Inside AI - Ep 6: Craving Productivity in AI Applications
In this episode of Inside AI, Sam Bradon is joined by Simon Revelman, Chief Information Officer at Craveable Brands, to explore how AI is reshaping the operations behind Red Rooster, Oporto, Chicken Treat and Chargrill Charlie's.
Simon walks through the commercial pressure driving the work: tight margins and cost-of-living headwinds mean every lever, from loyalty and pricing to wastage and labour, has to be optimised, and AI is what makes tracking those levers fast enough to matter. He details how AI has already reshaped franchisee recruitment, cutting a manual, handwritten process down to a same-day turnaround, and argues the real gain isn't novelty but speed to market.
The conversation turns to what underpins it all: data. Simon is direct: agentic AI is only as trustworthy as the data lake beneath it, and without centralised, consistently defined metrics, neither a person nor an AI agent can be relied on to get the answer right. He also sets out how Craveable Brands prioritises AI use cases, treating them as part of normal project delivery rather than a side experiment, so the benefits land back in the business rather than staying siloed in IT.
Looking ahead, Simon outlines a vision built around self-serve virtual agents and proactive coaching for franchisees, alongside early trials of AI-driven drive-thru ordering, while stressing that the human still makes the final call.
A grounded look at how AI earns its place in a high-volume, low-margin business, one use case at a time.
Runtime [00:29:32]
Our Speakers
Sam Bradon
Sam Bradon is a seasoned professional services leader with deep expertise across sales, delivery, and customer success. With a track record of leading large consulting teams and solving complex client problems, he helps CTO Consulting clients design and deliver innovative, strategic, and impactful business solutions.
Simon Revelman
Simon Revelman, Chief Information Officer at Craveable Brands, is a seasoned technology executive with deep expertise across IT strategy, digital transformation, and enterprise systems. With a track record spanning retail, franchising, and quick-service restaurant operations, he helps Craveable Brands and its network of franchisees design and deliver data-driven, AI-enabled solutions that improve efficiency and drive growth.
-
Sam Bradon: Hi, and welcome to another episode of Inside AI, the podcast where we focus on providing unfiltered, real-world examples of AI in action. I'm your host, Sam Bradon, and I'm delighted to have Simon Revelman with us today, Chief Information Officer at Craveable Brands.
Simon, I'd love to hear a bit more about Craveable Brands and yourself, please.
Simon Revelman: Sure. Craveable Brands is a franchisor that you've probably never heard of, but you've heard of the brands we own: Red Rooster, Oporto, Chicken Treat and Charcoal Charlie's. We operate around 620 restaurants across Australia, selling chicken, chips, salads and all sorts of delicious meals.
Sam Bradon: That's lovely. I should say I'm a big fan of Chargrill Charlie's; it's one of my favourite restaurants. Before we dive in, we should just say that these are very much our own opinions, rather than those of our various organisations.
But to get to the matter at hand, let's start with the business problem you're really using AI to solve, Simon. What's happening in your market? Are you facing headwinds, or what's really going on?
Simon Revelman: Good question. It's a tough market at the moment: cost-of-living pressures are all through the media, rising prices, and pressure on families trying to make ends meet. That has an effect on us, with customers spending less or coming into our restaurants less often. So we need to find ways to maximise our share of wallet, what customers spend with us versus what they spend at our competitors.
Data is obviously very important to understand what's going on, and then being able to use AI to tease out trends helps us make quicker decisions based on that data, so we know where to pivot on product, pricing, promotions or loyalty. There are other levers too, such as wastage, labour and inventory, that can all make a small difference which adds up when you pull them together.
Sam Bradon: That's great. And you mentioned loyalty there. How important is developing customer loyalty, and moving away from just short-term incentives? How do you build that from a long-term perspective?
Simon Revelman: Loyalty is very important to us. We know our loyalty customers come in more often and spend more each visit, so when you multiply the number of loyalty customers we have, it has a big effect on the bottom line.
What we're working through at the moment is determining what's beyond the standard spend-and-earn facility that most brands have. What else are customers looking for that would make them come back more often? They may want to be rewarded with merchandise or events, or recognition like gold, silver and platinum tiers, with freebies as is normal with dollars to spend on different transactions.
