Webinar: Revolutionizing retail pricing with agentic AI

Date: 8 May 2025

Location: Online

This 30-minute webinar will explore how agentic AI can revolutionize pricing strategies in the retail sector.

Participants will gain practical insights into innovation opportunities, key challenges and success stories around AI-driven pricing optimization, including a live product demo showcasing Peak’s capabilities — offering unique insight into how AI is powering perfect pricing for leading retail brands.

The Speakers

Tom Summerfield

Retail Director, Peak

Lucy McCann

Product Manager, Peak

Chris Ashley

VP Strategy & Partnerships, Peak

Watch now

Right. Hey, everyone.

This is fun. Thanks for joining.

Really quite a lot of you today, which is really, really cool. Thanks for giving us the time.

What we’re gonna be talking about today is AgenTic AI and how it is impacting, or it’s going to be impacting, like, retail and pricing specifically, because there’s a lot of buzzwords and stuff going around in this space, right now. By way of introduction, I’m I’m Tom. I’ve I’ve led the retail go to market at Peak for just over five years, and actually was a customer of Peaks. Before that, I’m a retailer by trade. I’m joined with, today by Lucy McCann, a senior product manager at Peak, also an ex retailer, and is currently leading our pricing work. And Chris Ashley, who is VP of strategy and partnerships at Peak.

I think I hope you don’t mind saying it again, Chris. I thought pretty much I thought a really proper thought leader in this space and and many of us, from from spending, you know, ten years consulting, enterprise businesses on on, technical on technical stuff. So, we yeah. Peek, if you just in case you don’t know, Peek are an AI company.

And actually, incidentally, we were an AI company before the whole world started calling themselves an AI company.

And we were recently really cool, acquired by a much larger company than Pete called UiPath, who are a leader in the RPA space. And actually, a combination of our products is gonna be enabling this stuff for for a lot of organizations going forward. So we might reference some of that, as as we go through it. So let’s get moving.

We yeah. So all all of this is, set on a backdrop in retail that is, like, feels more challenging than ever. There’s always this sort of, like, slightly hyperbolic, like, doom and gloom going around, but it does feel quite real. Again, all of a sudden, like the, there is we just as we were coming out of a cost of living crisis and that is still gonna persist for a while, we were not quite out of it, for sure. There’s now global disruption, you know, it it’s never ending the amount of things that retailers have to deal with and all of it is put in just ends up putting squeezes on margins basically. Consumer demand, rising costs, all the usual stuff off, and it’s not fun for a lot of retailers.

So, you know, this is this is a a gloomy backdrop for some. And I think in the world of AI as we’ve as we see it, there’s a lot of people trying to explore it which is really good and necessary.

The but that we we do see a lot of, what I would personally, like, deem a slightly more gimmicky AI work happening.

And Peak don’t try to indulge in any of that. We focus on some, like, fundamental pillars of businesses that, like, may means that your how much EBIT you make at the end of the year changes, as opposed to sort of efficiency gains, which actually come as part of this anyway. But you can see the types of use cases that we focus on in in in retail specifically. We also play manufacturing and CPG, but things like how, brands and retailers rebuy products or they replenish their stores, you know, you can see. And we’re gonna we’re just gonna focus on on, on the pricing space today.

Yeah.

Yes. So I guess, why aren’t retailers talking about Agenti k I yet? So as a product manager, it’s a big part of my job to be speaking to customers and prospects and, going to retail events, just to understand, like, trends and tech appetite and the adoption and what’s going on.

Tom and I were at, the retail tech show, last month, I think it was.

And, yeah, we went to lots of talks, spoke to a lot of the vendors there. But something we noticed that there isn’t much discussion at the moment about Agenstic AI yet in that retail space.

And we’ve been thinking, like, what is it? Why why is that? Is it a lack of understanding of its capabilities, or is there something else here? I know Chris, and Tom, you’ve both been running some really cool master classes to support customers and prospects recently.

