPEAKER_02: if this change in how AI affects the speed The visibility of everything
- 14:51SPEAKER_02: to measure AI's impact gives you the best chance in these potential negotiations.
- 15:00SPEAKER_02: Hiding it is a bad thing, but it'd be even worse if you just simply didn't know.
- 15:05SPEAKER_02: So instead of being almost sneaky, you're being ignorant.
- 15:12SPEAKER_02: You don't want to be in a position where you're just oblivious to its impact.
- 15:17SPEAKER_01: So three actions from takeaway one then for agency owners.
- 15:23SPEAKER_01: Measure.
- 15:23SPEAKER_02: Yeah, so measure,
- 15:24SPEAKER_02: and I think that AI can be your best friend when you have all the data,
- 15:28SPEAKER_02: like your timesheet data, which we'll get onto at the end of this episode.
- 15:34SPEAKER_02: You can use your sort of analytical frameworks to give yourself the best possible chance.
- 15:39SPEAKER_02: And I think a second one could be the fact of maybe measure AI and humans
- 15:44SPEAKER_01: separately for now.
- 15:45SPEAKER_01: Or that could be an option.
- 15:46SPEAKER_01: I do like that.
- 15:47SPEAKER_01: Yeah, I'll take that.
- 15:49SPEAKER_01: So I'll take that as 1A and 1B.
- 15:52SPEAKER_01: Number two action for this, I would say, is...
- 15:57SPEAKER_01: Think about it as simple as it might be,
- 15:59SPEAKER_01: think about resource planning doesn't necessarily need to mean people these days,
- 16:02SPEAKER_01: it could mean agents, it could mean AI.
- 16:05SPEAKER_01: And the third one is make a conscious decision or at least have a plan for
- 16:13SPEAKER_01: if you are improving efficiency and therefore saving time and money,
- 16:17SPEAKER_01: what are you going to do with that?
- 16:18SPEAKER_01: Are you going to use that to Improve your own business margins.
- 16:23SPEAKER_01: Are you going to present that to a client or discuss that with a client and Help
- 16:29SPEAKER_01: and sort of pass those savings as it were onto them or maybe some sort of compromise
- 16:34SPEAKER_01: between the two but having a strategy is better than on that or at least having
- 16:39SPEAKER_01: a view for now is better than wandering into it.
- 16:43SPEAKER_02: So confront the uncomfortable reality because not acting is a decision in itself and
- 16:50SPEAKER_02: that would be the worst thing because then you wouldn't be able to do
- 16:53SPEAKER_02: the number one takeaway of measurement.
- 16:57SPEAKER_02: Interesting times,
- 16:58SPEAKER_02: but hopefully if the agencies all stick together and some
- 17:02SPEAKER_02: of these events we'll be hearing in a couple of months,
- 17:05SPEAKER_02: maybe there'll be a little bit more insight from the guys at Teamwork
- 17:08SPEAKER_02: or any other timesheet business and we'll be making sure we stay tuned.
- 17:14SPEAKER_01: So the second takeaway that we thought was really important and insightful was
- 17:20SPEAKER_01: the point that a number of people made,
- 17:21SPEAKER_01: but was the point sort of saying that We are in the process
- 17:25SPEAKER_01: of AI transformation here,
- 17:27SPEAKER_00: whether we like it or not.
- 17:29SPEAKER_00: And we're already,
- 17:32SPEAKER_00: the room,
- 17:33SPEAKER_00: the presenters in that conference
- 17:35SPEAKER_01: Weare already seeing
- 17:37SPEAKER_01: that transformation stuttering,
- 17:39SPEAKER_01: if not failing, because of lack of ownership and lack of governance.
- 17:44SPEAKER_02: It was Prodigies,
- 17:46SPEAKER_02: the two guys at Prodigies who did a great presentation and brought some stats to
- 17:51SPEAKER_02: the forefront.
- 17:52SPEAKER_01: David and Bashir.
- 17:56SPEAKER_02: Yeah, two thirds, was it,
- 17:58SPEAKER_02: of sort of agencies or maybe three quarters of agencies that, a large proportion.