Something we're playing with at the moment is using AI to differentiate between those customers, so that we don't have a one-size-fits-all loyalty product, but something tailored to individual needs. That makes it more worthwhile for each of those different cohorts to join up and stay loyal to our brands.
Sam Bradon: And I think the benefit of personalisation from AI is certainly a key thing. I've been in that space for the last 10 years, where rather than just sending outbound marketing, it's very much focused on what you know about the customer, and not just what you can sell to them but how you maintain their loyalty. There's a lot of trying to maximise that lifetime customer value, isn't there?
Simon Revelman: Yeah, there is. And it's not necessarily about finding new ways, but speeding up the process to market to those different cohorts. Whereas we would have had a data person going through and working out those customer segments, then working with marketing to determine how we want to address them, now we can use AI to quickly put together emails on various promotions or creative, whatever it may be, to go out and target those customers with the most relevant offer to attract them back.
Sam Bradon: And certainly from experience, the more personalised that email or outreach can be, the more likely you are to get a positive response.
Simon Revelman: That's 100 per cent right, and it's not redefining anything we haven't done in the past. It's just speeding it up, and we're able to do a lot more individualised promotions because you don't need staff to sit there and come up with them.
Sam Bradon: And you can do that a lot quicker. Obviously, the other interesting thing for you, beyond the end customers who use those brands, is that you're a franchisor, so I'd imagine the franchisee experience is very important as well. How do you make sure that experience is as frictionless as it can be?
Simon Revelman: Sure. Our franchisees are our customers, so it's very important for us to take care of them. Their success is then our success, because the more they make, the more Craveable Brands makes. So what we're doing at the moment is providing a lot of services to make it as simple as possible for them to run their business.
If I look at it from an IT perspective, we've got a time and attendance system that automates rosters for them to check against, to make sure it suits what they're thinking, and it brings in learnings on labour savings. We've saved a couple of per cent there, helping them structure staffing based on historical and future sales. We've also got inventory systems to work out when they need to reorder and how much. At the moment, it's providing metrics to them; in future, we'll look to automate that so they just tick it off and it goes through, with data in our centralised system where we know everything about the franchisee, what assets they have, and so on.
We're looking to build in incident management, contractor management and service calls, just to make it easier for franchisee partners to get the help they need when they require it, without necessarily having to speak to anybody. And on the financial side, we're exploring AI coaching in our reporting. At the moment they get a report and need to go through it themselves, but we want a coach that can tell them, "you're over-indexing on chicken, you're under-indexing on resourcing," so they get pointed straight to the correct area, and then provide recommendations too, so they can improve, get more money to the bottom line, and run their stores more efficiently.
Sam Bradon: Because I'd imagine you're in an industry with lots of data points, and margins are quite narrow, so every little percentage point you can save or increase is really important. And I think the ability to act quickly is probably very important too, isn't it?
Simon Revelman: Acting quickly is very important in our industry. We're open 24/7. We're affected by weather, for example, and by other things that cause people to come in, or to order online instead. Promotions can drive sales up quite quickly. Having a tool that can help coach the franchisees to anticipate that, or make changes based on it, is very important indeed. It's something we're not doing yet, but something we're definitely wanting to start trialling very shortly.
Sam Bradon: I can imagine. And when you start to look at how you're going to trial or develop some of these things, how is AI actually affecting how you work with consultancies and technology vendors? Certainly from our perspective, we see a big change in how we have to operate with organisations because of how AI is affecting consultancies. What are you looking for now when you're engaged with a technology partner or a consultancy?
Simon Revelman: Historically it's always been about relationship, gut feel, past experience, or reference checks from other people who have used that supplier. That's how I've normally chosen suppliers, as well as price and availability and all the normal stuff. Whereas now everybody is trying to sell AI. You've got your big software companies doing it, your consultants doing it, your solution integrators doing it. Everyone's got a KPI to get a certain amount of AI sales they're trying to close, and a lot of people are approaching us and other organisations offering to do X, Y, Z and deliver X, Y, Z uplift, but none of it's necessarily proven.
