As the hospitality sector continues to wrestle with rising costs and shifting customer behaviour, accurate forecasting has become more important than ever.  

On HRC's TechX stage, a panel of industry leaders discussed how AI and data are transforming forecasting and, in turn, profit margins across restaurants, QSRs, and the wider foodservice supply chain. 

Moderated by Peter Martin, Co-Founder & Executive Director of Peach 20/20, the session featured insights from Brendan Kelly, Head of Projects at JKS Restaurants, Erika Biggadike, Director & Head of Supply Chain at 4C Associates, and Tim Doubleday, Group CFO of Burger King UK. 

Labour and Sales: The Core of Hospitality Forecasting 

Brendan kicked off the conversation by pinpointing labour as a key area where forecasting can deliver value. “Labour costs are one of the biggest costs we’ve got that we can control,” he said. “Being able to understand what demand is going to be and to look at that a few weeks ahead is really important. Every pound is being checked and double-checked.” 

Despite a wealth of labour management software, Brendan noted that much forecasting remains manual. “A lot of it is still done with Excel sheets. I think AI is really going to fundamentally change what we can see in terms of forecasting ahead for labour costs.” 

Tim echoed those concerns, citing Burger King UK’s £100m+ labour cost base. “Every penny, every hour counts,” he said. Their approach is built around classifying products as complex or non-complex, mapping the time it takes to produce them, and then building labour forecasts accordingly. 

Yet forecasting sales remains the primary input, with labour and procurement flowing from there. “Sales is the key,” said Tim. “If you get the sales forecast right, your labour and procurement planning will follow.” 

But in an evolving multi-channel environment, sales forecasting has become increasingly difficult. “Home delivery has probably gone up 400% in the last five years,” Tim said. “That sort of thing throws all the historical data out. The tools are good, but it’s the quality and relevance of the data that really makes the difference.” 

Supply Chain Forecasting in Focus 

Erika brought a cross-sector perspective to the discussion. As Head of Supply Chain at 4C Associates, she works across various industries and has been using AI in forecasting for nearly two decades. For her, effective forecasting starts with clarity on the business problem. 

“Labour and sales are one thing, but we could also go into commodity pricing, promotions etc., it’s about collating all of that as data and understanding what outcome you’re trying to change,” she said. “You need to understand what you’re trying to predict and what the drivers of that prediction might be, and that tells you what data you actually need to collect.” 

AI as Advisor, Not Oracle 

While AI plays a growing role in data processing and trend identification, human judgement remains critical. “Ultimately, [the tools] just give you guidance. You still need to layer on personal judgement,” Tim noted.  

At Burger King UK, restaurant managers often outperform automated forecasts once they have context. “They’ll say, ‘Actually, I think this is what will happen,’ and nuance that.” 

Brendan agreed: “Our GMs know their clientele, they know the time of year, they know the weather. They can say, ‘The sun’s out, we’re going to be rammed, I’m getting extra bodies in.’ They have the flexibility to adapt.” 

For operators like JKS, the variety of venues, from Michelin-starred establishments with months-long waiting lists to casual dining pubs, adds another layer of complexity. “You can’t take what you do for a Michelin-starred restaurant and apply it to a QSR,” said Brendan. “We need to spend much more time on the venues where demand fluctuates.” 

Smarter Site Forecasting 

Forecasting plays a vital role in site development. Burger King UK is set to open 30 restaurants this year, and Tim highlighted how tech has improved their forecasting accuracy for new openings (from 70% to around 90%) by analysing demographics, traffic cams (particularly for new drive-ins), and even mobile phone data. 

“You have to use tech and AI to analyse that data,” he said. But it must be refreshed regularly. “Otherwise, data that’s 12 months out of date is worthless.” 

JKS Restaurants faces different challenges when opening new sites, particularly internationally. “A lot of our brands are very different,” Brendan explained. “We spend a lot of time building out a forecast: projecting sales, estimating labour requirements, and setting target ranges for each brand.” 

But they also focus heavily on the operational phase. “It’s not technically forecasting, but it’s integral to being able to hit the forecast,” he said. “When we open our doors we’re offering a 100% guest experience even if there’s 50% off food. We’re not asking people to come in while we’re still only half ready.” 

Democratising Forecasting: Tools, Mindsets and Starting Small 

While the conversation centred on sophisticated applications of AI, the panel emphasised that forecasting is accessible at all levels of hospitality. 

“It’s still predictive,” said Erika. “It’s about anticipating what will happen, identifying risks, and making sure you’re prepared. It comes back to a business mindset: ‘Where do I want more certainty about the future?’ That’s where you focus.” 

For smaller businesses without access to enterprise-level tech, Erika had practical advice. “Start by identifying your biggest challenge, then think about what influences it. Play around in Excel. You can do so much with Excel and Python, and as your business grows you can look into more advanced tools.” 

For Tim, it’s about finding the right fit: “You can spend a fortune on systems, but tools like Fourth, Harri or Workforce will take you 80% of the way. They work for single-site operators just as well as big chains, just at different price points.” 

Real-Time Flexibility: Forecasting in a Volatile World 

With second-by-second sales data available at the fingertips of many hospitality leaders, the question becomes how often to adjust and tweak teams, rotas and stock.  

Tim explained Burger King UK’s approach: “We review forecasts three times a week. The labour schedule is set a week in advance, but then we reassess on Monday, and again mid-week. The big risk is pulling labour early in the week when sales are down then being caught short on the weekend.” 

Their systems allow real-time visibility. “I can open my phone and see minute-by-minute sales across 300 restaurants,” said Tim. “But it’s not just about having data, it’s about knowing when to act on it.” 

At JKS, the approach is similar, though slightly less formalised. “We do rotas every two weeks, but our GMs check them daily,” said Brendan. “They monitor reservations, look out for disruptions like a tube strike, and adjust staffing accordingly. It’s a near-daily cadence.” 

Forecasting the Future: From Prediction to Prescription 

As the discussion drew to a close, the panel turned their attention to the future. 

Tim spoke about the role AI is already playing in maintenance. “Ice cream machines are a constant problem. When they go down, you lose sales and you pay call-out fees. So we developed BK Buddy, which takes all the equipment manuals, puts them online, and layers in AI and voice recognition. A manager can say, ‘My ice cream machine isn’t working,’ and it guides them through the fix.” 

The result? “Fewer call-outs, fewer machines offline, and more sales. It may seem small, but those little changes add up.” 

Brendan predicted a shift towards unified systems. “Right now, each department has its own tools that don’t always talk to each other. I think we’ll soon see generative AI that gives a centralised view of the whole business. Less duplication, more integration.” 

Erika took the long view: “So far, we’ve mostly talked about predictive analytics or what will happen. The next step is prescriptive: what you should do to influence outcomes.” 

She offered an example: “AI might tell you that the price of a particular ingredient is falling. It could then suggest you keep your product price steady but increase promotion in certain regions to boost margin.” 

These prescriptive systems won’t just forecast, they’ll guide decisions in pricing, procurement and operations. “It’s about showing where you can experiment, and where you’ll get the biggest return.” 

As the hospitality sector continues to evolve, data-driven forecasting powered by AI is moving from a ‘nice-to-have’ to a necessity. Whether you’re running 300 sites or a single restaurant, the future belongs to those who understand not just what will happen, but what to do about it.