The Orchestration of Winning the Recommendation, with Jing Feng, Co-founder and COO at Bluefish

Transcript

Our transcripts are generated by AI. Please excuse any typos and if you have any specific questions please email info@digitalshelfinstitute.org.

Lauren Livak Gilbert (00:00):

Welcome to Unpacking the Digital Shelf, where industry leaders share insights, strategies, and stories to help brands win in the ever-changing world of commerce.

 

Peter Crosby (00:25):

Hey everyone, Peter Crosby here from the Digital Shelf Institute. As the squished funnel of the AI shopping conversation continues to proliferate, we are seeing strong data that indicates conversion rates are higher than traditional search-based journeys. And with studies showing that 60% of consumers blame the brand itself, not the site or retailer they got the information on, the stakes for getting answer engine optimization right are rising. So Jing Feng, co-founder and COO at AI marketing platform Bluefish, joined us to share best practices and the cross-organizational orchestration required to win the recommendation. Welcome to the podcast, Jing. We are so delighted to have you on. Thank you for making the time.

 

Jing Feng (01:11):

Absolutely. Hi to both of you. Great to see you again. Thank you for inviting me.

 

Peter Crosby (01:15):

Of course. You and your team sit in Agentic commerce every day in the flow of it, the stream of it, and you do a lot of time to try to understand, take a lot of time to try to understand the outcomes and the impact it is having for brands, which is a really popular and pressing question these days, especially as people make their 2027, what the heck are we going to do about this plan? So there's a lot of talk about all this Agentic, AEO, GEO, how AI is affecting the consumer journey. So just give us your take on everything.

 

Jing Feng (01:53):

Great.

 

Peter Crosby (01:54):

Yeah, all of that stuff and what really matters right now. What should folks be paying attention to?

 

Jing Feng (02:01):

Yeah, definitely. I think you called that out, Peter. A lot is happening all around us. I think to us as marketers, a lot of marketers are experiencing this intense pressure. I think the primary context, and given the limited amount of time we have, I think the biggest key thing is just that the consumer journey is collapsing and it's collapsing into effectively a single interface. And I'll go into what that means, but essentially discovery, comparison, and increasingly even the purchase itself is happening inside that AI conversation. And that's before the shopper even reaches your brand site or a digital shelf or et cetera. And so that has pretty major implications for a marketer because it means, and just some interesting stats for you, a third of consumers now are starting with AI tooling instead of traditional search, that number's increasing. I don't even think that accounts for folks who are on Google and getting the AI overview, so it's actually far more significant.

 

(03:16):

I also don't know if you guys saw this, the new Shopify data that was just released. Huge, what they're showing in their data for everyone who's listening, for research intensive purchases, AI referred shoppers, that traffic is driving 2X the conversion relative to traditional search. That's huge. And then on the other main category, they refer to it as taste-led categories. It's basically consumers who already know their preferred product or brand. They're saying that AI referred traffic is bringing 1. 3X the number of net new customers relative to search. That's a 30% increase on discoverability just based on the channel alone. And this kind of leads back to what I was saying is that AI is the full funnel. This is a really important takeaway for brands. It's as much of a brand exercise as it is a performance marketing exercise. And there's a really big mechanical shift that's underpinning all of this, meaning historically as marketers, we are used to thinking about these different channels, different touch points all along a consumer journey.

 

(04:39):

But in AI or now with AI, I should say, all of these individual channels and touch points are training the models on how to behave and who to recommend essentially. So it means that models, not just retailers or shoppers are now deciding what products appear in what order, at what price, et cetera. So I think this is probably the biggest thing is this mindset shift that needs to happen, that AI is not a new tactic, it's not just a new channel, it's a whole new audience that's going to make decisions on behalf of your consumers. It's your growth lever. You have to treat it that way because there are consequences if you don't given the growth and the brands who are building their orgs and who are orchestrating their marketing strategy around this shift. These are the brands that are already pulling ahead, but these are the folks who are going to retain advantage.

 

(05:48):

So does that make

 

Peter Crosby (05:49):

Sense? Yeah, absolutely. And it makes me think everyone's trying to figure out whether this activity can go under the incremental growth column. And so when you talk about those Shopify stats, in my somewhat blurry understanding is that the fact that those conversion rates are going up at the rate that they are and it's new customers, assuming that proves out across platforms and over time, that's considered incremental. Would you describe it as that versus just somebody switching the channel or the way in which they bought?What do you and your customers talk about when you think about how

 

Jing Feng (06:35):

To

 

Peter Crosby (06:35):

Justify this investment?

