Albert Heijn’s Sjoerd Holleman on AI Shopping Assistants and the Future of Grocery CX | Groceryshop 2026

Hello, everyone.

This is Omnitalk Retail.

I am Ben Miller, host of grocertalk, and I'm coming to you live from Las Vegas from the Vuzion podcast studio.

That's booth 913 at grocery shop.

And joining me now is Sjoad Hollerman, Chief Product AI and specialty Brands Officer at Albert Hein.

I'm sure most people listening today will know the Aholdales business, but they might not know the Dutch brand specifically.

So just give us a quick introduction to the division for which you work.

Yeah, yeah, I work for Albert Hein, which is one of the great local brands of auto Hess al clearly global retailer, being in Europe, in many countries and in the east coast of the US Actually work for Albert Hein, member of the board there, responsible for strategy and deployment of technology and AI and as such, visiting Grocery Shop and having a keynote.

Perfect, perfect.

So, yeah, that's right.

You were on stage yesterday for an interview.

And your interview is all about how AI is increasingly helping you shape shopping experiences.

So look incredibly relevant to the conversations we've been having.

Break this down for me.

What are the key things that you've done at Albert Hein over the last couple of years to use AI to help your shoppers?

Yeah, I think what we pride ourselves in is that we see ourselves as a food tech company at Albert Hein.

So we're not just grocers, we're food tech.

So we have over a thousand internal capability, people that work in software engineering, product management, data science, machine learning engineers.

And so we doubled down on AI long time ago already to make sure that we are ahead of the curve, helping our customers.

Not just putting technology for the sake of technology, but really making a difference in the lives of our customers.

Okay, so let's unpick some of that because I really want to go into some examples.

Let's know if you were one of the.

You were one of the first global grocers to launch an AI shopping assistant called Stein.

Yeah, that's right.

So because you've had this for longer than other people, what have you learned?

What functionality are people actually using on the assistant?

I think it's a very good question.

So when we launched Stein already 1 and a half years ago, so we were sort of ahead of the curve.

Yeah, but we really started with that what's for dinner question that everybody, everybody know.

Well, you know, I know what every is more of a puzzle every time.

Yeah.

So one of the things I said is when I was young, my parents just put the.

The food on the table, you just eat whatever you served.

Yeah.

And nowadays it's more complex.

Right.

So, so people wanting more protein, some people might be allergic to gluten, some people want to be only 20 minutes in the kitchen.

So you just could ask Stein, give me, help me, help me out with this problem.

And that's what for dinner.

And Stein will give you all sorts of recipe alternatives, put you on your grocery list, and really help you through that difficult decision that everybody has every night.

Okay, so I'm going to ask you the key question.

Insho, you've had this functionality for a year and a half.

Lots of people have got tech saying that this is what it can do.

Are you seeing customers actually using that functionality?

Yeah, I mean, what we see is that we have over 1.1 million active users, so really, really driving change in shopping behavior.

What we also see is, I thought, okay, is it really going to make a difference for all our shopper types?

And we actually see 41% of all users being families with children.

So it's not a Gen Z only feature.

And we actually see 13% of users being elderly over 65.

So it's really mainstream usage, which is, I think is something to be very wary about, that it's not just a nice shiny toy, it's for everyone to use.

Okay, and have you seen an impact on basket size?

Do people who shop like this, Is it smaller baskets?

Because it's very specific use cases, are they adding into the full meals?

What sort of behavior are you seeing?

Yeah, what we see is that Stein users actually spend 56% more than your average customer.

Okay.

In all fairness, our primary customers were early adopters.

That's what I was going to ask.

Let's gonna eliminate that.

But at least we see that it's really, when they're early adopters, they're really buying into it.

And we see changing in behavior.

We also see questions being asked like, how do I prepare a pumpkin?

What are the best recipes to use with eggplant or gourgettes or whatever the people using.

So we could actually help them also with their cooking experience, not only with their shopping experience, which made it a richer environment.

Okay.

I'd like to understand how they fit into the broader ecosystem.