So it’d be kind of interesting to understand, like, what response you’re seeing here. But as well, I guess, Tom, being in retail historically as well, like, what what do you think?

Yeah. And historically is probably the right word at this point. So, I have a slightly cynical perspective on this. This is my personal view, is that a lot of retailers are just think they’re too busy to really, like, try and indulge and get in their head around this. Like, that’s not true of everyone.

But over the years of it’s been interesting coming from the industry side to this side of the fence as it were to see, you know, the privileged position of seeing lots of different types of businesses and different appetites. Ultimately, if someone’s talking to peak, there’s a cause and effect that whether they’re probably a bit interested to find out how they could transform some stuff.

But, yeah, I think the the I think with this feels like it’s probably in the super future for a lot of retailers and the reality is some of their you know, some of their competitors might be starting to embrace it. I think there is a whole understanding piece though, Chris, and I’ll I’ll pass to you here because we’ve got a few slides, I think, that that talk to this. Right?

Yeah. Absolutely. I think, I think you’re spot on when when it comes to understanding. And I think one thing we’ve observed over the years when we’ve been going into boardrooms or, Chatham House sessions with CEOs of retailers around the UK and further afield, the shared definition of AI in boardrooms today, it isn’t that great, and it it’s partially a byproduct of the fact that there’s so much happening in in the AI space currently and over the last couple of years.

But we think it it’s really important for leadership teams, many of you on the call, that that lead merchandising functions, pricing functions, or or lead entire retail businesses creating the shared understanding of of AI and what it actually means in your business can have kind of seismic effects on on performance as people become more more, capable of having conversations about it across the business. So what we’ve got here, is just a really brief explainer, of AI from our perspective. So when when we think about the field of AI overall and when someone says AI, what we’re really talking about is the general field that aims to build systems that can perceive reason and act effectively trying to emulate aspects of human intelligence by leveraging, a machine’s capabilities.

If you jump to the next one, Tom, we’ve then got machine learning, which is a huge subset of the broader field of artificial intelligence, and it’s one of the the primary approaches to build AI systems. And it’s a general class of techniques that allows machines to learn typically from large volumes of data. So Tom mentioned earlier, we we do a lot, of work in the in the rebuying space for retailers.

And under the hood of that rebuying capability is a demand forecasting capability capability, a very sophisticated, multi echelon demand forecast.

And and the primary capability that is being leveraged there is machine learning, learning from all of the swaths of historical transaction data, within your retail business over many years and then predicting and learning patterns, within that data and predicting things about the future. So that that’s an example of machine learning. We’ve then got deep learning, which is a subset of of ML.

And this is all about using huge, huge, volumes of data. So significantly larger than just a few years of transaction data, and leveraging the artificial neural network models, to to, again, learn patterns from data, but it tends to be used for much more sophisticated use cases like, in in a retail environment. It might be things like a digital twin simulation of your business.

For, normal consumers or everyday people, we’re we’ll often see deep learning models leveraged in things like self driving cars and stuff like that. So really powerful capability, requires huge compute.

And then underneath deep learning, we’ve then got our newest favorite field of AI, generative AI. This is obviously what’s captured kind of consumer, hype and attention over the last couple of years with the advent of foundational models, like, OpenAI, and ChatGPT, Anthropic with Claude, Google with Gemini, and others also exist. And these models, have been leveraged with a simple user input to then create text, image, video, and more. And this is a huge capability that’s come about from a lot of innovation and r and d in the deep learning space.

And just when we thought we were starting to get our heads around, all of these different fields of AI capabilities.

A new emergent capability has arrived this year, we think, called AgenTek AI, which Lucy referenced earlier at the, the retail tech show, not having that much attention on it currently. But when we look further afield globally, you can see Jensen Huang, CEO of NVIDIA, Mark Zuckerberg, CEO of Facebook, Sam Altman, CEO of, Open Eye AI, and a range of other thought leaders positing that twenty twenty five is the year of AI agents and Agentech AI.

