Hey.
Something weird just happened, and it's going to sound unbelievable, but I swear it is real. I was just doing my grocery shop — not a big shop. It was a small shop. You know? The small shop that you do when you've already done your big shop, but then you've somehow forgotten all of the key ingredients for every meal that you planned.
Anyway, I'm there. I'm standing in the queue for the self checkout. I got my arms absolutely chock full of things half falling out because I decided at the start that I didn't need a basket, and then I definitely needed a basket.
But, anyway, I'm just standing there, and this man suddenly appears right behind me and gave me a start. And I I dropped my broccolini, and that was annoying. It was absolutely it was the last bunch of broccolini they had. But this guy, he just appeared out of nowhere wearing what I would say were very slightly futuristic looking clothes, and he shoved something into my arms.
And that made me drop my 3 passionfruits on top of the broccolini, and I was not happy at this stage. I was just about to complain to the guy, and he just whispers right in my ear, “listen to the platypus”.
And then he disappeared like a teenager when it's time to clean the kitchen.
So weird.
So I put down all my shopping, and I saw that what he had left me was a USB stick, a USB-C, which is unusual.
So I got home, and I immediately plugged into my laptop, of course, in absolute blatant contravention of every Help Scout security policy. Sorry, Pilar.
Then on this USB was a recording of a corporate meeting—and it sounds dull and you're not wrong—but this was a meeting…from the future.
I mean, not very far into the future, 2028. Just enough to explain the clues, I guess. So I started watching it, and, well, kind of halfway through, I decided that, you should all hear it too, and you'll see why. So what I'm gonna do is play it in for you now. I might have to pop in a couple of times to explain some of the more visual things, but I'll be back at the end with a few thoughts.
From the 2028 video
Hello, and welcome to your first day of orientation here at Platypus. I am so excited to see such a lot of you support professionals joining the Platypus support team in this November 2028 intake.
I'm Mike Peterson. I'm a support onboarding specialist here at Platypus and it is my absolute privilege to help you get started in your new roles on our various support teams. Now you've all got an amazing week of orientation ahead of you, so I won't take up a bunch of your time. I'm just gonna use this first session to give you an overview of the job of support here at Platypus and how we work together to support our customers and what you might find different from some of the places you've worked in the past.
Now, at this point, you were already supposed to have your own wearable AI work buddy, your PipBoy.
It stands for performance improvement plan beginning of year. But unfortunately, the driverless delivery truck, it's stuck in a standoff. It's on the highway with some government drone pigeons, and they've been there 11 days now. I'm sure it's gonna be a result, but until then, you've just got me to talk you through this all. So please forgive my rustiness. It's been a little while since I had to run one of these without some AI help.
I'm sure you're all excited to just dig into that work, but just let me give you a little bit of background context.
Now, most of you will already know this, but Platypus is a conglomeration of several smaller software companies stitched together, and that's actually why we're called Platypus. When the actual platypus was first “discovered” by Europeans in Australia, people genuinely thought it was a fake. I mean, it had a a duck's beak. It's got a beaver's tail stitched onto the body that has somehow otter's feet on the bottom.
Now in a way, we here are the corporate equivalent of a platypus, where all those different parts smooshed together into an unholy looking beast that still somehow works beautifully in its environment, which in the case of the actual platypus, of course, is a riverbed and in our case is the small to medium business software market.
And, of course, the platypus itself is a monotreme or an egg laying mammal. One of the things that you'll support here is “monitoring” software. That's just that's just our little joke. That I mean, our marketing folk made that up. They are so funny. You'll meet them later. #TeamMonotreme.
Yeah.
Anyway, here at Platypus, we value that sort of diversity. You know, whether you are personally more of a beak or you're a bill or a tail, even if you're a bit of an egg pouch of a person, you're all a vital part of the whole.
Now, some of you in this room probably haven't done frontline support since, what, 2024?
Whoo. What a time that was. We sure didn't know what was about to happen to us. But aren't we all glad to have made it through to today? But since it has been a while for some of you, you will understand if I just go over some of the basics today and try to bring you up to speed on how customer support works in 2028.
Firstly, I would say don't be scared. Things really aren't hugely different than what you were doing 4 years ago. You know, we're still here answering customer questions by email, although there is less of that these days, and by chats and phone calls and messages. And the core of the job is still figuring out what people need, understanding that that's not always what they ask for, and then helping them get that thing.