- 18:02SPEAKER_02: Yeah, three quarters.
- 18:03SPEAKER_02: Three quarters.
- 18:04SPEAKER_01: Three quarters of agencies in their research,
- 18:06SPEAKER_01: which I think is available on their website if anybody wants to look at it,
- 18:11SPEAKER_01: just to give them a plug because they did a great presentation.
- 18:14SPEAKER_01: Three quarters of agencies that they had surveyed cite a lack of resource being
- 18:20SPEAKER_01: the main problem,
- 18:20SPEAKER_01: unsurprisingly,
- 18:21SPEAKER_01: actually,
- 18:22SPEAKER_01: but being one of the main barriers to Transforming any business,
- 18:28SPEAKER_01: I guess, ultimately, but in our case, agencies.
- 18:31SPEAKER_01: So three quarters have started a lack of resource because they haven't got people,
- 18:36SPEAKER_01: they're small businesses,
- 18:37SPEAKER_01: small-ish,
- 18:38SPEAKER_01: small to medium-sized enterprises,
- 18:40SPEAKER_01: agencies that we deal with, independent businesses.
- 18:42SPEAKER_01: I haven't got the luxury of having somebody sat around doing nothing to be able
- 18:46SPEAKER_01: to take on the job of transforming the whole agency and setting the AI strategy
- 18:53SPEAKER_01: and deploying it, have they?
- 18:54SPEAKER_01: Of course they haven't got that person, so they give it to somebody else.
- 18:59SPEAKER_01: They give it as part of a job to somebody else.
- 19:02SPEAKER_01: Someone like me, actually.
- 19:03SPEAKER_01: It's a part of the work that I do.
- 19:05SPEAKER_01: I think it's good that...
- 19:06SPEAKER_01: They could bring in somebody externally, couldn't they?
- 19:08SPEAKER_01: Yeah, like a fractional... That could be a solution, a fraction, which is work that you do.
- 19:11SPEAKER_02: But I think ultimately it's a case that AI,
- 19:14SPEAKER_02: as a new piece of technology, it can be applied in every area of the business.
- 19:23SPEAKER_02: It's not like it's a small-scale thing, like a small project, and so it requires...
- 19:28SPEAKER_02: A lot of strategy and like we were saying at the start,
- 19:32SPEAKER_02: it's ownership, it's deadlines and accountability.
- 19:37SPEAKER_02: Very similar to one of our previous episodes when we were talking about sales and
- 19:41SPEAKER_02: the fact that a permission structure a lot of the time is lacking
- 19:45SPEAKER_02: and therefore people will get consumed with delivery and trying to meet deadlines
- 19:49SPEAKER_02: of client projects.
- 19:51SPEAKER_02: But we're saying a similar solution here,
- 19:54SPEAKER_02: it's ultimately it's We need to have that accountability and have
- 20:00SPEAKER_01: named owners
- 20:01SPEAKER_01: to certain projects.
- 20:02SPEAKER_01: And to carve out, those guys were saying, you've got to carve out some time.
- 20:06SPEAKER_01: If you can't give this AI transformation to one person,
- 20:10SPEAKER_01: probably you can't give it to one person as a job,
- 20:13SPEAKER_01: then if somebody needs to do it as part of their job,
- 20:17SPEAKER_01: or as often happens, there's an AI Team, AI crew, AI group, or whatever it might be.
- 20:26SPEAKER_01: Champions.
- 20:26SPEAKER_01: Champions.
- 20:28SPEAKER_01: A number of people.
- 20:30SPEAKER_01: But the guys were saying that they're seeing that that can tend to fizzle out
- 20:34SPEAKER_01: because The owner of the business doesn't allow them or doesn't facilitate them
- 20:40SPEAKER_01: to carve out enough time in their week or month to spend quality time on this.