And we can also see it in some pockets of our customer base.

A lot of manufacturers and CPG companies are starting to give us a lot of attention, especially the big enterprise organizations.

I think in in the UK retail, and probably in the mid market retail, we’re not seeing as much attention, but I think that’ll start to that’ll start to change in the coming months because this capability, it is gonna drive huge, huge business performance uplifts, for a lot of businesses that get get on it early.

And we’ll talk through a little bit of why we think that that will happen shortly. But first, let let’s talk about what AgenTek AI is. So we’ve, my my mom, actually, works in business development, and she, asked me about AI the other day. She’s always known I’ve worked in AI for a while.

But she asked me the other day what what’s all this, kind of AgenTic stuff because she’d read an article about it. So we thought, how’s the best way to explain this, to my dear old mother, and just then to chat GPT to say, if I was trying to explain, AgenTic AI to my mom, how would I explain it? And this is what it what it kicked out. So AgenTic AI is like a helpful assistant that can figure out what to do and how to do it all by itself, unlike older, more traditional forms of AI, which just answer questions or follow predetermined instructions, one at a time.

AgenTek AI can make a plan, take steps, adjust if things go wrong, and keep going get keep going until it reaches the goal, almost like a little digital worker with initiative. So I I quite like this, and I think the key here, I always say that the the key is in the name, Agentech. It’s all about agency. It’s all about autonomy.

It’s all about a new capability that previously, when we view software agents or, software programs, we will have had to predetermine and preprogram the logic for that bit of software to achieve a certain task, and it might follow simple if this, then that statements. It might have a rules engine underneath, but it would have to be predetermined and told what to do effectively.

What we’re now seeing with AI agents and through this definition on the screen are software, software agents infused with AI at the heart of them, which effectively enable them to act with autonomy to achieve a goal without necessarily being told the exact path to traverse in order to achieve it.

So let let’s bring that to life and double click slightly in into what that looks like. So the anatomy of an agent at the heart of it is is the LLM, the large language model. I mentioned a few of the foundation models earlier, likes of OpenAI, Anthropic, others.

And the LLM provides the fundamental intelligence and natural language understanding capability that enables it to, function as the brain.

We then have four separate capabilities, within an agent. One is the ability to plan. So an agent, when it receives an object that if it will typically break that objective down into a series of subtasks. So if I said, help us write a webinar script, it might go away and break that down into, well, what do I need to do to write that script?

And it will operate almost like a team of humans if we sat down in a room together and said, let’s whiteboard this out. What what does the webinar need to achieve? What what are the goals of the webinar? What’s the optimal structure?

What’s worked well in the past? What didn’t work well? All that stuff would become subtasks that we then go away and try and optimize for. So that’s what an agent is capable of doing.

Next one, Tom.

We’ve then got a really important one called tool use, which I’ll come back to in a minute. But this is effectively the ability for the agent to use tools to accomplish tasks. So things like searching the web, executing and building Python code, accessing databases, whether they’re proprietary to you and internal within your organization or whether they’re external databases that are publicly available, and then interacting with APIs and other software tools. So this is really important, and I’ll I’ll come back to this in a second.

We also have profile. This defines the agent’s behavior, personality, and specific capabilities.

This is really important when, Tom and Lucy introduce themselves, both both ex retailers.

They have a huge amount, as many of you will have on this webinar, of subject matter expertise that’s been built up over many, many years, and you have IP inside your heads of how we tell businesses operate, the operating logic, the the things that, you know, you need to do to optimize your business and your margins or not.

Giving that knowledge to AI agents, is really critical and really powerful. That specific domain expertise can effectively enable it to start to emulate some of the thought patterns and capabilities that, we have, as experts in our field.

And agents are capable of of taking that knowledge and working with it. So things like your merchandising strategy, you know, stuff like that, would be really important to to ground, an agent in that knowledge. And then you’ve got memory, super important, say same as a human being. You know, my memory isn’t perfect all the time, but this this short term memory and immediate recall of things that have happened recently and then the capability to remember things over a much longer period of time.