So what I now look for is skin in the game from those providers: an unpaid proof of concept to test out whether those gains can really be realised, or a reward for realising certain metrics, so we get to try it risk-free, and if it works out, the supplier is rewarded. Or trial periods, free trial periods, so we're not on the hook for 3 years on a licence, but have 3 months, 6 months, maybe 12 months to really test something out first before we commit longer term. I think those are the most important things now, because a lot of people talk the talk but haven't been able to walk the walk yet, because it's just too early.
Sam Bradon: And following on from that, in that classic AI life cycle, where people are experimenting, then piloting, then rolling things out into production, how far along that do you think you are currently?
Simon Revelman: We're in all different areas of that. We have a number of licences in production across the organisation to help people be more efficient in their work, whether that's Copilot to help with email, PowerPoint and Word, or Cursor for our data team, for example, to help them troubleshoot and test, and some other areas where they're using it to be more efficient. We've also got ChatGPT in some areas, and one place we've used that is in a customised build we've done.
Sam Bradon: Okay.
Simon Revelman: Just like a company recruits staff, we recruit franchisees as well, and it's a competitive market. There are a lot of brands out there who all want the good franchisees, so we need to make sure we're putting our best foot forward and moving fast to impress them.
Historically, we did it like old-fashioned staffing recruitment. We'd call up once someone had applied, do an initial phone interview, then move them to a second round, an interview normally by telephone as well, with somebody handwriting down all the notes. Over time we moved to using Zoom or Teams but were still handwriting the notes, and we determined there was a much better way to do this. So we took the built-in functionality of Teams to turn that into a transcription, which removed the need for the recruiter to write anything down. That meant the interviews went longer, because they went into more depth on questions without anyone worrying about having to write it all back out later.
The other advantage was that we could then feed that into our instance of ChatGPT, which goes through and fills out all the interview questions and answers in our system, no matter what order they were given in the interview or whether they were given in different places for the same answer, and gives a sentiment score up or down based on historical interviews we've done. That allows us to move someone to the next stage a lot quicker, because we don't have to double-type everything back into the system, and we don't have to think about it as much. We've also increased consistency in our decision-making, because we're using the historical data, not just for each recruiter, but overall, to make a call on whether they're a good fit to progress.
Sam Bradon: That's a really good example, because it's got two elements to it, hasn't it? One is the efficiency side, but probably the more important thing is how quickly you can move through the process. Anyone who's been involved in sales knows the more quickly you can react to hopefully get someone to say yes, the far greater the likelihood you have of being successful. That sounds like a really powerful use case.
Simon Revelman: Very much so. And as you know from when you've interviewed for jobs, you like to hear back soon as well. So anything we can shrink there helps us look more in control, and like we're going to be a good partner to join with.
Sam Bradon: Definitely. You've touched on this a few times, and we see it with other organisations too: data becomes vitally important in an AI world, or when you're looking to adopt AI. How important is having that single source of truth, or at least a consistent data set, within your organisation as the basis for AI?
Simon Revelman: It's essential, in my opinion. We've got disparate systems with data, and some of them are pulling data into the others, but it would be very confusing for a person, or an AI agent, to go and work out what's going on. So it's important for us to import all of that centrally into a data lake, with agreed, defined metrics, so gross profit is the same in every system, sales are the same, revenue and profit are the same, that sort of thing. Because when it's all uniform like that, it then allows the agent to give accurate answers without having to guess or calculate on its own.
It's also important for other information, like consistency in first name, last name, address, suburb, transaction order ID, value and products ordered, that sort of thing. Having that consistency everywhere means that when it's building a report or analysis, the information it's using is reliable and accurate, and we all know that. If we let it off doing its own thing, you'd have to question every time whether what it was showing you was right or not.
Sam Bradon: I get that. And in your organisation, how have you approached identifying or prioritising what the next potential AI use cases could be? What approach have you typically used?
Simon Revelman: In the past, people would come to me and say, "what can we use AI for next," or "can we use it for this," and I'd have to say, well, I don't really understand your individual processes well enough to be able to determine that. So what we did is bring a group together and provide training on what AI is, the different types of AI, machine learning and agentic and so on, then go through where the good places are for it to work and where it's not so good, and have a brainstorming session where each person in my department identifies where it could work. You know, "we do a lot of manual reconciliation, that sounds like it could be a good use case," or "we come up with creative for TV ads, maybe that's not such a great use case."
