The recent back-to-school shopping period showed that things can change quickly in retail when a celebrity or creator can suddenly make one product the thing everyone wants overnight. During the summer, Hailey Bieber’s jeans sold out in a matter of days at the Gap.
While TikTok continues to shape which styles younger shoppers are looking for in the first place, for retailers and consumer brands, the challenge is keeping up with demand that can shift much faster than traditional inventory planning.
In an interview with Retail Insider, Bianca Anghelina, Founder and CEO of Aily Labs, talked about how AI can help brands spot these shifts earlier and make faster calls on replenishment while there’s still time to act.
There’s also risk in moving too fast. A viral product can sell out in days, but the demand behind it can fade just as fast, leaving brands unsure whether they’re looking at a lasting trend or a short-lived spike before they commit more inventory.
Question: How are social media and creator-driven trends changing the way retailers forecast demand and plan inventory for events like back-to-school shopping?
Answer: The biggest change is speed.
A product can suddenly become popular because of one creator or one social trend, and within days retailers can see demand move in a completely different direction. Traditional forecasting was never built for that. It relies heavily on historical patterns, but what happened last season may tell you very little about what a consumer is going to want next week.
The opportunity with AI is to connect what is happening externally with what is happening inside the business. Retailers can look at social signals, search behaviour, and consumer interest together with sales, inventory, supply and financial data.
That creates a much more real-time view of demand. Instead of waiting until a product is sold out to realize there was a trend, they can see it developing and start making decisions much earlier.
For me, that is where the real impact starts. You move from reacting to what happened to deciding what to do next while the opportunity is still there.
Q: How can AI help retailers identify an emerging viral trend early enough to adjust inventory and replenishment decisions before products sell out?
A: Retailers already have enormous amounts of data. The challenge is bringing those signals together quickly enough to decide.
Social media might show that a product is suddenly gaining attention. Then you have search behaviour, sales velocity, inventory levels, store-level traffic and supplier capacity all sitting in different places.
AI can bring those signals together and identify patterns much faster than a human team could.
Then you can move from insight to action. Where is demand accelerating? Which stores are likely to run out first? Can I reallocate inventory from another region? Should I increase replenishment now, or do I need another day of data?
That is the part I find exciting. Decisions that could previously take days or weeks can start happening in minutes.
At Aily, our AI agents can run 100,000 what-if scenarios in seconds. So rather than making one big bet on a trend, retailers can compare different options and understand the trade-offs before they act.
Q: What data signals should retailers be watching to distinguish a trend with sustained consumer demand from a short-lived viral spike?
A: I would never look at one signal on its own.
A product can have millions of views and still not translate into meaningful sales. What matters is whether you start seeing the same pattern across different parts of the business.
You want to know whether social interest is growing over time, whether search activity is increasing, whether people are buying the product and whether that demand is spreading across markets or customer groups.
Then connect that with what is happening operationally. Are stores seeing faster sell-through? Is inventory starting to tighten? Are consumers coming back and buying again?
When those signals reinforce one another, you have a much stronger indication that this is real demand rather than a moment of attention.
This is why I always come back to the idea of a single source of truth.
Marketing, sales, supply chain and finance should be looking at the same picture and making decisions from the same information.
Q: What are the risks for retailers of reacting too quickly to social media trends, and how can AI help them avoid over-ordering or being left with excess inventory?
A: Moving faster does not mean reacting to everything.
If a retailer sees a viral post and immediately places a huge order, the risk is obvious. The trend can disappear just as quickly as it appears, and suddenly you have excess stock, tied up working capital and markdowns at the end of the season.
AI should help you make a better decision, not simply a faster one.
Maybe the right action is to move existing inventory between stores before ordering more. Maybe you increase replenishment gradually. Maybe the data shows that social engagement is high, but actual conversion is still weak, so you wait.
Retailers need to understand why the AI is recommending a particular action, and which signals are driving it. It shouldn’t feel like a black box telling you to buy more inventory. The recommendation should be grounded in the business and help the team understand the trade-off so they can make the decision with more confidence.
Q: How do you expect AI-driven demand forecasting to change the relationship between social media trends, consumer behaviour and traditional retail inventory planning?
A: I think forecasting is becoming much more continuous.
Today, many companies still operate in planning cycles. They create a forecast, build an inventory plan and then wait until the next cycle to adjust it.
But consumer behaviour does not work in cycles anymore.
With AI, new information can constantly feed into the forecast, whether that is a change in consumer interest, sales, inventory or supply. That also means marketing, commercial, supply chain and finance no longer need to make decisions from separate views of the business.
Over time, agents will not only tell retailers that something has changed. They will help determine what to do about it and, within predefined guardrails, increasingly take action.
That is where I see retail planning going. Much more real time, more connected and much closer to the actual consumer.
More from Retail Insider:
