Beyond the Wave: GenAI, Guided Selling, & the Future of Search

Beyond the Wave: Gen AI, Guided Selling, and the Future of Search

Welcome and Introductions

Nate Roy: Thank you all for joining Beyond the Wave: Gen AI, Guided Selling, and the Future of Search.

Just a few housekeeping notes before we get started. You're most likely going to have questions as we go, so please feel free to drop those in the Q&A pod. It is a different button than the chat pod, so just be mindful of that — look for the button that says Q&A on your toolbar. We'll get to as many of those as we can during the session, and we've also reserved some extra time at the end.

A recording of today's webinar is going to be sent out in a follow-up email, so keep an eye out for that. If you find the session helpful, please feel free to pass that recording along to your colleagues or use it for your own reference.

So, who do we have speaking with you today? My name is Nate Roy. I'm the Senior Director of Product Marketing here at Constructor. I've spent about a decade in commerce technology spanning PIM software, product content syndication, and now search and product discovery. I'll be your moderator for today's discussion, and I'm really excited to guide what should be a really fun conversation.

We've got two all-stars with us today. First, I'm excited to introduce Kat Flowers. She is the Director of Customer Success at Constructor. Some of you on here might know her already. She's been in commerce technology also for nearly a decade, with experience spanning technical success and operations. Across her career, Kat has worked directly with hundreds of retailers and customers, and indirectly supported thousands. At Constructor, she leads Customer Success across the Americas and APAC, and she's personally partnered on initiatives that have delivered more than $30 million in incremental growth for our customers. Kat, please say hello.

Kat Flowers: Thanks, Nate. Hello, everybody! Super excited to be chatting with everyone, and of course sharing a lot of insights with Emily.

Nate Roy: And then finally, I'm thrilled to introduce our guest speaker, Emily Pfeiffer. She is a Principal Analyst at Forrester. Emily covers the full spectrum of commerce technology at Forrester. She writes the Waves for search and product discovery, order management systems, and commerce solutions. She advises digital business leaders on their commerce tech strategies, helping them make smart decisions about vendor selection, management, and long-term growth.

Before joining Forrester, she spent years in the retail industry as a practitioner, which I think is really unique — most recently as a VP of Marketing and Digital for the largest designer and importer of blankets and throws in the United States. What makes Emily's perspective so valuable is that beyond her research, she's lived the challenges and opportunities that brands and retailers face every day. If you've ever spoken with her directly, you know she's refreshingly direct. She tells you what she really thinks, and obviously backs that up with data and experience. She doesn't sugarcoat things, which is exactly why her insights are so trusted in this industry. She's someone I personally admire and respect a great deal, so we're very lucky to have her with us today.

Emily Pfeiffer: Thank you, Nate. That was an amazing introduction. I'm so glad to be here.

What We'll Cover Today

Nate Roy: For folks in the audience, the questions we're going to try to help you answer today: How and why did commerce search and product discovery emerge as a separate software category from commerce platforms? Is now the right time to start a search project, if you're thinking about it? How many of your peers might also be thinking about this in the retail space? What does the current market look like? How can you use the Forrester Wave Report to support an evaluation? And finally, how are generative AI and guided selling shaping the future, and what does it mean for the retail site?

Emily, I will kick things over to you to get into it.

The Evolution of Commerce Tech: From Pangea to Continental Drift

Emily Pfeiffer: I do want to talk a little bit about the role of commerce search — or site search, whatever we want to call it — in that greater commerce tech ecosystem. To do that, we're going to go back in time a little bit.

If you remember your natural history from school, the world was at one time all one continent, right? We called it Pangea, and it was this one thing, all the land stuck together. And over millennia, things started to move apart. We got this continental drift. This is exactly what happened in commerce tech.

Back in the day when we all started doing this, there was basically one commerce solution, one platform, and it did all the things — almost all the things. You kind of had payments and maybe email marketing separately, but for the most part, all the tech you needed was in one solution. Then as pieces of it got more and more complex, they started to split off and separate into their own pieces.

If you go back far enough, some of those categories — order management, PIM, CMS — all of these different pieces were actually just part of, not integrated with, but literally one solution. And so now we have these distinct areas, components, categories of tech that are separate and integrated.