- 20:47SPEAKER_01: It's done as part of the day job and that is already showing,
- 20:52SPEAKER_01: it's understandable,
- 20:53SPEAKER_01: but it's already showing Poor results because people are,
- 20:58SPEAKER_01: after a bit of momentum at the start,
- 21:00SPEAKER_01: struggling to maintain that momentum and going back,
- 21:02SPEAKER_01: of course,
- 21:03SPEAKER_01: to doing their client work or their production work or whatever it might be.
- 21:08SPEAKER_02: By definition,
- 21:09SPEAKER_02: AI agents and AI automations are going to behave the same way each time
- 21:15SPEAKER_02: when they're applied.
- 21:16SPEAKER_02: Of course, you can build a system, embed it, train people on it, and then And leave it.
- 21:22SPEAKER_02: And it will do the same thing each time.
- 21:24SPEAKER_02: But going back to what we said earlier,
- 21:26SPEAKER_02: there's still a need to upgrade these systems over time as well.
- 21:29SPEAKER_02: So I think it was the guys at ETM and Mossy as well and Teamwork were saying
- 21:33SPEAKER_02: that there is a need for the approval process, of course.
- 21:37SPEAKER_02: German verification is an important step in any automation or any agentic workflow
- 21:43SPEAKER_02: where exactly are we using the judgment to interject and to verify and
- 21:48SPEAKER_02: to make sure things are okay.
- 21:49SPEAKER_02: But I think that goes back to the ownership part.
- 21:53SPEAKER_02: There's actually a need once a project is embedded to schedule time to
- 21:57SPEAKER_02: then review
- 21:59SPEAKER_02: it and improve
- 22:00SPEAKER_01: it, which is another hurdle.
- 22:02SPEAKER_01: That's really important.
- 22:02SPEAKER_01: You've just reminded me of that.
- 22:05SPEAKER_01: In another world back in the day,
- 22:08SPEAKER_01: not too long ago,
- 22:10SPEAKER_01: deploying a software system was a tricky job,
- 22:14SPEAKER_01: was a difficult job,
- 22:15SPEAKER_01: was time intensive for people doing it in their new business and had its pitfalls.
- 22:20SPEAKER_01: It was one system and deployed,
- 22:23SPEAKER_01: and then once everybody was using it,
- 22:24SPEAKER_01: hopefully, couldn't leave it necessarily all on its own.
- 22:28SPEAKER_01: It needed reviewing, but it was running.
- 22:30SPEAKER_01: It was okay.
- 22:31SPEAKER_01: Whereas AI,
- 22:31SPEAKER_01: as we were talking about yesterday,
- 22:33SPEAKER_01: is a wonderful You can deploy it,
- 22:37SPEAKER_01: but it's changing that fast that within weeks
- 22:39SPEAKER_01: or months what you have deployed is not necessarily broken,
- 22:42SPEAKER_01: but it needs updating, doesn't it?
- 22:45SPEAKER_01: It needs improving and it could potentially do far more for you.
- 22:48SPEAKER_01: So that AI champion or champions in the business, that's another challenge of it, isn't it?
- 22:56SPEAKER_01: It's not just one person,
- 22:59SPEAKER_01: a set number, right,
- 23:00SPEAKER_01: in the next three weeks we're going to deploy this and then we're going
- 23:02SPEAKER_01: to walk away.
- 23:02SPEAKER_01: It's an ongoing job.
- 23:04SPEAKER_02: Yeah, and part of the ownership framework,
- 23:08SPEAKER_02: the permission structure,
- 23:10SPEAKER_02: it can go the other way to tell people this system is now finished and it's built
- 23:16SPEAKER_02: and it's very valuable.
- 23:18SPEAKER_02: There is a tendency with AI, because it's easier to change, to be a perfectionist.
- 23:25SPEAKER_02: So there's also the structure around when is the right time
- 23:31SPEAKER_01: to move on and move on
- 23:32SPEAKER_01: to the next project.
- 23:34SPEAKER_01: Sometimes it never gets deployed, does it?
- 23:36SPEAKER_01: Which is the bigger problem.
- 23:38SPEAKER_02: Yeah, and so there's an ever-increasing need for sort of project management and giving,
- 23:44SPEAKER_02: allowing enough time for these systems to be built.
