Albertsons Is Reinventing Merchandising With AI | Fast Five Shorts

Albertsons is working to scale a new AI powered merchandise intelligence platform built on databricks with a goal of fully deploying it to merchants by the end of 2026.

Chris, get ready for this because I'm going to you first on this one, no doubt.

According to Grocery Dive, the platform is designed to bring pricing, promotions, product and placement decisions into a single space, replacing a patchwork of tools merchants had previously relied on.

It's built on Databricks Lakehouse for clean retail data governed through Databricks Unity Catalog and AI Gateway and topped with Databricks AI Agent Genie which lets merchants query the platform in natural language.

And if you had databricks three times on your bingo card, you win.

Today, Karthik Iyer, Albertson's group VP of merchandising transformation and AI, said the goal is training on years of clean transaction data like pricing trends between different Apple varieties to give merchants forward looking scenario based guidance like for example, how a dry summer might affect ice cream sales and shelf space.

Chris News and what's your prediction here?

Albertson says it's rebuilding how its merchants think using AI.

Is this the kind of AI investment that actually changes on the ground decision making or will it ultimately become another expensive layer that sits atop the same old merchandising processes?

I think this can absolutely change decision making.

Okay, this isn't guaranteed and here's why.

Bringing pricing, promotion, product placement into one environment is absolutely the right move and sounds great.

One other thing I really liked about this article was the fact that they brought in the merchants and store operators to actually help build this versus buying something off the shelf.

Because you know that adoption is going to be a huge challenge here when it comes to how this is utilized in the organization.

The other thing I'd note though is who else was in that room?

There's a lot of other cross functional partners that come into these conversations and these decisions.

Marketing, space planning, category management.

You can name cross functional partners that need to be part of this process and I'm not certain they were.

Hopefully they were.

Plus 10 more too probably that you haven't even named yet, right?

Yeah.

And so here's why I have a bit of concern as to how this actually happens.

Adoption is the first thing that I would call out.

We've all been part of organizations where there's been consolidation efforts with new systems, tools or dashboards were rolled out.

And yet what is the one thing that you probably saw in those same organizations?

There were teams or individuals that continued to use their own spreadsheets, their own processes because they work in very different ways.

That comes down to training and change management.

Like, how does Elliburts ensure that the AI platform is used effectively and the hard work being done to rethink the entire processes around it?

You can't just bolt on a new system to a bunch of existing processes and assume it's going to work.

You really have to put in the hard work to rethink the entire process and embed this in a way in the organization that can seamlessly allow you to make decisions that actually make it to the floor.

Yeah, yeah, I'm.

That's that.

Yeah.

Those are really great points, Chris.

Like, I'm.

I deliberately want to go to you first because I think if I was to ask anyone what they thought about this headline, I'd probably call you and be like, what do you think?

Because I think it's really hard to do this.

Like, I'm actually.

I'm probably even more skeptical of it than you are in.

In terms of how you just couch that, because, One, I think it's really hard to do.

Two, but the.

Here's the thing for me, and this is why I started Omnitalk, too.

The use cases dropped in the press release.

And, Shelly, I think you're gonna.

I'm guessing you're gonna agree with me on this one, because I can see you smiling in the background already.

The use cases dropped in this press release.

How a dry summer might affect ice cream and shelf space.

That is such bs.

I mean, come on.

Ice cream is in freezers, which are the most immovable store fixtures out there.

It's also dsd, so there's, like, very little impact that you're gonna have anyway.

And they can.

You can reallocate the shelf space as.

As simply as possible.

So if those are really the use cases, I think it's doa.

But the one caveat I would say is I don't think those are the real use cases.

Those are not why you're doing this.

The real use cases are the use cases no one wants to talk about.

But, Shelley, what do you think?

I 100% agree.

The use case aside, did not inspire confidence that they are building this for what really, merchants need to know in terms of this.

If you're using AI in merchandising, if it should.

Merchants have.

Traditionally, I'll just say this.

We're always trying to predict the future.

That is our job.

We are trying to predict the future.

So if this technology is not helping you predict the future with big, huge questions that are going to move the needle on your P and L, then you're not building the system in the right way.

So while there were several things that, like, sounded great in this press release, like we're using data bricks and we want clean data and we want to inform our merchants of this, it's absolutely the deployment here that I would say is.

And are they asking even the right questions?

Are they building this so that they're asking the right questions?

At the end of the day, this should raise the floor for every merchant.

But it's going to be the great merchants that use technology like this to raise the ceiling and to really dive deeper into how they can drive their category performance.

And so, for me, this comes down to the fact that the skill set of the merchant might evolve.

If this is actually, you know, technology that's going to be adopted and used by Albertsons, they're still going to need people with exceptional judgment, exceptional curiosity, because those are the folks that are going to outperform.

If we're relying on the same type of talent and skill set to leverage this technology, I think there's a different conversation.

So I think there's a talent conversation alongside an AI conversation.

That makes sense here.

But the press release based on the quotes that came out, it all sounds good until you start talking about what you're actually imagining it to do.

And then that's where I lost confidence.

Yeah, right, right.

Yeah.

It made me think of, like, all those things when AI search was first coming out, like, all the things you could do with AI search on your website that no one's actually going to do.

But that sound cool.

But, Chris, what do you think?

Last word here.

Yeah, one last thing.

Back to your ice cream example.

The other thing that I think this potentially does is actually creates a risk of paralysis, like the types of scenarios that you described and the amount of data that could be coming their way potentially actually slow down decision making.

Yeah, yeah, that got me thinking, too.

Like, Shelly, I'm curious, you know, you being.

You being the former CEO of the group too, like, not all merchants think and learn in the same way, too, or process information the same way.

So how do you think about that from the executive chair in terms of these tools and these technologies?

Because in some ways, it almost feels like we're trying to get all merchants to do their job the same way.

And I don't know that that's the best thing either.

Or maybe AI helps prevent that to some degree, too.

But how do you think about that question before we move on?

So, as A leader.

If I was, if I were leading merchandise teams today, my, my one on ones with my buying team would be just asking them how they're thinking about their business and what questions they're asking about their business.

Because that's how we coach people to be more curious and ask the bigger questions of what's going on.

And so it's actually shaping thinking.

And the more that we do talk about that in large groups and model the way and show that example is how you really drive that transformation.

So for me, I think about it in a way of like, how can I prompt my team to continue to ask bigger questions?

Because they're not going to spend half their day, half their week pulling data together anymore.

They don't have to do that.

They can actually spend that time with tools like these really asking bigger questions about customers and geography and weather and all of these different things that impact their category.

They can spend more time on supplier relationships, they can spend more time looking at products.

And so that's where you really drive the team forward.

It's like, how are you spending your time now and what kind of questions are we asking?

And those questions can get so much bigger and better about customers now if you're not spending your time on, you know, all the manual data gathering.

Right.

The other point it makes me think about too.

And then we'll move on to the third headline is like is, you know it also this, this, this headline also when you dig, dig into it as we have also highlights the divide still between digital merchandising and in store merchandising.

And a lot of what's being talked about here are store merchandising actions.

But in reality those actions are better placed in the digital online world, which is something that Albertsons probably gets to a degree but doesn't understand as well as Amazon and AI's applicability to run everything in that way, to react to the situations, to, to do what needs to be done in terms of inventory placement and reacting in season is so much easier and so much more fluid in the online space.

But bringing that into the real world has some real issues and dependencies.

So.