There was a time when we talked about suites, when some large vendors tried to provide what they would say was "end-to-end" or "all-in-one." The value proposition kind of failed. This idea that you can get all the pieces from one vendor when some were built and some were acquired, and they're actually just integrated but not really unified — it didn't deliver. We've learned over time that we have to get these pieces from many different vendors to complete that commerce tech ecosystem.

A lot of the pieces that used to be just sort of embedded and included in commerce are not really good enough anymore. We need the standalone solutions to replace them.

Why Standalone Search Matters Now

Because commerce was the Pangea — because it was this all-in-one and all of the functionality was together in that one solution — there are these vestigial pieces of functionality that are still there. Sometimes good enough is good enough. Not usually, and not in certain areas anymore.

I kind of have the privilege, from the position that I'm in, to say that with confidence. Because I am the analyst who evaluates commerce solutions and commerce search. So in this case, I can tell you that search in particular — it's time for most businesses to look at standalone solutions.

In the last B2C wave evaluation, I had a criterion called Personalized Search. The only way to earn the highest score was to rival at least the worst of the standalone commerce search and product discovery solutions. I can't tell you specifically what the scores were, but I can tell you it's nearly impossible to find a commerce solution that has search functionality that can rival even the lower end of the leading search market. What's included is just not good enough anymore.

As tech changes, as it gets smarter, as vendors put out new functionality, digital businesses have to evolve as well. They have to do replatformings, they have to upgrade, they have to sunset certain pieces and add new ones. Commerce solutions are still really big in the scheme of things, and replatforming a big thing is hard.

As a practitioner — I like to say a recovering practitioner — I have all that experience, the love and the baggage that comes from having been on that side of things. Nobody wants to replatform. There are benefits to it, but it's the worst. And we're in tight times now in a lot of ways.

So, smaller changes are more appealing. We want to make incremental changes that bring us a shorter time to value and a really clear ROI. As those continents have drifted apart and split up, we've moved to smaller pieces of tech that make up that commerce tech ecosystem. It's actually possible to leave some of the bigger bits in place and swap out others.

Do You Need to Buy Search Separately?

Nate Roy: Follow-up question for you on that, Emily. I think a lot of the folks on this call probably came in here with the question: do they need to buy search separately from their e-commerce platform? Sounds like you're saying yes, but I don't want to put words in your mouth. How do you see that decision evolving, and what do you think the trade-offs are that matter the most?

Emily Pfeiffer: My answer is usually yes. I do have some clients who have an extremely simple and small catalog, with really clear purchasing paths where product discovery isn't really a thing. It's more a matter of "I know what I need to buy, and I just need to purchase it." So there are times I have conversations with my brands and retailers, and I'm like, you know what, what you've got in your commerce solution is good enough. But almost never anymore.

Almost every single time the conversation is: yeah, I think it would benefit you to bring in a standalone commerce search and product discovery solution. I would never generalize to an entire market — I would never say definitely for everyone. Every conversation I have is specific to what my clients need. But I can tell you that it's become extremely unusual for me to say, "Yeah, no, what you've got is good enough for you today."

These standalone solutions have evolved with a focus on search and product discovery — that's the focus. And when vendors are focused on a broader breadth of functionality, it can be really hard to keep up with a very fast-moving market, and that's what we've got.

Implementation: How Fast Can You Get Live?

Nate Roy: That makes a lot of sense. One of the other ideas that you brought up was this idea of incremental, smaller projects that generate a return. For Kat, this is something that a lot of our customers actually bring up when they come on board with Constructor. It's usually perceived as a lighter lift and less costly than a complete commerce replatform, but it can add a lot of value and improve metrics. Just from your standpoint, how long does it usually take to stand up somebody on search and product discovery, and how fast do they usually start seeing results?

Kat Flowers: That's a fantastic question. And really quickly, I want to just button up the use case as well — something Emily was mentioning was how search has evolved from just a means for finding something to now being an actual part of the product discovery process. People are different, and finding a solution that's able to facilitate the difference in behavioral patterns and how people actually use search today is so important.

One thing I'm really proud about at Constructor is that we have a really flexible timeline. Whenever you come on board and say, "Hey, Kat, we need to be live within 2 months" — awesome, we get you live in 2 months. On average, that's a 6-to-8-week timeframe. But we do have some really ambitious businesses. I've worked with retailers in the past that have set really high goals for themselves and were able to get live in a matter of weeks, with a record being 9 days.

As an outlier, it demonstrates the kind of willingness a lot of these retailers have. Once they discover the importance of a product discovery solution and partner, they can hit the ground running and see value relatively quickly, because as the experts, we take over a lot of that manpower for them.

The State of the Commerce Search Market

Emily Pfeiffer: So I wanted to talk a little bit about this market — commerce search and product discovery. The market has come together, and I'll show you a little bit about how that happened. At Forrester, we like to define, clarify, and educate about the state of things.

We find that this market is established. We look at emerging, established, and mature markets. This one is established, which means it's not brand new as a market — it's weird, though, because a lot of the tech is. We've had standalone search solutions for easily a decade, but not in the form that they're in today. So it's established but not yet mature.

What we see a lot of the time in an established market is a lot of M&A activity — lots of acquisitions, lots of things coming together.

In this established market, we find a trend, a primary challenge, and a top disruptor. The main trend is Gen AI — the trend in a lot of places right now. It's still difficult to prove the value in a lot of ways. This is an area where it's very cool, it's got big cool factor. But making sure that we know what's happening, making sure that we're measuring before and after we implement things, making sure that we can evaluate the benefits of Gen AI — we're still there. It's hard to prove the value.

Audience Q&A: Is Standalone Search Worth the Integration Complexity?

Nate Roy: This is super helpful, Emily, and we actually just got a really good question from one of our audience members. This attendee said: "If a merchant is operating on a unified commerce platform, why split off search to a more specialized solution? While it might offer enhanced search and AI discovery, doesn't this create unnecessary integration complexity and workflow disruption that can outweigh the potential conversion benefits?"

Emily Pfeiffer: That's a great question. I think it's really smart to look at all of the potential benefits and risks — benefits and costs. If the additional integration, complexity, and vendor relationship isn't worth it, it's not a good choice.

What I'm finding right now is that the vast majority of the time, it's worth it. I sit in this privileged position where I don't actually have to do anything — I sort of advise, observe, interview, research, and analyze. But I have a view of the market, not just the vendors, but my brands and retailers that are making these decisions. And I don't see a lot of regret. I don't see a lot of "it wasn't worth it, I wish I hadn't done it."

I actually see search — and sometimes even just modules — as an approachable way to evolve the commerce tech ecosystem in a way that actually brings provable, measurable results. It doesn't have to be this whole big thing. It can be adding a module or two from a vendor.

There are some organizations that are like, "We're a single-vendor shop." They want that end-to-end experience, maybe it's a procurement issue. I get that. But right now, the state of the commerce market and the state of the commerce search market are such that the functionality that comes with commerce, by and large, does not rival what you get from a standalone. If it's good enough, stick with it. Absolutely. But it may not be good enough, in which case I think the investment is worth it.

Core Use Cases: What a Complete Solution Must Include

Nate Roy: This person actually just asked a follow-up: "What specific search limitations or conversion barriers would need to exist in the current setup before specialized search would become worth that integration investment?"

Emily Pfeiffer: I don't know what to say to you right now except I hope you're my client, because if you are, you should schedule a call with me, because it completely depends on which commerce solution you're using, how long you've had it, how it's implemented, whether and how you've customized it. It depends. Every commerce solution has a different set of functionality around search, and so the gaps are different.

Let me jump ahead to the core use cases — the required use cases in commerce search and product discovery solutions. As an authority in the market, I'm telling you that if a solution doesn't have all of these things, it's not a complete category solution.

Search Configuration is the ability for a non-technical practitioner — a user, not a developer — to configure how search results show. This includes configuring search results and also potentially category pages, any product listing page.

Search Personalization — personalizing search results to the user — is a functionality that I find is hard to find in a strong way in commerce solutions.

Merchandising is pretty common in commerce solutions. The ability to assign products to categories might be pretty solid, but I don't often see it automated within commerce — for example, automatically assigning products to a "blue shirts" category as soon as they show up.

Facet Management — facets are the filters on the side when you're searching for something. Every commerce search and product discovery solution has some level of facet management, which might be weak or questionable in a commerce solution.

These are just the basics. If these are gap areas, then it's an easy decision to add a standalone solution.

What Constructor Customers Say

Nate Roy: This is super helpful and a great lead-in. Kat, how does this match up to what you hear from customers? What are the top limitations when people come on with us?

Kat Flowers: That's a fantastic question. At the highest level, the ask is: how do we know it's time to make that jump, or what are some identifiers that our current setup isn't ideal? Anything that is not decreasing or shortening the distance between your customers and the product they're looking for — whether in a discovery sense or a utility sense — is not the ideal solution for your specific business.

Typically, utility-wise, a lot of limitations that come through whenever my team starts working with a retailer include lack of functionality within the tooling. Maybe they weren't able to slot things where they want to slot them in their previous solution, or they didn't have great visibility over why products were placed in a certain place.

Then we have more technical limitations — something Emily was talking about — where anytime they need something done, they need to send it over to their engineering team to do activities on their behalf.

And then further to that, the thing just not working. Customers these days, whenever they're searching for something, might be using 3–4 words outside of the actual product information. You need a partner that's able to support the way that people are changing how they search, and that becomes a more native functionality challenge versus a utility challenge.

AI Visibility and the Top Disruptor

Emily Pfeiffer: I want to complete the thought here. In this market, the biggest challenge is visibility into AI. Gen AI is trending, it's very cool. But the visibility into the control that the AI has, the impact that it has when we want it to take over for human decisions — it's just early days, and we're still working on building the visibility and the results into the solutions.

I can't believe I got away with listing the top disruptor as I did, but they let me publish it. Usually when we say what a top disruptor is, it's a thing that a couple of vendors are doing and we see it really shaking up the market. In this case, I'm looking to the future. I think there may be a vendor or two that guesses right about how shoppers actually want to shop — Gen AI, agentic, those kinds of things. I think we're on the cusp of vendors introducing new ways to bring a really generative, positive shopping experience into the owned environment on a website.

Extended Use Cases: What Leading Solutions Offer

We talked about the core use cases — those are required. The extended ones are capabilities that any one of these vendors might provide in some combination, but they're not required to.

Product Data Enrichment is really interesting to me. I see commerce solutions literally enhancing the product data in various ways. And they're not sending that enhanced data back to the PIM — they're keeping it there in the commerce solution, so it's actually improved for the experience outside of commerce, outside of PIM.

Product Recommendations outside of search — maybe a "you'll also like" carousel, maybe in an email.

Guided Selling processes are things a lot of us have experienced as consumers but maybe don't know what it's called. It might feel like a quiz. If you're shopping for car parts and you choose the year, make, and model to ensure compatibility. If you're shopping for skincare products and you put in your age and skin concerns — those are guided selling processes.

Conversational Selling and Shopping Assistance — maybe it's agentic, but we're pushing it. Every vendor in this market is building it, in pilot, or has it live. But again, really early days.

Testing is focused on the results of these changes. Search configuration, personalization — testing enables you to try things, see what the results are, and set a goal: does it improve our changes toward the goal?

Content Search — not just product search, but how content might sit alongside the product. It might be knowledge base, manuals, or instructional content.

The M&A Landscape and Constructor's Unique Position

I love my metaphors, so I'm going to tell you that commerce tech is a lava lamp. It breaks apart, and it comes back together in different bubbles all the time. We had pieces that were maybe embedded in the commerce solution, maybe considered standalone and separate, maybe add-ons, maybe they didn't exist before. But they've all now come together to form this new market. And "come together" often means acquisition.

The vendors in the original commerce search and product discovery Wave a couple of years ago — you can see their evolution over time. 9 out of 10 of them had major acquisitions. The 10th, arguably, is the newest tech. Acquisitions have ups and downs. It's a quick way to gain architecture, functionality, or expertise. But when you've got an existing solution that then acquires other actual solutions, there's always that awkward phase of integration — hopefully more than that, hopefully unification. But it makes for a little bit of a messy market for a while.

Nate Roy: There are a lot of really good points in the last few slides. First, for audience members who are wondering how much of the core and extended use cases Constructor covers — the answer is all of them. We provide tooling for everything Emily listed on those slides.