 

Jing Feng (06:37):

A hundred percent. And I think that's the real question, Peter, which is how do you, from a CMO level and a budget perspective, how do you allocate funds? I think the reality, and this is part of the challenge, is that it is both incremental and basically reshaping all of the existing tactics and budgets. So that means that brands and marketers are being put in this uncomfortable position where they have to invest and they have to justify the incrementality, but they also have to understand that if you don't, other competitors are basically taking a bite out of your existing market share and they will and in significantly compounding ways because of how Fast AI is developing. So I think that's why it's uncomfortable right now because you have to find that extra budget to invest. You can't necessarily say it's incremental yet. I think everybody is looking at different data sets to prove this out.

 

(07:50):

And there is this mystery of the disappearing traffic because a lot of traffic has disappeared going direct to brand sites, but the referrals aren't necessarily there, but that's because people are learning outside of the ecosystem that you control. So now you really have to think about, well, what does my brand.com even do? Is it still education if folks have already learned the majority of what they want to learn? So what do I want this part of the journey to now look like and how do I optimize for that? And so the reality is that there probably needs to be a investment across the board to reshape how every channel is thinking about AI and is optimizing and thinking about AI as its own channel. But it's not always clean cut, especially if you're coming from a CEO's office of how should I invest? What's the ROI?

 

(08:52):

Because the reality is those numbers don't exist anymore because the chain has been broken. And so that's the kind of uncomfortable place that we're in.

 

Lauren Livak Gilbert (09:02):

Reminds me of the early days of e-comm when we're like, we know it's important, please

 

Jing Feng (09:07):

Give us resources. Or social. Yeah,

 

Lauren Livak Gilbert (09:10):

Exactly. And social's still going through its own

 

Jing Feng (09:13):

Identity. Manifestations.

 

Lauren Livak Gilbert (09:16):

I though that's a better way of putting it, yes. But in the same way that most brands who are investing in TikTok are spending money, are not necessarily making money, but are seeing the effect on Amazon or on other channels. I think that's really the same for AI. It takes money to invest in looking at AEO optimization and creating content and things like that, and you'll see the effect. It just might not be a one plus one equals two type of situation, and that's hard to explain.

 

Jing Feng (09:41):

Yes. But everybody's seen this chart, the adoption in social, mobile, we though that moved fast. This is moving, I think the chart is at three or four X the pace. So we're having all of our brand customers, retailers are having to make that decision now in a much more condensed timeline, and that's very uncomfortable. But the impact, I should also say, the inverse of that is the impact is also really, really condensed and fast. And I think people should not forget about that because if you're an established brand and you have a bunch of challenger brands around you, now's the time for them to swarm and vice versa. So this is really critical to think about.

 

Lauren Livak Gilbert (10:38):

Yeah. And I like how you talked about how it's a part of your strategy and you have to think about it holistically. There's a lot of conversations around, oh, well, is SEO dead? Or how does AEO mix with SEO and what does that look like? And I see them as both distinct important features of your overall strategy, but how do you talk about that with brands? Are they competing? Are they working together? How do you chat through that?

 

Jing Feng (11:06):

100%. It's definitely working together. I think most folks understand that. We hear Google saying good SEO is foundational to good AEO. I think that's well established. At the same time, we also know, and LLMs are also out here trying to tell us that it's not the same, that AEO is much more broad. It encapsulates that full funnel. It's model driven. It's not algo that's searching to match keywords. AI is, there's just far more compute. It works differently. It's more powerful. It registers intent, it has memory. And so I think the trap that sometimes people can run into is they assume that SEO equals AEO and you can just take the same playbook. But a really important distinction there is that search is, or traditional search has historically been pretty deterministic, meaning there's an input of a string of keywords and you get a particular list that comes back.

 

(12:27):

AI answers are by design, highly personalized, not meant to look the same, even if you prompt one after the other one time after another with the exact same prompts. So the implications of that are super important. As an example, Peter, I'll call you out if you and I both put in the exact same prompt for good skincare. We're

 

Peter Crosby (12:54):

Practically the same person, James.

