AI best practice: lessons from Picnic, the fastest-growing online supermarket

08 januari 2024 • 5 min read

Notebook, titled field notes

2023 was the year of AI experimentation, and with most organisations on the same boat, it was fairly low stakes. Now in 2024, it’s top of the agenda and will be a pivotal marker in leading business strategies. 


So with every business aiming for greater productivity and faster time to market with AI, how do you stay competitive? According to Daniel Gebler, CTO and AI pioneer at Picnic (the world’s fastest-growing online supermarket), “it’s not about who jumps in fastest and pulls out the biggest idea. Like most new year’s resolutions, lasting success is about consistency, holding yourself to high standards and remembering why you started.”


Alongside AND’s GenAI Strategy Lead Leena Pankhania, Daniel shares the best-practice AI approach that’s secured Picnic as industry pioneers, and the standards he’d apply to ensure success in your next AI project.


Get your AI priorities straight


Don’t create a solution in search of a problem

The best designs are the ones that effectively solve a real problem for their users. Effectively incorporating AI into your business will depend on correctly identifying those problems, and aligning your initiatives accordingly. 

For Picnic, there were a few problems that presented an opportunity for AI. Fraud at checkout, extended drop off times due to an inconsistent supply chain, and unnecessary waste (be that in time or products) as a result of suboptimal warehouse stocking. AI and machine learning (ML) initiatives were able to significantly improve these issues; cutting costs and improving customer satisfaction. Set yourself up for success by making sure your initiatives are solving an existing problem, and that the outcome improves brand experience for your customers or team.  

Put your mission first and let your data support 

We all love collecting data, and we’re all trying to understand and maximise its value. However, it’s good to make peace with the fact that you will simply never have enough data. Instead, look at the entire value chain of your organisation and pinpoint areas where AI can bring the most value, either to your customers or your team. Prioritise the data that can power those high-value projects, considering both short-term gains and long-term transformational approaches.  

Data science has to come before AI 

If you’re building AI and ML prototypes, looking at your data science first will ensure you’re going in mission-first, aligning your AI initiatives with specific challenges and tangible business objectives.  
A good example of this in action is at Picnic, where the team sought to solve safety issues in its large fleet of vehicles. With a solid data foundation, Picnic built a tool that could identify anomalies in driving, and feed back information and improvement points to drivers. Ultimately, these improvements in safety and service levels will contribute to their customer experience.

Focus on your launch first, then scale later 

The world of AI is fast-moving, and competition won’t wait around for you to design for what you want five years from now. Start with your MVP, and use it to gather feedback and learnings to take forward. Don't try to come up with ways to scale up that version to a million from the start, because most likely you’ll require another 10 iteration rounds before launch.



Evolving technology means lifelong learning

Continuous and experiential learning

In the swiftly evolving landscape of AI, ongoing learning and adaptability are imperative. Whether you’re a leader or an individual contributor, it’s essential you invest your time in continuous education to navigate the impact on roles and processes. Given AI's rapid evolution, hands-on experience with algorithms, tools, and architectures is crucial for skill development; and building a team culture that enables this learning will ensure agility and innovation. 


Great products come from small teams

The phrase “too many cooks” comes to mind when trying to foster a fast-moving, innovative culture. Instead of throwing all of your talent into one idea and losing momentum to the difficulties of managing it, build small, focused, cross-functional Agile teams that are dedicated to specific projects and experiments. Combined with strong engineering skills, you will then create multiple innovation hotspots within your organisation.  

Keep focusing on closing the digital skills gap

As technology advances, the digital skills gap will continue to grow; even in groups that appeared proficient pre-AI. Identifying and developing future-ready skills is not just about navigating the technical side of AI and ML; it’s about the skills that power big ideas. Problem solving, leadership, and creative ideation will all be key to success in an AI future - make sure your teams are ready by equipping them today.


Innovate, but prioritise the ethics of AI

Ethical considerations are key

As much as AI presents exciting opportunities; it also comes with risks of bias, privacy concerns and regulatory oversight. It’s critical that you prioritise ethical use over innovation, by developing, adhering to and regularly evaluating clear guidelines. Learn more about building an ethics framework for your business. 


The human touch is invaluable

Even in highly digitised environments, adding a human touch is appreciated by consumers. Consider AI as a tool to enhance human capabilities, rather than replace them entirely. Daniel explained that at Picnic, they learned that 80% of their tasks can be covered or improved with AI and ML. However, the remaining 20% is an opportunity to differentiate themselves as a company through their human aspects.

Societal impact matters

Recognise the societal impact of AI deployment and aim to contribute positively. This includes considering the effects on the workforce, addressing skill gaps, and promoting responsible AI practices. For example, the starting point of Picnic’s business was to rethink the logistics process. Picnic owns the entire supply chain, from food production to delivery. The technology around this process includes the mobile app, obviously. However, 80% of the technology in the process is about logistics, such as the fulfilment in the warehouse, order picking, planning, and delivery. Many of these tasks are tedious, repetitive tasks that require lower skill sets, AI can help to automate or improve this type of tasks.

Learn more about AI best practice from the experts at Picnic and AND Digital

This is just a snippet of the conversation on AI best practice. You can tune in to the full webinar, on-demand now. 

AND Digital are solving problems with AI across every sector; what problems could we help you solve?  Learn more about our AI team and explore our work


Related Posts