She made another really good point about Gen AI kind of being unproven, and I think it's worth pausing on that. She's absolutely right. There are a lot of vendors right now that are slapping generative AI and agentic AI labels on products that look really impressive on the surface, but if you dig deeper, they're not actually solving real problems.

When you're evaluating technology, one of the fundamental questions you should be asking every vendor — especially for Gen AI — is: what results has this actually produced? One of the things that's really important to us at Constructor is that we focus our R&D only on purposeful applications — places where Gen AI or agentic products are truly the best solution for the shopper's need in that moment. And we don't just say that. We validate all of this through A/B testing.

The last thing I want to note: Constructor is the only vendor on this list that hasn't grown through acquisition. We've built all of our capabilities natively. And I think that's really important, especially with the pace that this market is evolving at.

Kat Flowers: I'd like to add commentary on that too. Whenever we're thinking about Constructor, our capabilities, the application — something that's been so crucial to so much of our partnership, especially being in such an evolving space where businesses are adopting a lot of this technology for the first time.

It's one of the first times in retail site history that merchandisers are starting to take their hands off and see what this technology does out in the wild, interfacing with their customers. Constructor has eyes and ears on every single part of the site. It sees every click, it sees every time something works or doesn't work, and it augments to make decisions based on that. That core logic from search spans to all of the extended functionality we were talking about.

Being in a position where we haven't acquired anything — we have so much history, insight, and connectedness with all of our products, because they're built in the same family to support our family of retailers.

And to prove how valuable they are, we're huge proponents of A/B testing. We want to test not just our product versus somebody else — we want to test our product versus nothing. If customers are going live for the first time with a chat solution, we want them to test what it looks like for customers with versus without that experience. So retailers that work with us really know that any decision they've made, any cool new technology they've brought on through Constructor — they know that it works, because the A/B testing doesn't lie.

Market Adoption: Where Digital Leaders Are Investing

Emily Pfeiffer: Let's talk a little bit about adoption. When we're asking digital businesses about their software initiatives for this year, the top two are experience-focused, which I love to see.

User experience — in this case, the customer's experience interacting with the website and with the brand. Number two is employee experience. A few of us have mentioned the tooling today — what a non-technical practitioner, a user who's not a developer, can do in a UI. I love to see this focus, and it's really exciting to me to see it so high on the list.

I always encourage my clients, when considering a new vendor, to bring the users to the demo. Take over the mouse. Be like, "Hey, you just showed that really fast, can I try?" That usability matters, because it's not just if it's a few extra clicks or not that intuitive. It's actually psychologically harder to do the work.

Increasing SaaS adoption, moving to the cloud, moving to subscription-based managed solutions — it's the way most things are going. Better personalizing customer-facing web and mobile experiences — we just talked about personalization during product discovery and how important that is.

But look at number 5: consolidating systems. There's a real push to have fewer vendors overall. And adding more specialized point solutions is much farther down the list. I get it that we want to simplify. This is one of those places where, if the search isn't good enough and you need to add it, look for a vendor that covers as much as possible in search. You don't need to add many different solutions anymore because the market has come together. Finding a vendor that checks a lot of your boxes is a way to add the functionality that you need without adding needless complexity.

The Importance of Vendor Culture Fit

Kat Flowers: I'd love to add something to that. Something we all feel really strongly about in this space is that you can even take it a step further and find the vendor that can teach you things — they can tell you about your customers, keep you informed on the market, how it's shifting, what other businesses are doing. I would even add another line under that: the most important checkbox says find the vendor that listens. The vendor that works with you, knows you, and can teach you a thing or two about yourself.

Emily Pfeiffer: I think the biggest kind of sleeper criterion in vendor selection is culture fit. Not every vendor's culture is right for every business — that's the point of culture fit, it's subjective. But yeah, finding the right vendor for your business, the kind of partner you want — totally agree, really important.

Investment Trends: 94% Maintaining or Increasing Spend

We know that this is an area where digital leaders are investing. We can see a really significant increase — 51% are saying they want to maintain, 43% are saying they want to increase their investment in commerce search and product discovery in the next year. That's massive. Maintaining or increasing. And we're in a time where a lot of budgets are shrinking. Seeing the vast majority of the market saying they're either holding on or adding budget for this is really meaningful right now.