 

Jing Feng (12:57):

Yeah, exactly. We can be pretty sure that we're going to get very different - I

 

Peter Crosby (13:02):

Believe that is true.

 

Jing Feng (13:05):

So from a data and data integrity standpoint, it means as a brand, for the same prompt, you might show up 80% of the time for me, but literally 0% of the time for Peter. So you have to make sure that you're optimizing for intent, that you're optimizing for the way that AI and AEO works, because if you're just transferring that playbook and you're just focused on SEO tactics and just focus on the prompts, then it means that you're essentially getting bad data and then you're going to make poorer decisions and you're going to get worse ROI. So there are really important implications to treat them as distinct things where having both is a necessary part of a good marketing strategy.

 

Peter Crosby (13:54):

Yeah, we would call that the context layer of data that you need, that the use cases, for whom, at what time, what's happening, what's the weather? Any of that stuff has to be available to the AI to be able to tap into,

 

Jing Feng (14:11):

Depending

 

Peter Crosby (14:11):

On what it knows about you and the context of you, they have to have the context of the data to then serve that up. And that's a big task. It is.

 

Jing Feng (14:22):

Putting

 

Peter Crosby (14:22):

That together is a lot of work and

 

Jing Feng (14:25):

Needs

 

Peter Crosby (14:25):

To be automated, right?

 

Jing Feng (14:27):

A hundred percent. Yeah. And we refer to them as hidden contexts because it's below the surface. The interaction that I have at this point with my AI, my new BFF is I can ask a simple question and then it's essentially asking me, "Oh, but what about this? I remember that you like X, Y, and Z. I remember that you live in a state that's high altitude, has a lot of sun exposure, therefore we recommended this." So the hidden context really comes out in that conversation. You really want to make sure that your products are optimized for those hidden contexts.

 

Peter Crosby (15:08):

So when there is incorrect information about a brand online, a survey from a company called Ritthum, R-I-T-H-U-M, found that 60% of consumers blame the brand itself,

 

Jing Feng (15:20):

Not

 

Peter Crosby (15:20):

The site or retailer they got the information on.

 

Jing Feng (15:24):

Or the AI, yep.

 

Peter Crosby (15:26):

Or the AI. Yeah. And so that's super challenging because you were talking about it earlier, the brand has just less control, less exact control, and also the volume of what's required to win that recommendation is so much higher. And how do you suggest that brands handle that in a sort of lower control environment?

 

Jing Feng (15:51):

Yeah. Well, first I was going to say, isn't that a super enlightening statistic that they don't - Enlightening,

 

Lauren Livak Gilbert (15:58):

Horrible. Yes. Bad, scary. We've got a lot of adjectives on that

 

Jing Feng (16:02):

One. Yeah, totally. Yeah. Well, I think the same study literally also said that 90% of those consumers abandoned the cart after blaming the brand. So not a super forgiving environment, which really has really important implications. So maybe I'll back up and just say, it's interesting that we talk about visibility in AI a lot. I think that's a starting point. But once you layer something like this data in, you can immediately see that greater visibility can also equal greater brand risk if the information is incorrect. Now, this is really critical for marketers to understand because a lot of. Depends on the vertical, but a lot of marketers don't have this risk framework or safety framework that they've really built out yet. So this is an important area for marketers to seriously think about, especially when we relate this back to what we talked about earlier about it being full funnel.

 

(17:16):

So if you're not being included or discovered because of incorrect assumptions in the upper funnel, then you're not going to show up in the lower funnel. But back to your question around control, because it is a scary citation. So the thing that I would say is that you actually have more control than you think as a brand. And this is actually by us looking at inaccuracies at scale. And what we've found is that across verticals, across brands, about 30%, it goes up and down based depending on the vertical, but about 30% of those inaccuracies can actually be traced back to the brand's own pages, meaning a meaningful share is actually within the brand's control. So it

 

Peter Crosby (18:17):

Is

 

Jing Feng (18:17):

Their fault. Yeah. It's just that the sites are so hard. Lots of point fingers.