So for listeners who aren't as familiar, you know, Albert Hein, market lead in the Netherlands, who are incredibly familiar, you know, loyalty card, high penetration.

How does this AI assistant fit into all the other touch points that Albert Hein, has with his customers?

Yeah, I think that's a good question.

So, so many Years ago, we doubled down on our app first.

So we have 5.5 million active app users in which the AI is embedded now.

So bear in mind 5.5 million is one in three of the Dutch population entirely actively using our app and it really fits in there.

So our app is used by not only e commerce customers, but also customers going into physical stores and being helped there with their company.

So it's a true omnichannel experience.

Okay, I'd love to know more about that.

How is your app helping shoppers in store?

It's really well, it starts at home, so also AI and agent E commerce really starts at home.

Right.

So it's preparing your list, preparing your what's for dinner, what am I going to focus on and your promotions and really helping you along that journey, providing more healthy alternatives.

So it's the discovery phase that's same for every customer whether you go to physical stores or do your e commerce stores.

Okay.

And I think what we're seeing now is most importantly what we're building now, and it's going to go live this year, is that we're not using Stein anymore as a you ask Stein and Stein will help you.

Stein will help you proactively.

So we're moving beyond and I think people are talking about the future here.

We're trying to build that right now is highly personalized meal planning for you, waiting for you in your app, ready on your basket.

Only you have to check out.

So that's one of the features that we're building right now.

Secondly, what we're building is pre filled essential baskets.

So all your essentials we all know every week, every other day you're buying similar stuff that you're just replenishing to say the word as a customer.

Why don't we pre fill that basket and allow you also to speak to Stein by adding stuff, changing your preferences.

I mean, we see examples of people saying, I used to have full fat Greek yogurt, I'm on a diet now.

Changed my historic behavior and you can actually talk to Stein and make those changes on your essential list.

And lastly, we're making it much more conversational on all elements of customer service.

So Stein, can you change my E Commerce order moments to a different moment?

Can you change the store, my favorite store, because I'm going to a different one.

I need to make sure that you know that I'm going to that store and really make it interactive as a conversational agent rather than just only answering questions that you're asking your agent.

So we're moving beyond that and I think one thing to emphasize is all about persistent memory.

So we're building the persistent memory that's underlying all these use cases so that Stein knows what your preferences are.

What have you done historically?

What behavior have you seen, what have you bought with us so far?

And all that information will be persistent memory underlying the app so it will know you more, better and more than you would ever seen before.

Okay, so I think this is taking me into another area that I want to talk about which has been a really big talking point on the show and that is what the grocers need to do to develop their own AI assistant capability versus what should they expect from shoppers who are using, you know, the assistance the chatgpts of the world.

How do you see what shoppers should go outside for versus what they should stay within the Albert Hein ecosystem for?

I think it's a very good question and I think my, my true and honest opinion is that I think we're uniquely positioned to win and lead that race in agentic commerce because we have trusted brand.

I mean we're 140 years old.

Yeah.

We're a household name.

We've built that trust with our customers over time.

We've got a deep customer relationship.

I think that leads to very, very rich first party data.

Not just transactional data like some others might have, but we have really rich, as I just discussed with a persistent memory rich first party data.

So we know you better than anybody will ever know or at least for the short and midterm and in addition to that, I mean we have a strong own brand offering.

So in the Netherlands we have over 50% own brand share sale.

Some incredible products.

Incredible products.

So we're proud of every own brand product.

I think that's something to emphasize.

And also customers are seeing that because they're buying 50% of their shopping there.

And then lastly grocery is about large combined baskets of 30, 40, 50 articles spread over fresh ambient frozen which is really complex.

And then you have the store network and the last mile because we do 12% of our business is home delivery ecom.

So not pick from store but home delivery.

So we have all that infrastructure and that knowledge of the customer and the trusted brand.

I think with our in house capability of truly building AI at scale as we've proven so far, I think we can win this race and be the leading party in asente commerce.

Okay, so is that so what is the kind of customer experience end goal that you think is going to shake out?

Let's say we're having this conversation in five years time.

What's going to be happening now?