- 23:46SPEAKER_02: And another part of this core topic that we're discussing here was the governance piece.
- 23:52SPEAKER_02: So a lot of people were rightly sort of concerned about AI's sort of invasive nature
- 24:00SPEAKER_02: and to build anything that you embed across a whole business You have to make sure
- 24:06SPEAKER_02: that it's reliable and safe and that people feel confident enough to use a system
- 24:10SPEAKER_02: that they aren't going to mess anything up accidentally.
- 24:14SPEAKER_02: So that's a big part.
- 24:15SPEAKER_02: I mean, we've got an episode on governance,
- 24:17SPEAKER_02: I think episode four maybe,
- 24:18SPEAKER_02: one of our earlier episodes,
- 24:20SPEAKER_02: but it's important in this topic because that's the start of kind of creating
- 24:26SPEAKER_02: a new system and all of this project management and ownership,
- 24:29SPEAKER_02: it's okay, who's owning the
- 24:38SPEAKER_02: The security of course,
- 24:43SPEAKER_02: but
- 24:45SPEAKER_02: who takes responsibility for making sure that a system is governed before it's
- 24:51SPEAKER_02: then created and
- 24:53SPEAKER_01: actually over time.
- 24:54SPEAKER_01: Some people I'm sure are experiencing this,
- 24:56SPEAKER_01: but the danger is you've got these AI champions in the business doing
- 25:00SPEAKER_01: the best part-time as it were in deploying AI.
- 25:03SPEAKER_01: And people start to assume that they are the governance people,
- 25:08SPEAKER_01: and to a degree they are,
- 25:09SPEAKER_01: they take that on that mantle,
- 25:12SPEAKER_01: but they're not spending a great deal of time on it,
- 25:14SPEAKER_01: they're doing many other things as well,
- 25:15SPEAKER_01: and then something potentially might go wrong,
- 25:18SPEAKER_01: God forbid,
- 25:19SPEAKER_01: and suddenly they're in the frame for being the people responsible for it,
- 25:23SPEAKER_01: when in reality,
- 25:25SPEAKER_01: They were just trying to do the best and put something in that was in
- 25:28SPEAKER_01: the time they had.
- 25:29SPEAKER_02: And it could be tools that are vibe-coded over a weekend.
- 25:33SPEAKER_02: It's shiny objects that people see on LinkedIn and X,
- 25:36SPEAKER_02: which can be,
- 25:37SPEAKER_02: frankly,
- 25:38SPEAKER_02: a nightmare when you think you're And all of a sudden there's this new system
- 25:42SPEAKER_02: and it can be very compelling and there's an urgency there to go,
- 25:45SPEAKER_02: I need to start a new project and embed this now.
- 25:49SPEAKER_02: I would say from my experience with helping agencies adopt AI,
- 25:54SPEAKER_02: Best strategy that you can have is to start on one project
- 26:00SPEAKER_02: and make sure it's carried out in full.
- 26:03SPEAKER_02: So we talked about governance being the start.
- 26:05SPEAKER_02: There's ownership for who's actually responsible for constructing the system
- 26:09SPEAKER_02: and identifying where AI can help.
- 26:12SPEAKER_02: Then there's the people who are responsible for maintaining that system as well.
- 26:16SPEAKER_02: And once that's become second nature and habitual where there's a process now
- 26:21SPEAKER_02: that is measured,
- 26:22SPEAKER_02: of course,
- 26:23SPEAKER_02: that goes back to the margin piece that we were talking about in the sort
- 26:26SPEAKER_02: of topic one of this podcast.
- 26:29SPEAKER_02: Only then should you move on,
- 26:31SPEAKER_02: in my opinion,
- 26:32SPEAKER_02: because measuring it,
- 26:34SPEAKER_02: as we've said, that's almost the end point and to claim margin.
- 26:37SPEAKER_02: But without that final step, There's no point doing it.