And we can see why. Only about 26% of software decision-makers say they already have a commerce search and product discovery solution. Of the ones that don't, about half say they will. My guess is that a lot of the folks in this room right now are part of that 49% group.

Build vs. Buy: "Please Don't Build"

Nate Roy: One of the questions I'd imagine someone might have is: I think I know I need to do something beyond my commerce platform, but a lot of people talk about building search in-house. What do you think about that, Emily?

Emily Pfeiffer: Please don't do that. Good gravy, please don't do that. There are a lot of areas of functionality that over time were not well served in the market, and so savvy businesses added it. This is not the time to do that in this area. There are now vendors that are really strong and that allow you to make a lot of customizations if you need to, but please start with a default.

Custom building at this point would be a long-term mistake. What you could stand up today might be really neat for a short time. But the market's moving so quickly, and if you're not a software company, if that's not your area of focus, I think it would fall behind really fast. I think it would be pretty impossible to catch up in the first place, but even if you did, keeping up with this market right now is kind of impossible. There are some areas where I recommend building, some areas of the commerce tech ecosystem where I think it's the smart thing to do. Not this one, and not right now.

About 11 years ago, there was this ASICS challenge. ASICS sponsored a runner named Ryan Hall — a really, really fast marathon runner — and they set up a truck with a treadmill on it at the pace that Ryan Hall kept up for 26 miles. It's a sprint for 26 miles. They put people in a harness and lowered them onto the treadmill to see how long they could keep up. Most of them had a great time getting flung off. I think this is what it would feel like trying to stand up a solution right now that can rival the standalone search solutions. A few years ago, these vendors did their acquisitions or built from scratch, and now they've had years to slam functionality into it, and it's moving that quickly. This is a buy-don't-build moment, as far as I'm concerned.

How to Use the Forrester Wave Evaluation

Let's talk about the Forrester Wave evaluation. This is the thing that I do for 3 different markets: commerce solutions, commerce search, and order management systems. Forrester produces between 80 and 100 of these a year. It takes about 4 months.

We use 3 primary inputs from the vendors. One is reference customers that we can interview — we have actual conversations with customers. Number two, we ask the vendor a lot of questions so that we can have concrete responses and compare them with the other vendors. And number three, there's a demonstration. It's big, it's detailed, it's demanding, and it's specific.

The purpose of having these three really concrete inputs is so we can develop an apples-to-apples comparison. We're asking the same thing of all the vendors, the same questions of all the customers, asking to see the same thing in the demos. The evaluation is relative — it's very detailed, and it's a snapshot in time evaluating the included vendors versus the other included vendors in that moment.

The result is a report that we publish with a scorecard, which is extremely transparent. I think this is really unique to Forrester — you can see exactly how each vendor earned each score and how it resulted in the major Wave graphic.

The vertical axis is the strength of the offering — that's the product and the functionality, all of those use cases. Higher vertically is a higher score on those specific criteria. Horizontally is the strength of the strategy — the strategy for this product in particular. As the dots move up and to the right, some make it into the leaders circle. We have leaders, strong performers, and contenders. And then optionally, there are halos around a dot for a vendor, which indicates above-average customer feedback.

The Wave Results and Constructor's Customer Satisfaction Halo

This is only the second evaluation in this market. The last time, acquisitions were still happening during the eval — it was really kind of sloppy, all the pieces coming together. Now we're in a place where vendors that have done acquisitions are trying to unify, remove any overlap in functionality, and simplify their solutions. It's extremely fast-paced.

Nate Roy: You mentioned this halo as well — this measure of customer satisfaction. Constructor has a halo here; we scored highly in customer satisfaction. What do you think that comes down to? Kat, what do you think customers value most about working with us?

Kat Flowers: Absolutely. Something pivotal in our customer success strategy that folks might not feel everywhere with all SaaS vendors is that we work for people. Our customer success team puts the people-first emphasis in everything that we do. Yes, we work at Constructor, we are here to help you use our product in the best way possible. But we're also product discovery experts and digital transformation experts.

We have a personal investment and a personal interest in a business doing well. Every single person on this team wants to see our businesses succeed and shine. Does it change the money in my bank account? No, but it changes the money in the customer's bank account, and that's ultimately what we're shooting for.

