From cleaning robots to smart kitchens: How Compass Group India embeds AI across operations

India’s global capability centres (GCCs) are racing to embed AI into software engineering, analytics and product design. But inside office corridors and cafeterias, another layer of automation is quietly taking shape. For Compass Group India — the Indian arm of $46-billion facilities and food services major Compass Group — digital transformation is now as much about cleaning robots and energy dashboards as it is about cloud ERP. 

“We are one of the leading providers of food services and support services, which is nothing but facilities management,” says Manish Mamtani, CIO, Compass Group India. “Globally, we serve around 5.5 billion meals per year. India is one of our top growth markets.” 

India ranks third globally within the group by transaction volumes. The company operates across 1,100–1,200 sites in eight major cities and industrial hubs, employs around 45,000 frontline staff, and works with over 120 GCCs across 300 plus sites, running more than 270 GCC cafés. 

That scale — in a business that runs on single-digit margins — has forced Compass to rethink how AI and robotics can be embedded into physical infrastructure. “Every 0.1% is very, very important,” Mamtani says. “You have to come up with very different operating models.” 

From pilots to production: 250 cleaning robots 

Compass plans to deploy 250 cleaning robots across GCC sites over the next 12–18 months, with 150 targeted this year. Around 50 robots are already in its fleet; 30 are live in production. “Six months back, it was experiments,” Mamtani says. “Now it is not an experiment anymore. These are real operational use cases.” 

The core business case stems from a recurring workforce gap. “There is always a gap of 5% to 15%, depending on the season — especially during festival seasons,” he explains. “Our starting business case was that there is a gap, hence there is an opportunity, hence there is a revenue gap. Why not do it through a mechanism which is much more reliable?” 

Instead of selling robots, Compass offers them “as a service,” taking on the capex burden while optimising deployment. “Between 2 to 2.5 workforce adjustments, we are able to offer a robot at a reasonable operating margin,” he says. “If deployed smartly, we are able to get an immediate ROI.” Internal tests comparing robotic cleaning with the best human output showed results “three to four times better” on measurable quality parameters, he claims. 

Water and chemical usage have also fallen. “It gives better cleaning, lower water consumption and customers are quite happy with it. It literally works.” AI Moves Into Energy and Utilities Automation at Compass goes beyond cleaning. “We connect UPS systems, gensets, and energy meters. We measure water quality, water consumption, gas consumption — all of that is part of our smart facilities initiative,” Mamtani says. 

On the robotics side, optimised charging and energy usage have delivered 5–7% efficiency gains compared with earlier manual equipment. But at a building level, AI-led monitoring is delivering 10–15% energy efficiency improvements in sites where the system is fully implemented. “In smart facilities, we are seeing much better efficiencies,” he says, adding that the ambition is to scale this intelligence layer across our entire GCC locations over time. 

Why GCCs are prioritising physical AI 

For many GCC operators, automation has largely been associated with software, data and process optimisation. But Mamtani argues that talent attraction has shifted the focus toward workplace infrastructure. “There is a war of talent,” he says. “To attract the best talent, they have to ensure that the working environment is very good.” 

Today’s GCC workforce expects seamless digital and physical experiences. According to the company, 91% of food orders across its cafés happen through self-service channels — mobile apps or kiosks. 

“People don’t stand in front of a point of sale,” he says. “At peak, we process around 80–90 orders per second. In a month, we process around 1.2 crore orders.” Employees receive real-time updates on order status — an Amazon-like experience inside office cafeterias. 

Meanwhile, cleaning automation, energy optimisation, air quality controls and CO₂ monitoring operate in the background. “The right temperature, light, oxygen levels — all these things are taken for granted now,” Mamtani says. “That is where players like us come in.” 

A Cloud-First, Multi-Cloud Stack 

Compass India’s digital backbone was built early. “Our journey on cloud started around 11 years back,” Mamtani says. “We didn’t have too much legacy. Whatever we did, we did it on cloud.” Within a year, the company eliminated legacy systems. It was among the first globally to deploy SAP S/4HANA on a public cloud. 

Today, it runs a multi-cloud strategy across AWS, Microsoft and Google Cloud. “Our stack is very, very powerful. It is very cloud native,” he says. “Depending on what is suitable, we pick that cloud environment.” Except for ERP, much of the application stack has been developed with Indian partners. “There is no 100% ERP available for this industry because the industry is small and highly distributed,” he explains. “So we said, why don’t we build our own?” 

On robotics, the model is hybrid — Indian startups where possible, global OEMs for specialised equipment. Compass works closely with startups, often helping them navigate infosec and enterprise compliance requirements demanded by multinational GCC clients. “We don’t follow a conventional way of working,” Mamtani says. 

“We always challenge the conventional way of doing things.” Workforce Impact: Role Upgrade, Not Replacement With 45,000 employees, automation inevitably raises concerns about job losses. Mamtani insists the focus is augmentation. “We are not taking out people because we already have a gap,” he says. “We are trying to fill that gap.” At the same time, roles are evolving. “A person who we used to call a cleaner, now we call them robo operators. Just think about the profile upgrade,” he says. Technicians are being repositioned as IoT specialists — “IoT commandos.” 

Over the next two to three years, he expects roughly 20% of workplace operations — excluding core hospitality touchpoints — to be AI- or robotics-enabled. “We are a hospitality company at the end of the day,” he says. “We exactly know which touchpoints need to remain human.” 
Needless to say, Compass’ 270+ GCC cafés generate rich behavioural data. AI models forecast attendance patterns based on weekdays, holidays, hybrid schedules and festivals. “Our forecasting is about reducing food waste and optimising costs without compromising quality and variety,” Mamtani says.

Food safety systems are also digitised end-to-end — from vendor selection to AI-powered grain sorters that detect foreign materials and halt conveyor belts automatically. “If there is a problem, we exactly know which vendor it came from,” he says. “We are able to take proactive action.” 

For India’s GCC ecosystem, the shift signals that AI adoption is moving beyond dashboards and dev centres. It is embedding itself into kitchens, cleaning routes and utility rooms — where efficiency gains may be less visible, but no less strategic. 

In a cost-sensitive industry running on thin margins, Mamtani says the calculus is simple: “If you follow the conventional way and just put money behind it, the ROI doesn’t come. You have to find new ways to make solutions really work.”