Once we work that out and have a list, we go through it jointly and determine priorities, agreeing, "yeah, everyone agrees that's the first place to start, that's got a lot of benefit, so let's look there." That way, everyone's on board together with our limited team, knowing they're in line to get some of the benefit of AI.
Sam Bradon: I'd imagine you'd have to do that almost on a regular basis, whether weekly, monthly or quarterly, because not only is AI technology improving and increasing, but there are different impacts on your business, or different pressures the business may be facing at different times.
Simon Revelman: Yeah, and we don't treat it as something separate. It's just another one of our projects, and we always re-evaluate our project list and determine if the priority is still right for what's happening in the business, or whether we have to move or cut things. I don't want it as a separate stream, because then, with limited resources, you're competing to get things done. It has to fold in with everything else the company's doing, and we make sure we're keeping on top of it all the time.
Sam Bradon: I think that approach makes a massive amount of sense, because it needs to be part of your normal delivery. You also run the risk, if it's almost like a special team on its own, that it becomes like a science experiment sometimes, in that you don't always get the benefits coming back into the business as a whole. So having it as part of your ongoing delivery approach makes perfect sense.
Simon Revelman: The other thing is, in silo, you've got team members, say in IT, who are very keen and are off playing and building things, but there's no buy-in from the business yet, or any visibility of what they're doing or trying to achieve. So it's best to bring everyone under the same umbrella, get agreement, and then go forward, so that by the time you want to move to UAT and production, everyone's been involved all the way up to that point. You don't get that classic scenario of, "we've got a great solution, now we need to find the problem it's fixing."
Sam Bradon: I get that. And how have you approached the rollout or education of your staff, both in how to use the AI tools and in how to feed recommendations back into the organisation?
Simon Revelman: I think I've been a bit slack in this area, to be honest. We let people work it out for themselves. Copilot went out to everybody to use, and there were people who specifically asked for Claude or Gemini, or who we figured must already know how to use it, and gave them that. But now we've gone back and realised we're not necessarily getting good value for money there, and there are licences sitting underutilised.
So, starting in a couple of weeks, we're having lunchtime learning sessions, where you come along and learn what each of those products can do, what prompts to use, what good practice looks like, and what to watch out for and avoid, so people can start using it better to their benefit. Because, like any technology, some people are keen and work it out themselves, and others don't know and are a bit scared by the change, so they don't jump in yet. What we're trying to do is bring everyone up to a certain level, set them free to see how they go, and then go back and repeat that at a higher level.
Sam Bradon: That makes perfect sense. So if you look at where you are at the moment and do a bit of crystal-ball gazing, looking three years into the future, what are your aspirations as an organisation?
Simon Revelman: I think my biggest one is self-serve for everybody: virtual agents that can speak to all the different systems and answer any questions people may have, whether that's someone in a store or at one of the support offices. For example, in a store, we've already built an MVP where you go to the agent and ask, "what temperature does chicken have to be cooked to," "what points are available on a burger through Red Royalty," or "how many hours can a 15-year-old work in a fortnight." You can throw random questions at it and get answers back with links to the documentation, rather than having to trudge through the documentation system, search for the relevant policy or process, and then open it up and read it. It's all very time-consuming, so I want that one-stop shop inside the agent.
And in combination with that, for the right user role, there's a coaching aspect: proactive alerts that your roster is too heavy for next week, or that you're running a promo and your roster is too light, or that you need to order more because there's a promotion coming, or that an item is on run-out and needs clearing. All of that proactively tells the franchisee what needs looking at, to direct them in the right direction and help them be as successful as they can be.
Sam Bradon: So it's almost an agentic coaching model. You're taking the data you've amassed and helping the franchisee make real-time decisions, and potentially even make some decisions on their behalf.
