Vidora’s AI Spotlight Series is a series of conversations with leaders, academics and pioneers in the world of artificial intelligence about where AI is today, and how it will change the way we do business.


Stella & Dot is one of today’s fastest growing social selling companies. Recently named an Inc. 500 Fastest-Growing Company, the San Francisco-based business develops boutique-style jewelry and accessories. that These are then sold socially through independent Stylists and in-home trunk shows. Stella & Dot is one of the most innovative businesses in the fashion and social selling space. This comes from the way they combine elements of eCommerce with AI, social media marketing and personal service to attract both customers and Stylists.

Stella & Dot VP of Product, Meera Bhatia, has been at the forefront of Stella & Dot’s social selling model. We had the opportunity to talk to her about how the space may change moving forward, and if AI might have a role to play in that. We also spoke about what its limitations might be. In doing so, we covered ideas from predicting churn, to influencing how products get sold, to the creative process itself. Here is our conversation:

On AI’s possible role in predicting churn & improving retention

One of the areas we’ve heard a lot about in terms of managing customer relationships is understanding how to predict and prevent churn. Is that a priority for a business like yours?

MB: Yes, certainly. I think that’s the number one priority for just about any business. One thing that is unique to Stella & Dot is that predicting and preventing churn isn’t just something we think about in terms of the end customer. Since we have over 40,000 Stylists, retention is just as important when thinking about Stylists.

How does Stella & Dot think about Stylist loyalty today? What goes into driving that?

MB: Well people decide to be Stylists with Stella & Dot for a range of reasons. For some, this is a fun way to make some money on the side, while for others, being a Stylist is a full-time job to provide for their families. The signals of loyalty look really different from Stylist to Stylist, so it can be challenging to predict whether an individual Stylist will churn or not.

Do you think AI could potentially give some signals in terms of understanding Stylist churn?

MB: There’s a lot that’s TBD with AI. Ultimately if a Stylist isn’t selling, they are going to churn – that’s the obvious part. But there are many reasons why a Stylist isn’t selling. Maybe they’re not booking trunk shows. Maybe they’re still developing sales experience. Usually a Stylist isn’t selling for a number of reasons – so identifying trends early for each Stylist, and then being able to act on them quickly, is where a predictive engine could show some value to us.

The impact of reducing churn with AI

When thinking about either your Stylist base or your customer base, if a predictive engine could reduce churn by say 1-2%, what would that mean for your business?

MB: Of course those kind of reductions in churn would be massive. Because we want consistent interactions between Stylists and customers, reducing churn in any given area ends up having a big impact on purchasing habits down the line. Improving retention is additive in social selling, so getting people signed up as Stylists, bringing customers on board, and then keeping them around, are all crucial metrics for Stella & Dot. Whether AI can ultimately deliver that or not is another matter.

Do you think that there is a role for AI and Machine Learning in predicting, understanding and ultimately automating engagement to improve those metrics?

MB: Well a system that could do those things would be the holy grail. But whatever AI provides in terms of insights must come alongside the ability to change the outcome. That’s absolutely key. Stella & Dot has plenty of data on our Stylists and customers, so if AI can help us make better decisions there, that would be useful.

What about looking beyond churn – can an AI help in other areas?

MB: It’s hard to know how much AI can provide – it’s possible. Perhaps AI can have a role in helping us identify great Stylists and helping them become brand ambassadors. There are different benchmarks that Stylists can hit before they reach what we call “star level”. Being able to detect and predict which Stylist is going to hit “star level” at an early stage would be very valuable for us.

On the role AI could play to influence selling

Stella & Dot is seen by many as a pioneer in fashion retailing because of this unique approach of having Stylists that sell to customers. Would you be able to walk us through what that model looks like?

MB: The Stylist-customer interaction is at the core of our business. Usually, someone who wants to be a Stylist will set up a trunk party that they’ll host for their friends and family, with a range of Stella & Dot products available at them. Those kinds of informal get-togethers go a long way to bringing in customers, and also encouraging more people to be Stylists themselves. For customers, those parties help them get a feel for our products, and most importantly give them an avenue to begin having a relationship with their Stylists. Our role is to help our Stylists nurture those relationships.

Aside from looking at questions of churn among Stylists, do you think AI has a role to play in improving the selling process itself?

MB: Whatever role AI could potentially play in the future, I think it would have to be as a tool to empower Stylists. In addition, it should assist their relationships with customers. At its best, perhaps AI has the potential to be something like a Stylist’s personal assistant. That would mean AI helps them make better choices for their customers. Depending on where things go, AI may be able to empower the Stylists with trend data, and recommendations for the end customer.

Will AI replace salespeople?

What about in other businesses, where priorities are different? Can you see Stylists or other salespeople being replaced by an AI?

MB: In the fashion space, I think that personal interaction really matters. There are other companies that have these ‘fake Stylists’. People can tell the difference. As a result, I think that has a big impact on customer satisfaction.

Over time though, AI is going to get better and better at mimicking human interactions. Let’s say it does eventually pass the Turing test for a fashion business – would that still make people angry?

MB: It’s hard to say. I think the distinction lies in some customers wanting a person to guide their shopping experience. There are others who are happy to buy from some online store, but we don’t cater to them. For these people, I guess human relationships don’t matter to the experience. As a result, I don’t think they’d care about whether that AI can pass a Turing test or not either way.

What about pricing? How does that happen today, and do you think AI can improve it?

MB: With a lot of companies the pricing process is somewhat static. As a result, there may be some potential for AI in that area. There is certainly data that isn’t always capitalized on. Quality, popularity and trends are among those. Maybe AI can help make the way businesses price more based on real-time data than it is today.

On AI’s potential role in the creative process

What about the products themselves? If pricing and selling products is static today, would that also apply to the products that are promoted to end customers?

MB: Well in general I don’t see much of a place for AI in the creative side of the business. Perhaps an area it could be useful though is portfolio strategy. Stella & Dot puts together its product portfolios in a really collaborative way between our design team and our Stylists. To go into more detail, this means Stylists vote on the products in our line. Because of this, they can give us real-time market feedback. We combine that information with historical data to understand trends of what has sold well in the past. It also helps us understand how to present products. A predictive intelligence framework could help augment that part of our product strategy. If that is the case, we could also leverage data in more dynamic ways.

And what about the creative process itself – is there a role for AI in actually making these products?

MB: I don’t think so. That’s similar to the line I draw with the role AI can play in customer relationships. Just as with our Stylists, I think the limit of AI is capped at being a tool that our creative team can leverage. I don’t see a place for it in making creative decisions.

So as you view those things together, how do you assess the areas in your business where AI can make an impact?

MB: Whatever you’re selling, you want to empower people to make the best purchasing choices. In addition, you want them to have the best relationship they can with the person who’s selling to them. Any role that AI can play in helping to make that possible would be valuable to a business. There’s a lot to be seen in the coming years about whether or not that is possible, but it’ll definitely be interesting to keep an eye on.

You can find out more about Stella & Dot here

Read more from our AI Spotlight Series:



Want to Learn More?

Schedule a demo and talk to a product specialist about how Vidora’s machine learning pipelines can speed up your ML deployment and ultimately save you money.