 

(18:24):

So not everything is just wild hallucinations and they do exist. They 100% do, but there is some control that you can exert. We also found 64% of those AI errors are due to incorrect numbers. So this was something that surfaced quite a lot. So it could be a dosage, it could be a price, it could be a rate, which again, has super important implications if the consumer is never leaving that conversation and making a decision based off of everything that's showing up there. But again, you can ultimately, if you can trace where the incorrect information is being generated from, then you can do something about it. And that is why from a Bluefish standpoint, we're so focused not on just if something is correct or incorrect, but on providing that traceability back to our customers because that's where the action can actually come in so that you can go and figure out which of my own first party pages are providing the incorrect information, which third party pages are just out there wilding out.

 

(19:46):

And of course, there's a percentage that's going to be just hallucinations that is maybe out of my control today, but hopefully in the future, this is something that I can directly influence with AIs as well.

 

Lauren Livak Gilbert (20:01):

As someone who used to work with an OTC brand, the comment you just made around dosage just made my antennas pop up because even if you as an over-the-counter brand don't want to engage with LLMs, you need to see what they're saying because if someone is. To the stat, if the consumer's blaming the brand versus the LLM and it has incorrect dosing information, I mean, we're getting into territory where you should not take medical advice from an LLM, but again, it is the number two use case of LLMs, I think, from data -

 

Jing Feng (20:34):

Medical. Yeah, related, for sure.

 

Lauren Livak Gilbert (20:37):

I guess the point I'm trying to make is you still need to know what LLMs are saying about your brand,

 

Jing Feng (20:42):

Even

 

Lauren Livak Gilbert (20:42):

If you are not doing anything about it.

 

Jing Feng (20:44):

Yes. And especially around these very sensitive gray areas around risk and safety, 100%.

 

Peter Crosby (20:53):

Yeah. I'm not sure that the risk team at their company would say even if you don't do anything about it would be unacceptable. I mean, once you know, you kind of have to do something about it, right?

 

Lauren Livak Gilbert (21:03):

You would hope so.

 

Peter Crosby (21:04):

Yeah, we would hope so. Yeah, I know. We live in a very difficult - The key

 

Jing Feng (21:07):

Takeaway for the audience is don't just listen to LLMs when you're searching for dosage.That's

 

Peter Crosby (21:13):

One of the takeaways.

 

Lauren Livak Gilbert (21:14):

You heard it here.

 

Peter Crosby (21:15):

Yes.

 

Jing Feng (21:17):

Saving countless lives, Peter and Florian.

 

Lauren Livak Gilbert (21:20):

That's what we do here at the DSI, save lives. But Jing, do you have any examples of any brand experiments around Agentik Search that have worked, maybe haven't worked? Our audience always loves to hear those fun examples.

 

Jing Feng (21:36):

No, examples are important because I do think that there's a lot of talk around AEO, but then very quickly the question, especially as it relates to budgets and investments is where's the attribution? Does it work? So I can walk you through an example of a global beauty brand, for example, that we work with. So this is a beauty brand that sells within retailers on Amazon. They have their own e-com site, so really across the board. They have a few brands that are within a similar division or product line, but they really wanted to. Of course, they started out same as everybody else by just wanting to understand discoverability across the board. But very quickly, and I'll walk you guys through what happened, they were able to pinpoint some really critical aspects that then influenced how LLMs were presenting their brand, especially on Amazon. And that actually resulted in really significant increases in actual sales and ROI.

 

(23:00):

So essentially, they're a Bluefish customer, of course. They started with just basic audit and benchmarking of what's happening across the different LLM surfaces. In the platform, they could then see that a number of their PDPs, especially on Amazon, were lagging competitors. And as a result, they were just not surfaced as highly within Rufus responses. And so then using a lot of the diagnosis data and analytics in our platform, they were able to audit those PDPs and then just pinpoint the very specific narrative gaps that they were not strong enough on relative to competitors. And a lot of this was around. It was a variety of medical/dermatological claims, essentially. And so we worked with them to figure out what's the right way to craft that narrative to really showcase the right representation of their products. And I want to be super clear that this is not about gaming or abusing the models.

 

(24:23):

I think this is really important because hopefully you guys heard and everybody else also heard at Google Marketing Live recently, they were really clear in saying, "Look, if you're going to go out and use AI to create a thousand pages on certain topics, we're going to market a spam, and if you don't stop doing it, we're going to punish you for it." So you don't want to be doing that. What we're really trying to do is understand how LLMs are extracting information from the sources that you're providing to train their models. And so once we were actually able to do that, then we implemented the right changes and that actually yielded, I think it's over 8% increase in visibility, which actually correlated to a 20% increase in product views and an overall 9% increase in actual ordered revenue. That's a huge lift, think about it 10% over a matter of months.