But there definitely have been some changes over the last 4 years. So let's just run through it in these 3 areas. How customers get helped in 2028, onboarding and growing in your support role, and the place for humans in service.
Okay. So at Platypus, our customers come from all over the world and we offer 24/7 support using the hybrid human-AI model that's pretty much become the industry standard. So we don't have human staff in every time zone, but our AI system does offer live response coverage all day long.
That AI, which incidentally we call Parry, to avoid copyright issues from the surprisingly litigious lawyers of Phineas and Ferb. Parry is able to directly respond to a chunk of the questions itself, but it can also escalate to a human on call team when necessary.
And actually on that, just, one little note I've been asked to share with you. Please try your best, not to swear at the AI agent, you know, if it does wake you up. I know it's not human. You can't hurt its feelings. It doesn't have feelings.
But at Platypus, "We Strive to be Inclusive".
You’ll have read that on the wall in the break room on the way in.
Also, there was one former support agent. They they'd just been woken up when they were on call. They thought they were talking to the AI. They gave it a little blast.
It was actually the CEO. And, well, it's just easier not to do it.
Anyway, so we provide multichannel support, and, here's how it breaks down.
Mat from 2024: Hey. It's Matt here in 2024. You can't see it, but there is a pie chart being shown on the video.
It shows the different support channels and what percentage of incoming support they all represent. And the biggest one's self-service. That's about thirty eight percent, then email at twenty six percent, live chat at sixteen. I'll put this in the show notes for you.
Right now, back to the mysterious recording from the future.
2028 video:
So that's probably pretty close to what you're used to. But over the last few years, there definitely have been some shifts, and the biggest change that you'll notice is that shift to AI mediated self-service.
You know, of all the many, many promises that those AI companies were making back before the AI bubble burst in in 2025, the one that they really did deliver on was that conversational interface. That ability for people to ask a question in the rambling and vague and weirdly phrased way that we all do most of the time and then have a computer translate that back into a reasonable search query. That really has been a support game changer. You know, we see much higher hit rates for people who are able to find the right answer that's already there in the knowledge base.
But some of those teams and some of those companies that went really hard on AI only customer support, most of them failed. Some of them pretty spectacularly, at least for any of the markets where there is some margin to spare. And most of those companies, they have had to backtrack.
But conversational interfaces, they're everywhere. Today, almost all our support interactions start with the customer describing their problem in plain language to an AI either by text or by voice. And from there, they'll most often be given some self help resources.
That might be a knowledge based article, it might be an on the fly custom generated video. Sometimes it's a setting or a tool that we can bring out of the product and show right in the chat interface so that the customer can describe what they're trying to do and the AI can make that change for them with their permission.
Lots of those support interactions are triggered, from inside our apps, you know, places where the AI already has some context about who the customer is and what they seem to be trying to do, contextual help really has been a primary tool for support teams. It's been the rebirth of Clippy.
Although, of course, not many of you are old enough to know the original Clippy and for that, you should be grateful.
So in practice, no matter which support channel customers choose to use first, the first response that they get is probably going to be an AI mediated self-service response. And we will find that about 30% - 40% of the time at Platypus, that will resolve the issue satisfactorily, you know, in general. For some of our products, that number is a lot higher. For others, it is much, much lower. It really does depend on the type of product, the type of audience, the form of support questions that people need to ask. But even in the very best case, that leaves a lot of customers who still need to be connected with a human, someone like you all.
We've done a lot of work trying to make that AI to human transition as smooth as we can. We do still find that's the part where people are most likely to be dissatisfied with the support service. And actually, one of the ways here at Platypus that we differentiate our support from our competitors, your Wombats or Echidnas, is by making it easy for people to opt in to that human help. And we absolutely do gate it a little bit. You know, if you're on a free plan, our customers can't choose to speak to a human first. We'll have to manually escalate them into the human support cycle.
But if you're a paying customer, you can always find a way to reach a person.
Although, actually, not all of our customers are even human themselves.
What we're seeing is more and more AI software agents that are requesting help on behalf of their particular human. You know, so instead of our support AI working with the customer directly, it'll be the customer's AI agent dealing with our AI bot.
And I would say that's been interesting. You know, sometimes it works fine. Everybody gets what they need.
Sometimes though, I mean, you might have seen the VR news reports about that lady who woke up to a massive pallet, four thousand pairs of customized crocs dropped on a front lawn because she'd accidentally misconfigured her AI agent and it got stuck in a loop with our support bot.
Anyway, look, it's all sorted out now. Customer is happy.
Although, speaking of, in your welcome pack that you will have received, just have a look in there. You'll probably all have a pair of, custom printed crocs in a women's size seven.
I hope you can find a use for those. Personally, for me, I've used mine as just a fun planter for basil. Just an idea. You do you.
But back to the point. Mostly, you will be dealing with human customers and that will be after they've already moved through some sort of self-service. That is where your work as a team really starts.
Okay. So before we move on to the next section, the people ops team, they just wanted me to run you through some work from home guidelines. So just a second. Okay.
So, Platypus currently respects the right of our team to work from home. But please do refer to your email each day, before you come to work just for an update, just in case that policy has changed. We do of course have beautiful offices that you may choose to attend and what we've done is synchronize the in office and the home office rules. Right?
You'll find all the details in your handbook. Just a couple of key reminders if you are working from home, no fish in the microwave, please, during working hours.
There has been an increase. We now allow up to three five minute toilet breaks without authorization.
Thanks to everyone who filled in the survey about their habits in that area. That's a 50% improvement on the previous policy. So great stuff.
And, of course, no more than 2 personal items on your desk exactly as you would be within the office. Although, of course, at home that really means within the angle of view of your webcam. So no more than 2 personal items there, and, this applies to everybody but including at home. The following emojis are no longer acceptable.
Eggplant, toilet, double eggplant, cry sob pirate. I'm not going to get into why. I think you probably know. Speak to your line manager for details.
And of course to keep you safe and productive, your facial expressions will be logged at all time just to make sure that you're sort of within the emotional boundaries defined for your role.
And of course, your secure pro home monitoring device will need to be on 24/7, just so that you are kept safe and surveil…safe and secured as you would be at the office. Because at Platypus, Equality is a Core Value.
Alright. In, part 2, I want to talk about what it will be like to get started in your new support job here and a little bit about your ongoing role growth opportunities.
So over your first few weeks after this induction, you'll work through some training modules, you'll shadow a veteran support rep just to see how they work. And by about day 3, you should be able to log in to every Platypus application, internal and external, that you'll need to do your job. And, of course, they will be biometrically secured. So we'll do that using the blood sample you'll be providing at the end of this session.
But, of course, you won't be working…
AI voice: “You're doing so great, Mike. Nobody is even wondering how you got your job. Why would you think that?!”
Sorry. Let me just, yep. Okay.
Right. Yes. The first inbox that you work with is filled with realistic questions, but not real questions. These are all questions that we've had our AI, Parry, generate based on real support questions from real customers. And you can open those questions, you can read them, and you can respond to them exactly as you would a real customer, but you'll only be responding to an AI that can have a conversation with you. So even if you go really badly off the rails, which I'm sure you won't, well, maybe you (points at viewer). But generally, no.
And it's okay. Right? Because you're not upsetting a human customer. Although having said that, you will possibly be irritating the AI that will later be used to judge your performance and that feeds into your salary review process. So do try your best.
But then your team leader, who will typically be a human, they'll keep an eye on your progress. They'll make the call as to when you move out of that training queue into the real support queue.
And in that queue, the work is gonna be mostly assigned to you by the AI.
It handles the triaging for priority and the sentiment analysis and the types of customer. And that's been a big improvement that you might not have seen a few years ago. We don't just do round robin. We have a much smarter way of sending you the sorts of questions that are most appropriate to your current skill level, the areas that you're trying to work on, depending on who else is in the queue and who can do what and how large the queue is. It's a much more intelligent sort of queue management than you might be used to. And there's, of course, also a human queue director overseeing that process. And you can always come back and you can practice with that AI whenever you'd like.
And the QA tool, actually is super helpful to you. So, you know, when you first get started on a on a support team, at first you're constantly worried that you're going to get fired because you don't know what you're doing and you're not getting the immediate feedback about am I quick enough, am I accurate enough? The QA tool and your personal dashboard, they are super helpful tools that will tell you how you're doing against the requirements of your job and how you're doing compared to other people doing the same sort of work. So you'll always know at any time where you need to improve, where you're doing really well, an area that you might want to do some training in, and the dashboard will show that to you personally.
And every support role at Platypus actually has a defined tier system. So whenever you're ready, you can work with your QA Pal. It will tell you what you need to do to qualify for the next tier in whichever area you want to be in. And that might mean you need to tag some specific examples of where you've done a thing.
