The DXP Catalyst Update - May 23, 2025

INTRO
Welcome to This Week’s DXP Catalyst Update

Can’t believe we’re heading into Memorial Day weekend already. It’s been a busy stretch at DXP Catalyst, with a mix of events and client work this week.

On the small business side, I attended the Greenwich Village & Chelsea Chamber of Commerce’s 2nd Annual Changemaker Award event Tuesday night, recognizing local leaders from across the downtown NYC business community. On Wednesday, I was out in New Jersey running a HubSpot Sales and Marketing Hub workshop for a financial services client I’ve been supporting since last fall. Fortunately, the NJ Transit strike had just wrapped up, otherwise, the timing wouldn’t have been great! I’m also exploring the launch of a second newsletter later this year focused specifically on the SMB space, covering digital trends and how we’re engaging with local companies.

On the enterprise consulting front, I’ve been thinking more about conversational search, Search Generative Experience (SGE), and Generative Engine Optimization (GEO). I recently spoke with the leadership team at Squiz about the conversational search capability they’re rolling out soon as part of their DXP. It builds on their Funnelback search engine and reflects the broader evolution of how users discover and engage with content.

So, for this week’s edition of The DXP Catalyst Update, I wanted to explore this topic further - what conversational search means, how it works, and what organizations should be doing now to prepare their content and systems.

LEADERSHIP GUIDANCE
Is Your Content Ready for Conversational Search?

Preparing for the next Wave in AI-Powered Experiences

Search is changing fast. For years, users were expected to adapt their queries to match how systems worked. They typed in keywords, scanned link-heavy results pages, and hoped to find what they needed. That model is giving way to something more intuitive. Thanks to advances in natural language processing and generative AI, systems are now expected to understand the user and deliver direct answers, not just links.

This shift has major implications for how organizations structure, write, and manage their content. It’s no longer enough to have information buried in long-form pages or PDFs. Content needs to be clear, concise, and structured in a way that machines can parse and respond to. That includes proper use of headings, metadata, and formatting that supports both findability and interpretability.

And it’s not just about improving search. Conversational interfaces, AI-driven chatbots, and virtual assistants all rely on the same content foundation. If your information isn’t accurate, consistent, and accessible, the experience breaks down regardless of how advanced the front end may be.

The goal isn’t just to provide answers quickly. It’s to provide answers that are useful, trustworthy, and aligned with the user’s intent. Getting there starts with how you approach content.

From Keywords to Conversations

Traditional on-site search relies on keyword matching. Users enter a few terms, see a list of links, and have to dig through pages to find what they need. Conversational search changes that experience by allowing users to ask questions in natural language and get direct, often personalized, answers. This can happen through a chat-style interface, a smart search bar, or even voice input. The goal is to reduce friction and deliver useful information quickly.

This shift is part of a broader evolution known as Search Generative Experience, or SGE. Still in experimental rollout across many platforms, SGE uses generative AI to produce summarized answers based on a range of indexed content. It allows for follow-up questions, references source material, and in many cases, removes the need to click through traditional search results.

Alongside this, a new approach to content strategy has emerged: Generative Engine Optimization, or GEO. GEO focuses on making sure your content can be accurately understood, interpreted, and surfaced by AI systems. While SEO is about ranking in search results, GEO is about ensuring your information is included in the answers that generative tools present to users.

GEO and SEO are not mutually exclusive. Clear, well-structured, and authoritative content still supports both. But GEO adds new layers of importance to things like conversational tone, clarity of entities and concepts, and the ability for AI models to accurately summarize your content. It also raises the stakes in how your brand is represented. If AI systems are answering user questions directly, your goal is to make sure those answers reflect your voice, your values, and your facts.

Why It Matters

Conversational search doesn’t just enhance the user experience. It changes what people expect from digital interactions. When someone can ask a question like “How do I get started with your service?” and instantly receive a clear, personalized answer, they’re far more likely to take action. This kind of experience removes friction, increases conversion rates, and reduces the volume of routine support inquiries.

It also gives your team more direct insight into what users are actually trying to do. Traditional search data forces you to guess at user intent based on fragmented keyword inputs. With conversational queries, you get full questions in natural language, which makes it easier to spot content gaps, streamline navigation, and address common points of confusion.

But for any of that to work, the underlying content needs to be accurate, consistent, and structured in a way that supports direct answers. That’s where many organizations fall short.

Where Most Organizations Fall Short

AI-powered search is only as good as the content and systems behind it. This holds true for both traditional keyword-based tools and newer generative AI models. For systems that generate summaries or direct answers, the stakes are even higher. If the underlying content isn’t clean, current, and well-structured, the experience falls apart.

Many organizations are still working with a patchwork of legacy content, siloed systems, and inconsistent standards. The result is a digital ecosystem where even basic questions can be difficult for AI to answer reliably.