It’s critical as AI agents are deployed within our business. They have the ability to recall past interactions, past, kind of, runs when when they’ve tried to optimize for a certain thing in the business. The ability to for it to remember what it’s done and learn from those experience and improve over time, is is really important and another core capability of the agent. I wanna call about really quickly, two things here.

When when we think about AI agents, this is my personal opinion. I think over the next two to three years, we’re gonna see an absolute explosion of the deployment of AI agents across effectively every business function, And it will probably start in some of the more administrative functions like finance and HR, but it will certainly, we’ll start to see agents, and Tom’s already referenced a few, be deployed in critical profit, and revenue driving areas of our business, like the merchandising function, the buying function, you you know, places like that.

And the things that will differentiate your agent within your business versus some of your competitors will be tool use and profile from my perspective. Everything else will become broadly commoditized, and we’ll all have access to the same capability.

But it will be the tools that you provision your agents, that will give them distinctive capabilities to achieve an an objective, and it will be the profile, the context, the subject matter expertise that you ground that agent’s knowledge in. They will be the defining factors that make an agent incredible and will help optimize vast swaths of your business or that make it a bit average and generic and doesn’t quite achieve what your your competitor’s agent is achieving. So I just wanted to call them out when when you’re thinking and having conversations about this over the coming months, years, really think about what tools are we giving the agent and how do we get the context, the expertise grounded in the agent to help it become really performing.

Cool.

So if we if we leap over to pricing and, the core purpose of this webinar and anchor everything we’ve discussed around that, what we’ve got on the screen here is a solution architecture for Peaks Pricing AI markdown application.

And this is heavily, heavily, centered around deep machine learning capabilities. So on the left hand side, we’ve got typical data inputs that we would grab from a retailer, in order to deploy this type of application. So things like customer orders and transactions, inventory, product data, location data, website data, promo calendar data, all really critical. We then pass that through to the application, and and the application then does a couple of things in this box in the middle.

It deploys a a demand forecast, the base demand forecast that predicts the volume of sales per product per week at a given price, typically the the price that it currently sits at. And then we deploy a price elasticity capability that predicts the changes in demand resulting from changes in price. So price elasticities, Tom, wrote a brilliant blog on this, which we can probably link out to after this after this webinar. But price elasticity is a really important, economic, theory that’s been around for a while.

And it as it says in the box, if if you know how much demand will be, created for a product if you change it, in price reduce it in price by one pound, one pound fifty, ten pound, fifty pound, then you can start to do amazing things with figuring out how do I liquidate my stock in the most effective way, how do I achieve the net margin targets I need to do with my current inventory holding, and you can create a an amazing lever that can help control your inventory pool.

And this is all around just applying using machine learning capability, a price elasticity score to every SKU by location combination within your within your retail network. So some products will be highly elastic, and that basically means if we apply a small markdown, maybe a couple of quid, it will create a huge uplift in demand. And Pete can help you predict precisely how much demand will be created down to a a a unit level, which, again, gives you an amazing lever. If something’s really inelastic, then you need to hit it with a massive discount in order to move the needle on demand.

So it it gives you a lot of visibility on what we’re gonna have to do here with our with our markdown strategy to achieve our targets. The demand forecast and the price elasticity then, both push through, what we call an optimizer model, which basically takes all of the business guardrails and business logic, like product versioning, exit prices, price rounding, product versioning, markdown targets, markdown budget, things like that. And it seeks to optimize, towards specific objectives, typically a margin objective and a sell through objective. So we’ll try and find the optimal trade off between achieving the most margin possible and the fastest sell through rate, in line with the target that you’ve set.

Those capabilities are then kicked out via an API, typically. Lot lots of our customers use it in a headless capacity, which basically just means you don’t see any user interface. You just choose you just see the output passed into your your tool of choice within your business, or you can use a a user interface, which I think we’re gonna show off today, where you can interact directly with the with the predictive output.

Next slide, please, Tom.

Yep. So when when we think about, Agentic, what we’ve just shown on the previous slide is a machine learning capability that interacts with some generative stuff actually, which we’ll talk about in a second.

But, when we look at the coming months of deployment, we think we’re gonna start to be deploying agentic workflows as opposed to just purely, machine learning and generative workflows. And what that looks like, we’re gonna visualize on the screen here for a a fully agentic workflow for markdowns and promotions. So what would typically happen in your business, that there will be a flag such as a product hits a rate of sale threshold, and it will trigger an AI agent, which, you know, could be Peaks markdown or promotions agent. And that agent will be instructed with a key objective to profit profitably optimize inventory holding at all key locations.

That agent, as we mentioned before, will then break that down, that objective down into a bunch of subtasks. So what, how do I best understand my SKU by location rate of sale? How do I factor for age of stock seasonality?

How do I figure out the optimal discounts, to quantify for, optimizing my my working capital, my margin, and so on?

And the agent will, then choose the most appropriate tool. So if we click on Tom to achieve each of those sub subtasks oh, sorry. And it it will also be grounded in all of the contacts we’ve we’ve given it. So things that I mentioned earlier, like the merchandising strategy, your current Wizzy, budget planning, category specific rules, other business rules, floor price, and minimum discount.

And then using that knowledge and using the objective and the subtasks it’s broken down, it will then select an array of tools in order to, achieve the goals that that we’ve asked it to achieve. So things like UiPaths bots, will go out around the business and grab the prerequired data, the the, things like store and AUR margin targets, the marketing calendar, all of the data we mentioned previously on the screen, transaction data, things like that, live sales data. It will trigger Peaks pricing API to forecast the forward demand on a SKU by location data and to simulate the, the price elasticities and then optimize accordingly.

And it might also use external marketing, market data, via APIs available. So it might take a look at market trends. It might look at competitor data. It might look at hyperlocal events that are occurring that we need to factor in because that will create a spike in volatility and demand at one of our stores.

So it uses all of that, and then the output is ultimately an optimized markdown and promotion, strategy at an individual SKU by location level where we’re seeking to optimize margin versus revenue trade off, and distribute all of those decisions downstream into into the critical, business systems of action where we can then execute. And the typical value output that we drive, and we work with dozens and dozens of retailers at this point, with, the pricing capability, three hundred to five hundred basis points in that trading margin improvement, which, Tom, I know from having chats with you, that there is just nothing out there, you know, in the market that achieves these types of uplifts in terms of EBIT margin, net trading margin.

So it kind of seismic in in its impact for a business.

Increased revenue and then, obviously, thousands of planning hours saved, normally spent in Excel where our teams are are liberated to focus on some of the the more strategic, creative, and difficult tasks that we’re set with within the the merchandising or or, kind of planning functions within our business. And then critically, all of this capability, all of the different, bots, different tasks that are being automated, different machine learning, models that are being deployed are all being orchestrated, by something like UiPath, who, now have an incredible capability called Maestro that enables humans, bots, and AI agents to be orchestrated seamlessly in really secure environments, with with what, we’re describing as controlled agencies.

So the ability to, really get under the hood of how all this stuff connects together in in a sophisticated way.

So we’re really excited about this capability.

Tom, I think I’ll hand back to you to take this onto the next bit.

Yeah. Great. Thank you. It’s pretty mind blowing.

And if people watching or anything like I probably still am, but definitely was. Like, it’s still I’m I’m getting it. It’s making these these graphics are quite helpful.

But we wanted to just show a little snippet of actually how it manifests on peak, essentially.

We’ve got a few minutes left on the call today, but, Lucy, you’re gonna take us just through this. So over to you.