You might need to do some additional practice sessions. You might need to do some independent learning. But once you have self qualified for that next tier, you can talk with your manager about upcoming opportunities. And there's also many other roles within the support organization that you might want to move into.
We'll cover that in the next section. But first, I just need to go over a couple of onboarding tasks with you. Okay. You should have received in your welcome pack all the usual forms you're going to fill in, but just pay special attention to your health care process, because as you probably know, Platypus is now a US-owned company and so we are required to follow the US government guidelines in that area.
Some of you might not have previously worked under that system, but basically what they've done is gamify it for maximum engagement. So that means, to select your health care plan, you're gonna play a game. Each of you will play your own game of, Healthlife 2. You'll have 3 lives and you wanna get as far as you can into that game because your final score will determine which health plan you're assigned to for that fiscal year. Good luck!
Secondly, just make sure you download and install “Cheerleader Gold”. That's our employee morale maximization solution. Pop that on your phone and then every 12 minutes, it is going to ping you with a morale boosting message or a GIF or a sound or a video. All of them customized just for you by the AI based on all the psychometric testing that we made you go through during the interview process. You probably saw mine pop up just a minute ago there.
Just a tip from me, make sure you acknowledge those in a timely fashion, you know. Don't you don't want to upset the cheerleader. Like, it starts off super friendly and helpful, but if you leave it too long, it can get a little…unsettling…let's say. But it's great.
Okay. Time for our final part of the session. This is about how you as a human support professional are going to work with our customers.
And a support team at Platypus, it might be smaller than you think, you know, given the number of customers that we're able to support. Definitely, the ratio of human support to the number of customers supported, that has shifted over the last few years. You know, with self-service and AI covering a lot more questions, we don't need quite as many frontline people as we used to have per team. So a typical support team structure in the Platypus organization looks like this.
So our teams are based around a self-service first model all underpinned by AI. Almost everything that in support inbox comes in first through self-service interactions.
The AI will answer some of them. Only what can't be resolved through self-service and AI will end up in the queue with the human staff. And what that means is that the questions that do come to our human support team, they tend towards the more complex, the more nuanced, the tricksier ones.
So on the chart it says, First Level support but really the first line of support here for people, it's probably more like tier two or tier three would have been in your past experience, certainly in terms of the complexity of the questions and the expertise required to answer them.
And you'll see that there are some other roles on the team. I'd say you're probably familiar with most of these. You've got your queue director there. They're the person who's working with the AI tool to monitor the overall support queues.
They might override some AI decisions on prioritization. They might tweak the way the team is working. And they'll also look after all the first level support folks. That's where most of you are going to start.
We do find that, a lot of our most effective support agents end up moving to the Education team.
If you do that, that's where you work with the the knowledge and the product teams on content, of course, but really the main role is to help customers make the most of our products but just on a one-to-many scale. So that would be running classes, writing emails, hosting webinars, building out samples.
They're heavily informed by their frontline work, of course, but that is where support teams can add a huge amount of value to a company.
Also within that group are the knowledge experts, the people that own the documentation and the way that the documentation is used across all of our platforms. So that's how it shows up inside contextual chats, how it's interpreted by the AI, how it's maintained, all of that stuff. And those people attend product meetings, they're part of the development process. The AI is a huge new consumer of, knowledge base and documentation and so, consumer of knowledge base and documentation and so it requires a lot of focus.
Similarly, over on the Support Operations side, there's people who are driving tools and processes keeping up with the current technology and they're selecting it, testing it, implementing it. That takes up a lot of time. So we find that the people who are sort of into the AI technology tend to end up in the operational roles.
QA. QA is another role, has changed a ton. It used to all be about how well are our people answering questions for customers and that is still a part of QA's remit, but now they're also responsible for the quality of service that's received by people who may never speak to a human, who only speak to the AI or only use the self-service tools.
So the QA team, they'll work with support. They'll also work with engineering and knowledge and ops and education. They will shape the output of our AI tools as well as the output of our human support team.
And account managers—been around forever, still exist. For larger customers, some people just want to pay more for access to a consistent support person. They're still mostly coming in through self-service channels most of the time but any support that comes through that, gate will end up with the same support person. It builds a relationship. A lot of customers don't need that constant attention but they're still willing to pay for it when they do want it.