 

(25:25):

So that's one example of having the right granularity and understanding where you really want to target those optimizations for which audiences and really breaking that data down so that you can basically create the right sets of targeted content to create impact.

 

Peter Crosby (25:51):

So going to that idea of gaming the system, which became. Well, I don't know whether it's fair to say it became easy to do in SEO land, but certainly

 

Jing Feng (26:01):

A

 

Peter Crosby (26:01):

Lot of people made a lot of money being the gamers of that system. Do you feel like when you talk about don't throw up thousands of pages of content, are you talking about then your. I'd just like to dig deeper on that. Should your investment in adding that context layer and adding those things that will get you into more considerations and up your recommendation, are those better off happening on your core site rather than a lot of new stuff spinning up? Or what are your takeaways from that kind of warning from Google Live?

 

Jing Feng (26:41):

Yeah, that's a great question. I think it's two things. I think one is just that it should be about quality, not quantity, is part of what that warning is implying. Then the second piece of it, I think it's also important, which is where do you deploy the set of information? So I think talking about the content itself, what folks are saying is the content itself should be more reflective of being actually useful to this new audience, which is an AI agent that's trying to be helpful to their human. That at the end of the day, that is what all of this compute is going towards, which is to actually be helpful. Now, the AI agents do work differently to humans. It's very practical in many ways. And in trying to be helpful, it is trying to understand all of your circumstances. And so what is helpful in messaging to this new audience is going to be different to what is helpful in messaging directly to a human.

 

(28:08):

It's less emotional. It's got to be use case driven is often what we see. You have to be specific. It's not about to my AI, it's not about how the tinted skincare or tinted sunscreen is going to make it easier to get ready in the morning. That's more of an emotional message. It's that it is SPF 50. The tint does not appear artificial. It doesn't rub off on clothing. What we see in terms of successful content is that it's very use case driven and very, very specific. That's where the performance that you're looking for is really going to come out. You still need to market directly to consumers. You still need to have that emotional messaging in order to retain that brand loyalty and to connect with the shoppers, but it does have to be both and it is going to be really different. And so that's the quality that I'm referring to.

 

(29:27):

Once you've determined what it is that you need to say, which audiences, what particular use cases are really important to these audiences, what are the things that I really need to get out there for the AIs to understand, then it's a matter of authority because at the end of the day, AIs are trying to parse through the entire internet and just like a human, they're parsing out maybe they're really outspoken And they're parsing out the slop, which is literally the warning. And they're saying, if you can actually be helpful, if it's authoritative, and many people agree with you because they're also trying to validate, then we will make this recommendation. And so that validation is super important. So this is where your reviews become intermixed with your brand, where social mixes in with brand with a capital B. This is where PR becomes really important and is all part of that mix to establish authority.

 

(30:42):

And that's not necessarily on brand.com.

 

Peter Crosby (30:46):

And that kind of brings me to my last question, which is who are the people at the table for all of this work? Because it is different. It is broader than I think. I don't think, maybe this isn't entirely true as I start to say it, but PR wasn't heavily involved in SEL. They might dip in and say, are there some terms you want me to go after or something like that? But it feels like this is much more of a sort of all hands on deck kind of thing. And I'm wondering in terms of your product, who are the users and then who are the consumers that may not be actual hands-on in Bluefish, or maybe they are in the way that they're able to get data but maybe not do a lot of it? I'd love to know who the players are in all this work and how close do they get to your product?

 

Jing Feng (31:39):

Yeah, no, it's a great question. The short answer is there's no one team that should be the only users. I was just at the MMA and we were talking about this, which is really just orchestration across marketing teams now that we're in this new paradigm where every channel affects the AI channel. So in terms of users, I would say we're still in relatively early innings of this space. So we have a lot of SEO folks who originally were handed this mandate. Increasingly, we see a lot more brand folks who are in the platform trying to understand how they can influence discoverability, what's the narrative that's being shared out there. PR teams are a hundred percent part of the mix, often cited. Content teams, affiliate teams as well, because a lot of these affiliate articles are actually being cited in AI and essentially grounding and influencing the responses.