Well, five years, I always, I mean two years ago I didn't know we were having this conversation, let alone five years from now.

But I think as I said, I mean it's going to be proactive towards our customers and not reactive.

Don't wait for your customers to ask you something.

Make sure you use everything you know about the customer and everything you know about your products and your organization to serve your customers better proactively.

So that's what I talked about.

Highly personalized meal planning ready for you.

Only you have to check out pre filling the essentials, making sure that you help your customers in all sorts of ways and really be there for your customer rather than, and be proactive rather than have them interact with you by their own initiative.

Okay, Sched.

How proprietary is this to Albert Hein and the Dutch market?

How transferable is it across the other market and the other banners that hold al operates?

Yeah, I think that's a, that's a very good question.

What we've built is if we, you look at the Netherlands, what we've built is we actually build a foundation.

So we wanted the one secure sort of governed foundation under anything we do with AI.

So it's not individual use cases, individual opportunities.

So we built basically a platform that's AI agnostic.

So basically whatever LLM people want to talk about, who's going to win that race, I couldn't care less as long as my platform and my foundation is agnostic for whichever one I can use.

So again, I don't know who's four weeks from now what's going to be the news flow.

I just know that I'm flexible, adaptable and agnostic of who wins.

And then the platform is built on top of that.

Having all the security in the platform, having the finops in the platform, the load balancing, anything you would do with normal software engineering to make it really scalable and super robust.

On top of that we have sort of reusable building blocks that our engineers, our in house engineers can use so they don't have to repeat and rework that every single time.

So their speed to market is really fast.

And on top of that we're building all the Agentix solutions.

But I think what's most important is, and that's pretty unique also for our company because we're great local brands, we're scaling this now and making the platform available to all European brands and all U.S. brands.

Right.

And we're actually going to co develop together from now on.

So it's going to be across the whole alto has group.

So this is going to be before the end of the year.

That's going to go live.

So that should help accelerate our position there.

Okay, so more news by the end of the year.

Really interesting.

Fascinating.

Before we let you go, I just want to look internally and I know Arro Delez has been developing AI assistance for its store associates.

So what do they do and what impact is it having on their day to day work?

Yeah, I think that's a great example.

I'd love to share some of that.

So at Albert Hein we have a very, very centralized model where we have central replenishment and we have a lot of, lot of people in our stores.

We have over 80,000 people that are often youngsters that do this as a part time job.

Right.

So having many youngsters in our teams that we, that we see as proud colleagues, but they're a high turnover of people going out to study and etc.

So how can we help them and empower them to help our customers better?

So I give an example.

My daughter used to work in a store.

She's 18 now.

She went off to study, but when she was 16, 17.

And I said to her on Saturday, you come back from store.

So how did it go?

She goes, I got so many customer questions.

I walked up to the lady working in the bread department every time to get the answer and then go back to the customer.

Okay.

So what we did was we built an AI assistant for our store employees to really help them answer questions for customers in their own app on their phone.

And so they can just type in the questions, get the answer, help the customers.

And the customers were actually really happy because they got straight away the answer.

So we're happy customers.

Yeah, the employees found that and the colleagues found that they're really empowered and they feel like they can help the customers rather than be out of their depth sometimes.

And then again for us it's productivity gains.

Right.

So they can focus on their work in the meanwhile.

So we were really happy with what we did there.

But also we always start with the real problem that we're trying to solve.

We never start with technology.

So the problem we try to solve is how do you help your colleagues to serve our customers better?

And we see some of the bottlenecks there.

We actually, funnily enough, we actually made a dummy AI and sit in the back room for three weeks in three stores whilst learning what they were asking so we told them it's AI, but it wasn't.

And we just learned what they're asking and that's how we can prioritize based on value to see where are the main use cases, how can we help them best, etc.

Without building something first and then hoping that it will work.

So it's a sort of division that we have.

Super interesting.

We could talk all day, but we need to let you go.

Thank you so much for joining us.

And thanks for the teams of Fusion and Grocery Shop for making our coverage of this conference possible.

And stay tuned for more.