- 26:43SPEAKER_01: David Bark at Inspire KBB mentioned earlier,
- 26:47SPEAKER_01: he was a big advocate for that and he showed a number of tools,
- 26:51SPEAKER_01: AI driven tools that he was using in his business.
- 26:54SPEAKER_01: Very interesting and obviously very forward thinking with all that sort of stuff.
- 26:57SPEAKER_01: But I remember one part of his advice was what you just said there,
- 27:02SPEAKER_01: try and take one project at a time.
- 27:05SPEAKER_00: Deploy it, get it right.
- 27:10SPEAKER_02: Make sure it's governed from the start and then measured at the end,
- 27:15SPEAKER_02: so you can translate all of that hard work and focus and dedication into margin.
- 27:22SPEAKER_02: Only then, I think, is the right time to jump off and create something new.
- 27:30SPEAKER_01: The recognition that AI transformation can fail without ownership
- 27:37SPEAKER_01: and true governance and the actions are for people.
- 27:43SPEAKER_01: If you have an AI champion or champions,
- 27:48SPEAKER_01: give them to carve out some time for them,
- 27:50SPEAKER_01: don't assume or don't allow them to Do it just as part of the day job or in between
- 27:56SPEAKER_01: the cracks of the day job.
- 27:57SPEAKER_01: This is important.
- 27:58SPEAKER_01: This is something that people should have dedicated time to.
- 28:04SPEAKER_01: What else as actions on that?
- 28:07SPEAKER_01: Take it one stage at a time?
- 28:08SPEAKER_01: Yes.
- 28:09SPEAKER_02: So treat a project With respect to your client's safety and your own safety in your business.
- 28:17SPEAKER_02: So make sure that you're doing the governance at the start because you want to build
- 28:22SPEAKER_01: a strong foundation ultimately.
- 28:24SPEAKER_01: And if you don't have anybody in the business that you If you either can trust,
- 28:28SPEAKER_01: has the knowledge,
- 28:29SPEAKER_01: capabilities,
- 28:30SPEAKER_01: has the time to do this properly, then look outside, use a quick plug for yourself.
- 28:37SPEAKER_01: Give me a call, yes, I'll be happy to
- 28:39SPEAKER_02: hear your situation and help you out.
- 28:41SPEAKER_01: Use an external AI expert like Callum, other experts are available, we think.
- 28:51SPEAKER_02: Hiring a fractional lead or an AI expert for agencies that's the value,
- 28:57SPEAKER_02: it's actually the knock-on effects of them dealing with
- 29:02SPEAKER_02: the project management themselves.
- 29:06SPEAKER_02: Your team members are very capable and they're heavily involved in the process
- 29:09SPEAKER_02: of embedding AI,
- 29:10SPEAKER_02: of course,
- 29:11SPEAKER_02: but having someone else who has the time to think about that and AI as a part
- 29:16SPEAKER_02: of your business development to actually develop
- 29:20SPEAKER_02: that ownership structure is very nice.
- 29:25SPEAKER_01: And so the third takeaway of many actually,
- 29:29SPEAKER_01: but the third one we want to talk about today was
- 29:34SPEAKER_01: use existing data that you have,
- 29:38SPEAKER_01: use AI to tap into that and uncover growth.
- 29:42SPEAKER_01: And this was something that came out very strongly on the day,
- 29:46SPEAKER_01: but you've got relevant recent Direct experience
- 29:52SPEAKER_02: of doing
- 29:52SPEAKER_01: just this for some of
- 29:53SPEAKER_01: the clients.
- 29:54SPEAKER_02: Exactly.
- 29:54SPEAKER_02: Yeah.
- 29:55SPEAKER_02: And it's my favorite topic.
- 29:57SPEAKER_02: If you're watching,
- 29:58SPEAKER_02: you might see me smiling here because it is the X factor when it comes to AI.
- 30:04SPEAKER_02: It's the difference maker and it's the idea of understanding the importance of context.
- 30:09SPEAKER_02: The ideal scenario is that you have an AI employee that understands the DNA
- 30:14SPEAKER_02: of your business and understands exactly what your team members do
- 30:18SPEAKER_02: and what you yourself as the founder,
- 30:22SPEAKER_02: if you're the founder listening to this podcast,
- 30:24SPEAKER_02: It understands as well because it's a processor of information.