Having a team of winning CSMs collectively creates a rocket ship where we are all aiming toward having every single business not only succeed and meet their goals, but go beyond that and become co-creators with us in everything that we do. It's not like we just show up to a meeting once a week, check the boxes, and walk away. We're really getting into war-room scenarios where we're preparing with customers for peak season, or we're preparing to launch something really interesting and generative on their site.

Every one of us — the biggest standout is that we're all very personally invested. And customers feel that. Businesses feel that. As a customer of vendors myself, I can feel the difference between somebody who sees me as a customer on the roster versus a business that they're interested in seeing succeed. I think that's the biggest differentiator our businesses know about us.

How to Use the Wave Report

Emily Pfeiffer: The last thing I want to say before we move on is that no Forrester analyst would say, "Oh, you need a new solution, here's the Wave graphic, look up and to the right." We have a massive scorecard that we publish. There's a whole report available.

Figure out what your specific needs are, look at the scorecard, figure out what you need. In the report, you'll find expert advice — we explain the way to think about the market and what you need right now. Review the portfolios and agreements. Explore whether incremental adoption might bring short-term relief and long-term gains — somewhere to start. You don't have to do everything all at once. Shorter time to value, get some ROI, get some buy-in before you move forward. Then identify the priorities in the functionality — look at the problems you need to solve, figure out what functions will solve them.

Forrester for clients has a digital experience that's really incredible. It's new within the past year, and it changes everything. It moves from a static report to an interactive experience. You can select the criteria that are most important to you, only show the ones that really matter, and it will actually change the score. Because the up and to the right — it's based on a non-existent hypothetical business that doesn't exist and isn't you. You don't need everything; you need what you need.

You can select the specific criteria most important to you, and it will give you a new score based on only those criteria. It lets you compare up to 4 vendors, see the full scorecards — what questions I asked the vendor (steal it for your RFI!), and exactly which score each one got and why. It's a level of transparency I don't see anywhere else.

Gen AI and the Future of the Retail Site

Okay, so all of that is very cool — these solutions exist and you probably need one. But we're hearing about Gen AI, we're having all these conversations, and this is the question I keep getting: Should we even bother? Is the retail site just going away? Is all of the shopping just going to happen somewhere else now?

Nope. No, it's not. It's not going away. I don't see that happening anytime soon.

We're going to have conversational experiences in the owned and the non-owned environments. Gen AI is generative — that means it's not a canned response, it's creating the response from the training it's been fed. Sometimes it's agentic. Agentic means autonomous — it's making decisions. If it's just "do this thing that I've instructed you to do," that's called assistive. It's not all the way agentic.

I'm being a little bit generous when I say that the ones in the owned environment are agentic commerce today. Only just barely. But I know it's only going to get more and more autonomous. If I publish saying "it's only assistive today," I'd be wrong by tomorrow, because it's changing that fast.

Common Issues with Current AI Shopping Experiences

In current agentic conversational shopping assistant experiences, there are some common issues I see across the board. I mean they're pervasive, because it's early days for these types of solutions.

I see unintuitive responses to prompts — sometimes the answers are kind of nonsensical. I see a lack of clarity in the scope or usefulness. I see inconsistencies in how it's displayed. I see answers that are not in line with product data — literally making suggestions or asking questions that are irrelevant. I see them sometimes adding friction to the shopping process, which is literally the opposite of the whole goal. Sometimes they're slow, and sometimes they're just wrong.

Best Practices for Deploying AI Shopping Experiences

Moving forward, if you're looking at bringing in these kinds of agentic, assistive, or conversational experiences:

First, ensure logical and direct responses to what the user is asking. You're not going to be in control of this — the vendors and the systems are in control of it. Your job is to test. Sometimes it's impossible to test too much beforehand — test it live.

If it's unclear in scope, provide reminders. Literally just say, "I'm a bot, and I can help you do XYZ." Make it clear.

Test hard for formatting consistency. I have seen the HTML just get stripped. Clickable stuff becomes unclickable. Formatted content with headlines and bold just gets wiped, and now you're looking at a bunch of text you can't do anything with.

Always limit responses to the actual product specs. This is huge — it should not be that creative. Make it better or don't bother. And test. Test, test, test. Things are changing, it's learning. Really check. Don't let it be wrong. Make it a trusted authority. Make it something that will actually improve the experience.