 

(32:57):

So it's a good way to use our platform to understand what's the ROI for certain YouTube influencers or for affiliate partners, et cetera. So it really is everyone, but it is still at the beginning, I would say stages of familiarity. But because AI as a paradigm is so full funnel, every team has a role to play. And I think that looking forward, and if there's one thing that we would share with CMOs and CEOs is you have to be looking at and investing in this orchestration as it relates to AI. Because you're right, Peter, historically, all of these teams have been relatively siloed, but in the world of AI, if you're not aware of what one hand is doing, they could be investing in certain topics that are going to negatively impact what the other hand is doing. And then if AI gets confusing messaging or just non-consistent messaging from a narrative standpoint, you're basically diluting your efforts in what will probably become the most consequential channel for marketers.

 

(34:22):

So this orchestration is super critical for CMOs to really think about and how to structure their teams and really enable their teams to be cross-functional, to be successful in this space.

 

Peter Crosby (34:39):

Because I keep thinking about the actionability of the data that you or anybody else who's in the recommendations. So you

 

Jing Feng (34:49):

Do

 

Peter Crosby (34:49):

An audit, you look for those opportunities, those become recommendations I imagine, and then the recommendations need to be probably prioritized. I don't know. I mean, I'm wondering,

 

Jing Feng (35:01):

Because

 

Peter Crosby (35:01):

Orchestration is about getting the recommendations into action. And when you talk about that number of teams, and who probably already have some flow of workflow or task management or something, how does all of that today work and what is your ultimate vision for how that

 

Jing Feng (35:20):

Happens

 

Peter Crosby (35:21):

At machine speed potentially? Yeah,

 

Jing Feng (35:24):

Absolutely. And I think a few things here. One is just that we work with Fortune 500s, and for each of these big organizations, they have their own intricacies. And so we basically can't fold everything in the same way to get to a particular shape. And so I think that it's going to be a transition similar to the digital media transition that's happened over the past decade or so. And we've seen a couple of models though starting to emerge. And one is a more centralized model where you have one team that not only has the knowledge and the time or the mandate, but they also have the resources to help execute. So we've seen this emerge as one kind of model. Another model is more of a distributed model where each brand has their own resources, their own ability to execute their own ways of working, but there's a centralized kind of governance committee or something like that, that's just looking across everything from a governance standpoint and less execution.

 

(36:54):

What I would say is that whichever path is the closest and most achievable for you as a brand is the path to take. The most important thing is to put one foot in front of the other and make moves here because this space is shifting so quickly and orchestration I think is going to be a big determination of whether brands win or lose just because the space is moving faster than we've ever experienced it before. And of course from a scale standpoint, Peter, you called this out, that's important and that's why from a platform standpoint, we provide those orchestration tools so that you can collaborate together and measure together and take action together in one place. But from an org standpoint, messaging to CMOs is plan for that today, how you're going to have that governance and execution, whatever version of that looks like.

 

Lauren Livak Gilbert (37:57):

That's what I'm also seeing from the org side of things. And I know I keep using e-commerce as a comparison, but it's very similar to the days of the early COE where it's like, "Hey, this is important. Let's put some people on it. Okay, great. Now we know it's important. We got everybody's buy-in, let's democratize it back into the business." So I think we're going to see a lot of that with AI where if you have a chief AI officer or if you have someone who's focused on AI, that's not a role that may necessarily continue for 10, 15 years. It's like, "Hey, let's establish that this is important. Let's put dedicated time to it. Let's

 

Jing Feng (38:33):

Educate - Learn the motions.

 

Lauren Livak Gilbert (38:35):

Exactly. And then it becomes part of the business because you shouldn't be thinking about AI as a separate channel. So I encourage people to just look at the correlations between what you've already been through and what's worked and just apply those to this new but faster change that we're seeing.

 

Jing Feng (38:53):

Yeah, 100%. We've been through it before. We're resilient as marketers. So it's all about just starting down that path and taking the lessons, like you said, Lauren, that we've already learned.

 

Peter Crosby (39:09):

So before we completely let you go, I just want to let our listeners know that bluefishai.com is the place to go to get really rich content and advice around all of these things that we're talking about from technology to orchestration to organization. And it's a great resource. And so I'd recommend folks do that. And so again, Jing, thank you so much for bringing all of that to our attention.

 

Jing Feng (39:32):

Thank you guys.

 

Lauren Livak Gilbert (39:34):

Thank you so much, Jing.