Common issues include:

• Duplicate, outdated, or conflicting information scattered across departments and platforms

• Unclear, overly promotional, or jargon-heavy writing that obscures direct, actionable answers

• Missing or inconsistent metadata, making it difficult for search systems to index and prioritize content

• Use of internal terminology or acronyms that external users don’t understand

• Content locked away in PDFs, intranets, or databases that lack proper structure or accessibility

• Outdated content that hasn’t been reviewed or refreshed in years, leading to inaccurate or irrelevant results

Even if your organization has the right information, if that content is fragmented, misaligned, or buried in inaccessible formats, your AI-powered search tools will struggle to find it, interpret it, and deliver it when your users need it most.

How to Prepare Without Waiting for the Tools

Even if you’re not deploying conversational search yet, there’s a lot you can do now to get your content and data ready. These steps will also improve your overall digital experience, regardless of the platform.

1. Run a content inventory

Map out where your key user-facing content lives. This includes CMSs, PDFs, help centers, knowledge bases, and even spreadsheets. Identify your primary sources of truth and flag duplications or inconsistencies. Bring in stakeholders from different teams to make sure critical materials aren’t missed.

2. Standardize language and terminology

Inconsistent naming creates confusion for both users and AI systems. If “clients,” “members,” and “users” all mean the same thing, choose one term and apply it consistently. Use plain language that reflects how your audience actually speaks.

3. Audit metadata and tagging

Systems rely on metadata to retrieve and prioritize content. Clean up titles, descriptions, tags, and categories. If applicable, align this work with structured data practices like schema markup to improve clarity and visibility.

4. Rewrite for clarity

AI performs better when content is straightforward. Use bullet points, short paragraphs, and headers. Avoid vague language or overly branded copy. Focus on writing as if you’re answering a real user question clearly and directly. 

5. Simulate real queries

Ask your website the kinds of questions your users might actually type. For example, “How do I get started?” or “What’s the difference between plans?” See if you can find the answer quickly. If not, that’s a signal your content needs work.

6. Prioritize high-impact content areas

Start with the pages users rely on most. These often include:

• Help and support content

• Product or service comparisons

• Onboarding instructions

• Contact or location information

• Step-by-step process explanations 

These are high-intent interactions where unclear or missing answers often lead to frustration or lost business.

7. Document and maintain

As you go, keep track of decisions like terminology standards, metadata conventions, and content ownership. Clear documentation makes it easier to maintain quality over time and onboard new contributors.

Start Small, Build Smart

One of the biggest mistakes organizations make is trying to apply conversational search everywhere, all at once. A better approach is to start with a pilot.

Choose one focused use case. Audit the content supporting it. Align terminology, add metadata, and write with answerability in mind. Then test, monitor, and iterate.

Use search analytics and conversation logs to identify gaps, unclear responses, or misunderstood queries. Let the data guide your next set of improvements.

Think of this like building a content product. It needs owners, processes, governance, and iteration cycles - not a one-time launch.

Where Squiz Fits In

One vendor making significant moves in this space is Squiz, whose upcoming conversational search feature is built on top of their enterprise-grade Funnelback search engine. With more than two decades of development, Funnelback provides robust multi-source indexing, customizable relevance tuning, and deep integration capabilities. It serves as a solid foundation for AI-enhanced search experiences.

Squiz’s conversational search isn’t just a chatbot interface layered on top of search. It applies a retrieval-augmented generation (RAG) framework, where a large language model (LLM) only generates answers based on content that has been retrieved and approved from your organization’s own indexed sources. This ensures that responses are grounded in truth and reflect what your content actually says, rather than information the model may have learned from elsewhere.

If the system cannot find high-confidence material to support a response, it doesn’t attempt to guess. Instead, it can fall back to traditional search results or prompt the user for clarification. This fallback helps ensure the system only responds when it has reliable content to draw from, reducing the risk of inaccurate or misleading answers.

Governance is also a central feature. Organizations can define which content sources are accessible, simulate how the AI would answer common queries before going live, and monitor actual interactions to detect content gaps or problematic patterns. Squiz incorporates a validation step to ensure answer accuracy before natural language responses are returned to the user.

For organizations concerned about content accountability, Squiz’s approach offers a high degree of control, visibility, and confidence as they explore or expand into conversational search.

Final Thoughts

It’s easy to assume you’re not ready for AI-powered search because the technology isn’t in place yet. But in reality, the bigger challenge is almost always the content. Tools can be purchased and configured in weeks. Getting content ready takes months, sometimes longer.

If your digital experience strategy involves any kind of search, support, or self-service functionality, now is the time to get ahead. Treat content like infrastructure. Build processes that ensure clarity, consistency, and accuracy. And when your platform partner introduces the next generation of search, the real differentiator won’t be the tool you choose. It will be the quality of what you feed it.