As a bonus: allow an easy in and out. Consumers did not decide that they would like their entire shopping experience to happen in a chat bubble. Chat isn't everything. Look for experiences that allow you to refine manually using the facets or conversationally by typing. Look for easy ins and outs from those experiences. And then capture the preferences — this is an opportunity to know what a customer is saying in their own words. Have a memory, remember what that customer looked for. It's amazing data to hold onto.

Audience Q&A: AI as a New Purchase Channel

Nate Roy: We just got another really good question in the chat. This person said: "As AI emerges as a new discovery and purchase channel, do you see parallels to early social commerce? Brands that launched on popular social channels gained significant advantages despite unproven ROI at the time. How do you evaluate the risk-reward of being among the first brands accessible through chat?"

Emily, I want to go to you first, but before you answer — Gen AI, at least with Constructor customers, is already proven. If you look at our website, there's a case study that just came out with Belk. They launched an AI shopping agent on their site with Constructor and saw search conversions double for people who interacted with that experience. We've had other customers using agentic AI see increases up to 200% in purchase rate and add-to-cart rate. There is definitely a first-mover advantage there, at least in what we're seeing in our own customer base.

Emily Pfeiffer: We have so many words to say about this, and we don't have that much time, so I'm going to try to go fast. We have to separate the owned from the non-owned experience here. The owned experience is the brand or retailer website where you have control over the tech that you bring in, and potentially chat or generative shopping experiences.

The non-owned environment — answer engines such as Perplexity, ChatGPT, and Google Gemini — that's a whole other thing. I do think there are some advantages to being an early mover there, but just like social commerce in the U.S. and North America, it's a tiny portion of the shopping population using it today. I do think that's going to change — I think it'll increase. I don't think it's going to break everything. So in the non-owned environment, it's a proceed-with-caution moment. I think it's going to be expensive to do it well, and I'm not sure of the ROI in the short term.

The last thing I'll say about that is to watch product data really carefully. Some of the requirements for user agreements with the answer engines ask you to turn off the filters that prevent their bots from scraping your website to get your product data. It's basically a DDoS attack. I don't love it. I don't know what else they're scraping. And it gives the merchant no control whatsoever over what product data is sent into these engines. I would really like to see integrations and optimization for those channels specifically.

Deploying AI the Right Way

Nate Roy: I love this. It goes back to things we were chatting about earlier — when you're going to deploy these experiences, they have to be purposeful. You have to know what problem you're trying to solve, not just deploying AI for AI's sake. And you have to test it. Kat, how do you see the folks that have deployed agentic experiences navigating this?

Kat Flowers: Great question. I have kind of two bubbles of experiences — hearing the leftovers from the previous life versus the new life now when they're going to try it again and do it the right way.

Circling back to what Emily called vendor culture — making sure you have a good vendor culture fit. A lot of solutions out there today have the product, and then a business can buy the product, but you can't really give somebody a car without them knowing how to drive it. I can give you this car, and you can either drive it into a brick wall, or I can give you the car and you can take it out on the highway, and it'll be a great experience.

Whenever we're going to deploy — whether it's a chatbot, a Product Insights Agent on the PDP, or something very generative and self-led — something that has to be clear for the business, and for us, is: What problem do you want to solve? How quickly are you trying to solve it? How are we going to test it? We as the vendor are going to put that in front of the customer, so we're going to work with the retailer to help them define those problems and define the approach.

The solution is no longer to just buy the thing and push it out. It's more of a slow, controlled process to make sure we're actually getting value and will help you look for that monetary gain. It's kind of 50/50 — from the past life, where businesses come over and they've got these older solutions that they've tried and failed, and now they don't really know about doing it again. And then we go for it, but we do it the right way this time.

Closing

Nate Roy: All right, we are at the top of the hour. Sincerely appreciate everybody who joined today. Thank you so much again, Emily and Kat, for joining and sharing everything that you shared. I feel like I always walk away learning new things when I talk to both of you, so this was no exception.

For anybody who has additional questions that maybe we didn't answer today, feel free to connect with me or Kat on LinkedIn. Shoot us an email with additional questions you might have — we'd love to continue the conversation after this. Enjoy the rest of your day, and we look forward to talking again soon.