Simon Revelman: I don't know about the "on their behalf" bit yet, but I'm sure we'll get there eventually. They're each running their own small business, so they should get a say in what they do at the end of the day. But having the coach make recommendations, backed by fact-based information, to help them decide whether to go the way they were thinking or the way the coach is suggesting, that's where I see us getting to. There's a lot more work to be done to make sure the information it provides is accurate, because we don't want to be giving any fanciful information, hallucinations. But yeah, that's where we're aiming.
Sam Bradon: And following on from that, if you look at hospitality in general, where could you see AI taking or changing that industry in the future?
Simon Revelman: I think there's opportunity in a few different areas. One is we have self-service kiosks in our stores that people use anonymously. What if the camera, without identifying the individual, could tell that's an older person, a younger person, or a family together, and automatically lay out the menu or promotions based on that customer cohort's purchasing history? So if it's a family, let's put the full chicken, chips and peas meal up; if that looks like a single person, let's show them a burger. Just that sort of thing, to try to get people purchasing what we think they should, to make it more efficient for them. Obviously, they still have choice.
Then for customers who are logged-in loyalty customers we know, whether by kiosk or our online channels, app and website, we could address the menu to their format, with a reorder option for the things they usually get, to make it a lot quicker for them, because people come to a quick-service restaurant (QSR) for convenience, so you want to make that checkout experience as fast as possible.
Then there's the loyalty piece we've already talked about: getting the right message to the right people at the right time. So if I know you come in at six o'clock on Tuesdays, I can get you an email at 5pm on Tuesday with an offer, just to make sure you choose us and not one of our competitors, and use it elsewhere to influence that customer decision.
Sam Bradon: I was reading that one of the targets people have talked about is AI-driven drive-thru technology, but I think it hasn't always been that successful so far, at least from what I understand. Is that something you and your team have looked into as well?
Simon Revelman: Funny you say that, we're testing that in our lab right now with a team from the UK. It's going very well: it can take an order, modify it, change items on the menu, and it upsells, but appropriately. Some of the ones in market that we've seen always try to sell you something you may already have in your basket; this one is aware of what's in the basket. It's already integrated into our point of sale as well, so the order goes through automatically and up onto the kitchen screen for them to start preparing.
It's early days, with a lot more testing to do to make sure it's good, but I see that in the next month or six months we'll go out into drive-thru in parallel with a person, just to see what it does with real people, real accents and real customer queries, and use that information to decide when we go into a pilot and run it on its own.
Sam Bradon: I'm personally intrigued by how well AI tools, and they will get far better, pick up different accents and variations, and even terminology. The terminology people use in the UK compared to Australia is actually very different, isn't it?
Simon Revelman: 100 per cent. Brands saying Sprite instead of lemonade, or Pepsi instead of Coke, that sort of thing, and it needs to know what to do, what to provide, and what to feed back to the customer. So that's all something we've got to keep testing, to make sure that whether you're in Sydney, Mount Isa or Geraldton, it can still understand you and all the quirks people have for different words.
Sam Bradon: I get that. Look, I think it's really interesting times we live in. Simon, I'm conscious we're getting towards the end of our time. Is there anything else I haven't covered that you'd like to talk about, around how you're seeing AI affect Craveable Brands?
Simon Revelman: I think what I see, more generally, is that AI is a partner to people in doing their jobs. I don't see it as the big people-replacer, at least in the short to medium term, but it's important for people to be aware of AI and to use it in their organisation, because that increases productivity, which benefits everyone who works there. The training we've been a bit slack on, which we're now pushing forward, is going to help with that, so people spend more time on value-adding tasks rather than menial ones. I think that's really important.
I think the jobs of the future are all going to go to people who can work with AI. That's going to be one of the questions in interviews now, not just for IT people, but people in general: where have you used it to enhance your abilities in your position, and what benefits did it bring? Regardless of what industry or department you're in, you're going to be affected by it. It's best to jump on board, get working with it, and understand how to make your life easier, because in the future that'll be very important.
Sam Bradon: I totally agree with those sentiments. Simon, thank you very much for your time, it was a very interesting conversation. I'm not fully across what happens in the hospitality space, so it was a really interesting insight into how you're currently using AI, and what you're planning to use in future as well. Thank you very much.
Simon Revelman: Thank you, Sam. It's a pleasure.