- 30:28SPEAKER_02: We want it to process your DNA,
- 30:31SPEAKER_02: your business's outputs rather than generic combined amalgamations
- 30:37SPEAKER_02: of data it's captured on Reddit and Wikipedia.
- 30:40SPEAKER_02: We need to give AI the right things to speed up.
- 30:44SPEAKER_02: And so agencies have many beneficial assets that if they aren't already using,
- 30:51SPEAKER_02: it can beat Almost instantaneous value.
- 30:54SPEAKER_01: Yeah.
- 30:55SPEAKER_01: And I think if they aren't already using is a fair point,
- 30:59SPEAKER_01: but I'd say I think many agencies and many businesses, again,
- 31:03SPEAKER_01: are at the moment at least blind to the fact that they have these things really.
- 31:07SPEAKER_01: True.
- 31:07SPEAKER_01: Or if they see them.
- 31:09SPEAKER_01: If they do know they have them,
- 31:11SPEAKER_01: they don't see them as assets to be used, they see them as just information.
- 31:16SPEAKER_01: So what we're talking about here is that agencies hold now lots and lots
- 31:21SPEAKER_01: of transcripts of meetings that they've AI generated of their online meetings.
- 31:26SPEAKER_02: And if you aren't already, that's almost instant value.
- 31:29SPEAKER_02: Of course,
- 31:30SPEAKER_02: make sure that your clients are aware that it's being transcribed,
- 31:33SPEAKER_02: but ultimately,
- 31:34SPEAKER_02: As long as the parties agree,
- 31:36SPEAKER_02: it's beneficial for everyone because it allows you to be more present in meetings
- 31:40SPEAKER_02: and capture notes.
- 31:42SPEAKER_01: So notes from meetings,
- 31:44SPEAKER_01: timesheets of course, we won't label that one but we all know they're there.
- 31:47SPEAKER_01: Information and commuting comes from Slack and or email,
- 31:51SPEAKER_01: it's all there to be used but people aren't using that stuff.
- 31:57SPEAKER_01: You've got some live examples, not marketing agencies, but a couple of recruitment agencies
- 32:07SPEAKER_01: Using AI are starting to use those pieces of information in a much better way.
- 32:12SPEAKER_02: Yeah, there's so much data available.
- 32:17SPEAKER_02: I, as one of my services,
- 32:19SPEAKER_02: help agencies build a company brain where they digitize the DNA of their agency.
- 32:25SPEAKER_02: So we've talked about meeting transcripts, time sheets.
- 32:29SPEAKER_02: Credentials is a big one,
- 32:31SPEAKER_02: cred decks,
- 32:32SPEAKER_02: pitch decks,
- 32:33SPEAKER_02: policies,
- 32:34SPEAKER_02: processes, frameworks, design materials, even people and their roles and strategy.
- 32:42SPEAKER_02: What are your yearly goals?
- 32:46SPEAKER_02: Make sure that AI is aware of them so in every single output for every single one
- 32:49SPEAKER_02: of your employees it's basing its responses on a certain goal
- 32:53SPEAKER_02: which is always important for AI.
- 32:55SPEAKER_02: So there's lots and lots of data that's their DNA and then of course there's
- 33:01SPEAKER_02: so many different softwares out there which you want that data as well
- 33:05SPEAKER_02: because that's nicely packaged and easy for AI to read.
- 33:10SPEAKER_02: Using either MCPs,
- 33:12SPEAKER_02: which is through connectors,
- 33:14SPEAKER_02: you'll see on Claude and ChatGPT,
- 33:17SPEAKER_02: where you can connect them straight up, which is one way of doing it.
- 33:21SPEAKER_02: There's APIs through code where you can access timesheet data,
- 33:24SPEAKER_02: and that's a large quantity of data at speed.
- 33:28SPEAKER_02: There's the CLIs, which is slightly more technical.