Now, you might find yourself in any one of these different areas over time and we'll figure out together where you're most effective, where you're most engaged and you might move around between these roles. What we found is that our sort of AI self directed learning makes it a lot quicker for people to get to speed and to be able to move around between teams.
So that's basically it and just to wrap up, I wanna talk about AI.
You know, we've spoken a lot about AI today and, of course, you're gonna get more of it in the rest of your orientation. But at Platypus, we do not rely on AI alone to talk to our customers. It's just still a little bit risky. You know, generative AI, it's never quite solved its hallucination problem. It's still confidently wrong sometimes like the drunk man at the party that you desperately want to avoid.
His name, by the way, is Garth. He works on level 3. This is him.
If you do see Garth this evening, I would say the best way is just ease yourself out of sight, don't make eye contact. Ideally, you want to get a salesperson between you and him. I personally found that to be the most effective method.
But the biggest problem is not when AI—or indeed Garth—gets it wildly wrong. It's when it's plausibly wrong. You know, the closer it gets to 100%correct, the harder it is for people to notice those errors. It's just difficult to pay attention all the time when most of the time you don't need to.
You know, it's like self driving cars. There's a robot there sitting next to you who says, “yeah, I can drive you. It's going to take 3 hours to get there and for 2h58m, I'll be absolutely fine. But there will be 2 minutes, can't tell you when, but you're going to need to take control immediately or we will both die. So let's go.”
And it's just not the way our brains work. We always have problems with support following customer service scripts and then not noticing that those scripts didn't exactly apply.
Generative AI has sort of aggravated that issue for a lot of companies. But that's why at Platypus, we are always self-service first, not AI first. You know, if we can't get people a self-service answer with the help of AI where it makes sense, with proper limits and monitoring and QA, well, that's why we still have all of you lovely human support folks here.
Some of our competitors absolutely are AI first. Maybe they boast about their 20 second response times. But good service is more than just fast service. Right?
Otherwise, McDonald's would have all of the Michelin stars. And, of course, in the 2028 edition, we all know they've only got 2 stars. So we are human-first here but we are humans amplified by AI. We think that is the sustainable model for high quality support.
Alright. I can see you're excited to get started on your support journey here at Platypus. Let me wrap this up. Customer support always been about people helping people, person to person but also people building tools and systems and processes to help people. And in 2028, more of those tools than ever before are built on machine learning and AI, but they're still only slightly more intelligent tools. That's what they are. They can't actually care about providing good service.
They can't care about anything no matter what your Cheerleader app tries to tell you. Fundamentally, we're still doing the same thing we have always done and that is great.
We're just adapting to the new options that are available to us and those tools can amplify your care. They cannot create it. You have to bring that with you.
Also, your swipe card, you'll have to bring that with you as well. Otherwise, you'll be stuck in the lobby. That's where Garth often holds his “meetings”.
Anyway, thank you so much for listening. Let me once again say welcome to the Platypus family.
And, legal insists I tell you at this point we are no contractual or moral sense, a family.
That's it for me. Don't forget to give your blood before you leave.
I'll see you around the office.
-> Mat from 2024
Wow.
What a glimpse into the not very distant future that was.
Do you think that voice sounded like weirdly familiar? I I can't place it.
But what do you think this, unidentified weirdo wanted me to hear about? I mean, I've been thinking about it. I think it was really a message for all of us because it seems to me it was it was both a warning but also an encouragement. You know? A warning that, yeah, these big changes are coming to customer support. All these new tools, new capabilities, they're gonna create a whole lot of adjustments, new ways of working.
But also it's an encouragement that customer centric businesses will still need to be investing in people and that the best customer experiences will still come from people helping people.
And if that is the future that we're heading into, does that change anything for us? If you were setting up a customer support team today, knowing what 2028 looks like, what would you do differently? You know, where would you invest your time and your energy? Because I think that the most impactful customer support teams will absolutely be using all of the available tools to help their customers achieve their goals. And really, that's always been true, hasn't it?
But to get from here to there through this period of technological change, we are gonna probably have to deal with a whole bunch of false paths and dead ends along the way.
But I think if we can keep our focus on the customers rather than on our own systems and our own issues, we will be able to spot those opportunities and the pitfalls, along the path to that point.
Will my mysterious platypus friend ever return?
I don't know.
All I know really is that I'm definitely gonna pick up a shopping basket next time, just in case.