Great. Thanks, Tom. And it’s very much if you remember the previous slide around that pink box and the decision in, and a slide prior to that, Chris talked about the architecture, and the core of pricing AI being that demand forecast and price elasticity. But at a really high level, pricing AI is designed to make more informed pricing decisions, finding recommendations that balance that revenue and margin and the business goals towards profit, all without kind of spending weeks and weeks in spreadsheets, which I think we’ve all been guilty of.

In the product today, you can very much interrogate and compare your different scenarios, and dig deep to make those best decisions, but it’s still very kind of manual in terms of comparing all of those different scenarios of each other. But I guess the future, and with Agenntic AI, we can take this one step further and start to reduce that human reliance on some of those elements, helping us to solve problems faster, and spend more time in the day to day rather than, the the kind of the general operations.

In terms of our current capabilities, we’ve got our analytics agent, which you can see on the screen now. So this is all about converting your plain English LLM, plain English into that LLM language that Chris was talking about before. It then takes the next step to figure out what what’s the best data tables to extract the information from. It then takes another next step to decide what format should be the best to present that output in, whether that be a graph, a table, or something else.

All of this might sound, like, pretty simple, and humans can do this and have done. But to do scale and to unlock that insight that we have never seen before, it takes a lot of time. So, yeah, the agentic systems will help to unlock all of that power that we’ve not seen before.

And I guess after that, there’s the possibility to action findings and take some next steps, from that. So in, the demo, we kind of saw that we’re asking questions around, flagging some high stock SKUs, high elastic products. But, actually, what do we want to do with that? Do we want to increase, the price, decrease the price?

Agenstic Systems can be built to help us action some of that and actually execute it as well.

But, yeah, jumping on to kind of what’s next in this space. We want to think about other agentic use cases, in terms of the jobs to be done and all those processes that retailers, have on a day to day basis. So in respect to, pricing decisions, one that comes up over and over again is scenario planning. So this is hugely complex and time consuming.

And at peak, we’ve been able to inject AI into how retailers conduct their scenario planning.

But very much using agentic AI, we can accelerate this, and speed that up to get better decisions faster by being able to, I guess, simulate hundreds more than you would have done on a day to day basis.

It’s all about the agents supercharging that process.

Yeah. And I guess another observation in pricing in retail is that at the moment, we are very much reactive. AgenTiCare could take us to that next level, and help us be more proactive, be smarter with price setting earlier in that product life cycle, and ultimately generating, like, more gross margin and other business goals that we’re, working towards as well.

Cool.

I think, we’re coming up on time actually, so we might have to keep this bit quick. But, like, I wanted to just throw it to you, Chris, to talk about some of the potential barriers. Maybe you could summarize some your thoughts on, like, why why you’re maybe not getting started here.

Yeah. I I think I’ll summarize it in three really quick, bullet points here. I think one one of the main limiting factors when we think about everything we’ve talked about today, You need to have a really good Internet understanding of your business processes in order to give an agent, the context in which it’s operating in.

And, I mean, I love getting in front of a whiteboard and mapping out business processes, guardrails, business logic, kind of operating procedures, things like that, but not everyone does, and it is time consuming.

I think the businesses that we’ve seen starting to get at this really quickly and drive value with AI agents and some of their key, kind of profit generating functions are the ones that had already done the groundwork in understanding exactly how their business operates and functions today so so that the agent understands, when when you’re, you you know, giving it an objective, the parameters in which it’s it’s trying to operate within and and optimize, for. So I think that’s really important. The the second point is really closely related, which is if you’ve got your business processes, you’ve then got all of your systems, that set underneath those processes, and you’ve got the data that is coursing throughout those systems and processes.

And I think you need to understand where data lives. You need to understand the limitations of some of the data capture within your business, and you need to start to plan for how you will capture some of the structured and unstructured data, that’s flowing throughout your business in order to in order to, enable AI agents to get their hands on on stuff that that will help it. Things that we come across or chat about all the time, Tom and Lucy, is probably, like, promo calendars, where are they stored, like, how frequently they’re kept updated, things like supplier catalogs, who are our suppliers, what what do they actually have available, those types of things that are really, really valuable structured and unstructured datasets that an AI agent could use to contextualize, what it’s trying to do, are really important.