- 33:31SPEAKER_02: But the core thing to understand is that there's so much data that you're aware
- 33:35SPEAKER_02: that You need to make AI aware of it as well and give it the keys as long
- 33:42SPEAKER_02: as it's governed and it's safe.
- 33:46SPEAKER_02: Give it as much data as you can,
- 33:49SPEAKER_02: but the only data you don't give it is the ones that aren't governed properly.
- 33:56SPEAKER_01: Governance is obviously a fundamental point,
- 34:00SPEAKER_01: but when you give AI access to this data,
- 34:03SPEAKER_01: the things it can do,
- 34:04SPEAKER_01: So if you can give it access to the data that you have on meetings,
- 34:06SPEAKER_01: on emails, on Slack, we talked in the
- 34:19SPEAKER_01: Sessions in the Hackers Conference,
- 34:23SPEAKER_01: people were talking about you can have an early warning system for unhappy clients.
- 34:30SPEAKER_01: You can use AI to pick up sentiment from emails,
- 34:34SPEAKER_01: from meetings,
- 34:36SPEAKER_01: that it can then alert you as the agency founder potentially
- 34:40SPEAKER_01: that this client is giving signals that they're not happy,
- 34:44SPEAKER_01: literally,
- 34:45SPEAKER_01: as in they're telling you they're not happy,
- 34:47SPEAKER_01: or maybe even more subtly,
- 34:49SPEAKER_01: Some of the things that they're not doing can give you an early warning.
- 34:53SPEAKER_02: In the same respect,
- 34:54SPEAKER_02: before you bring another one in there,
- 34:56SPEAKER_02: you can monitor the same resources like email and transcripts and identify client
- 35:03SPEAKER_02: So data can give you reactive information,
- 35:05SPEAKER_02: as in clients are unhappy, or proactive information.
- 35:27SPEAKER_01: Yep.
- 35:27SPEAKER_01: There's an opportunity.
- 35:28SPEAKER_01: The clients said in three meetings now that they need X or they were looking at Y
- 35:35SPEAKER_01: and nobody in the agency has picked up on it.
- 35:38SPEAKER_01: Nobody's responded.
- 35:39SPEAKER_01: Nobody said, we can look at that for you.
- 35:41SPEAKER_01: We do that for you.
- 35:41SPEAKER_01: For the clients, that's something that we could actually help you with when they should.
- 35:47SPEAKER_01: So upsells, yes,
- 35:49SPEAKER_01: but quotes,
- 35:50SPEAKER_01: of course,
- 35:51SPEAKER_01: the timesheets,
- 35:52SPEAKER_01: quotes based on actual time rather than guessing,
- 35:56SPEAKER_01: using the data in that way,
- 35:59SPEAKER_01: automating responses to leads and proposals, costing and all that sort of stuff.
- 36:10SPEAKER_01: There is a scratch on the surface here, but those are some of the things.
- 36:12SPEAKER_01: But the third takeaway was use your existing data,
- 36:17SPEAKER_01: use AI to use your existing data because it could really unlock growth,
- 36:22SPEAKER_01: could protect you from losing clients that maybe some of your junior staff
- 36:30SPEAKER_01: or whatever are not seeing the signs as well as you might if you were involved in
- 36:34SPEAKER_01: the meetings.
- 36:34SPEAKER_01: But I think that canary in the coal mine Early Warning System of Clients,
- 36:40SPEAKER_01: that could be a really powerful one.
- 36:42SPEAKER_02: Yeah, and that ultimately all falls down to context.
- 36:46SPEAKER_02: The more you can accumulate,
- 36:49SPEAKER_02: the more you can measure,
- 36:50SPEAKER_02: the more visibility you have to every part of your client relationships.
- 36:55SPEAKER_02: Even internal processes, something like a lot of time sinking to estimating new projects.