The final thing, which I guess is a bit of a theme of this series of webinars that we’re running, is around the leadership team themselves and their willingness and desire, their innate curiosity to educate themselves at a deeper level, around these technology capabilities because I think they they are gonna drive seismic changes in organizational design, in our ability to eke out more and more margin, from from the tech capabilities we’re deploying.

And I think every boardroom owes it to itself to to start to to understand these capabilities at a deeper level. To your point right at the start, Tom, everyone is mega busy, and it’s harder than ever probably to get your hands on some decent, grasp of the all of the changes going on within AI and, technology at the moment because it’s it’s moving at such a frenetic pace. But it’s so important, I reckon anyway, that everyone’s trying to just understand what’s going on and figure out how it’s gonna apply to their business. Otherwise, as unfortunately, you you know, a few businesses are gonna get left behind.

So so those three things, understand your your business processes, understand the data flowing throughout them. AI agents are gonna need that. And then as leaders, making the big capital allocation decisions, making the big calls for for the future of your your businesses, just try trying to educate yourself, you you know, on on what’s going on in this space and and trying to get deeper and deeper into what these capabilities mean going forward.

Yeah. Nice. Thank you. I think, we, we we see that yeah. The the pink box as we referred to before, that’s the machine learning bit that that’s that exists today. It’s the wider Regentic stuff here, isn’t it?

Lucy, we added just a few couple of closing thoughts here just on the how to so if we’ve if if we’ve removed the barriers, like, how do we successfully deploy this stuff?

Yeah. Very much. And I think speaking to, like, our experts on what we’ve observed in this space, these are the things that have been or the themes that have been, sticking out to us. So simplicity and focus is key. With this, it’s all about starting small and identifying where you can have the biggest impact with the lowest effort.

Collaborate and then automate. So this is very much about taking your teams on that journey and ensuring that they still have a level of autonomy.

Fully agentic decisioning won’t happen overnight, and this is where that human in the loop process, will win and help us to get where we want to get to. And we need teams to feel empowered by agentic AI, not re not replace them.

And pilot and iterate fast.

It’s a core product principle.

It’s not just about testing that technical feasibility, but it’s also about test the appetite for adoption within a business. There’ll be some decisions and processes that business teams are more likely to want to keep close to and others that they’re a bit more happy to relinquish that control on. So it’s about finding the balance there as well. And then the last one, keep learning. So encouraging experimentation and allow teams to organically identify those opportunities.

Those are the ones that are gonna be more successful, in the long run. And I guess to summarize all this, it’s a bit of a journey.

Just because we can use agents for a decision doesn’t mean that we should be smart with where and how will be key to the success, of this new technology.

Totally.

Right. We busted through the time that we’d advertised, so thanks for sticking with us, everyone online. There’s been a couple of questions come in, which we will have to follow-up with now.

Some good ones, like, some slightly cynical ones, which I quite like as well. I think, ultimately, this is happening, you know, this is happening right now. The some of the agentic stuff is, like, the next step, but the machine learning applications for this is is exists today.

Hopefully, this has been a sort of, you know, reasonably nourishing and useful time spent with us.

Please, like, keep an eye on our our various channels. We’ve got a couple of other resources there which we’ll link out, and we’ll be sharing the video afterwards so you’ll have that as a reference.

And we might even have a a little bonus, a little, extra bit of content that comes with with that.

Myself, Chris, Lucy, at your disposal for any other follow-up questions, please feel free to message us directly.

Maybe, Chris, when you’re not on holiday because you stepped in very nobly for Mark who is now on paternity leave, which so thank you for for helping us out today.

But, yeah, hopefully, this was, good fun, everyone, and we’ll see you somewhere soon. Thank you.

Thanks.

Stay in touch!

Subscribe to our newsletter to find out what’s going on at Peak