- 37:03SPEAKER_02: You can use your past project data,
- 37:05SPEAKER_02: actually per client,
- 37:07SPEAKER_02: if you've had a client for a long time,
- 37:08SPEAKER_02: you've got a new project that you're forecasting and sort of estimating for,
- 37:14SPEAKER_02: you can use the insights that you've developed in the past few years and use that
- 37:22SPEAKER_02: to make you the best case
- 37:26SPEAKER_02: Takeaway 3 was user-existing data with AI to untap
- 37:28SPEAKER_01: and unlock some growth.
- 37:28SPEAKER_01: The action for me on that one was think about data in a different way.
- 37:30SPEAKER_01: Data isn't just lists of numbers that are
- 37:53SPEAKER_01: So the action is to think about what have we got access to here and using AI to very quickly
- 38:13SPEAKER_01: Interrogate those things and pull us,
- 38:15SPEAKER_01: what sort of reports,
- 38:16SPEAKER_01: what sort of information could we pull out in a way that we've just described there?
- 38:21SPEAKER_01: What could be proactive information as in upsells?
- 38:25SPEAKER_01: What could be reactive information as in spotting issues with clients
- 38:29SPEAKER_01: that you might not spot necessarily in meetings that you've been having?
- 38:35SPEAKER_02: Like we were saying in point one,
- 38:37SPEAKER_02: where we could see a future that Agents' time is being tracked,
- 38:43SPEAKER_02: and then humans' time is being tracked.
- 38:45SPEAKER_02: There's almost two sets of employees, two types of intelligence.
- 38:49SPEAKER_02: The same way in which your employees need food,
- 38:53SPEAKER_02: water, and sleep to operate, AI needs data, that's its equivalent.
- 38:58SPEAKER_02: So if you don't give it data, it's not going to operate, it's not going to function.
- 39:03SPEAKER_02: That's how much severity you should treat.
- 39:11SPEAKER_02: It's an important place to cover.
- 39:15SPEAKER_02: That's how severe it is.
- 39:16SPEAKER_02: AI needs that data.
- 39:17SPEAKER_01: Totally.
- 39:19SPEAKER_01: Great soundbite there at the end, Colin.
- 39:20SPEAKER_01: Well done.
- 39:22SPEAKER_01: I guess we should wrap this one up.
- 39:24SPEAKER_01: That was three takeaways from the Agency Hackers.
- 39:27SPEAKER_01: Robots are coming.
- 39:28SPEAKER_01: October meeting in London.
- 39:31SPEAKER_01: There was many others.
- 39:31SPEAKER_01: We took loads of notes.
- 39:35SPEAKER_01: We came away with lots of great things that supported what we already thought,
- 39:40SPEAKER_01: things that challenged sometimes what we were currently thinking,
- 39:43SPEAKER_01: and ideas and thoughts that we thought we should help our clients work on those sort
- 39:50SPEAKER_01: of things even more,
- 39:52SPEAKER_01: or those are ideas that we could utilise within our own business,
- 39:55SPEAKER_01: Agents of Change, or with our clients as well.
- 39:58SPEAKER_02: Like how we think an agency should Have an element of focus and start smaller
- 40:04SPEAKER_02: and finish projects.
- 40:06SPEAKER_02: In this podcast,
- 40:07SPEAKER_02: although there's a lot of noise and a lot of different topics to discuss,
- 40:11SPEAKER_02: we think those are the three main focus points.
- 40:14SPEAKER_02: And yeah, it's been a great episode.
- 40:17SPEAKER_02: Good episode.
- 40:17SPEAKER_02: Pleasure.
- 40:18SPEAKER_02: Thank you very much.
- 40:19SPEAKER_02: See you on the next one.
- 40:19SPEAKER_02: Yeah.
- 40:20SPEAKER_02: See you later, guys.
- 40:24SPEAKER_02: You've been listening to Gareth Healey and Callum Healey.
- 40:26SPEAKER_02: Together, we are Agents of Change, which is also our URL, weareagentsofchange.com.
- 40:32SPEAKER_00: If you want support with growing your agency,
- 40:34SPEAKER_00: including how AI can help you future-proof your business,
- 40:37SPEAKER_00: visit us online to learn more about what we do and find out how to connect with us.
- 40:42[music]