Is your website a help or a hindrance to your customer service? (feat. Contentful)

Episode topics

• 00:00 – Is your website a help or a hindrance to customer service?
• 01:27 – Introduction: Questions for now by TELUS Digital
• 01:40 – Why does self-service fail for web visitors?
• 02:35 – What is composable content architecture and personalization?
• 04:11 – How does agentic AI bridge contact centers and digital teams?
• 05:37 – How do you sustain disciplined content models over time?
• 06:27 – What are the most common mistakes in content personalization programs?
• 08:09 – How do you align business intent with customer intent?
• 09:24 – What is the operational and CX impact of unclear web messaging?
• 10:50 – How do you drive operational efficiency and team autonomy?
• 12:00 – How does personalization shift across marketing and service?
• 13:09 – How does modular content enable localization and multi-channel flexibility?
• 15:04 – How do you leverage first-party data and behavioral signals?
• 15:48 – How does agentic AI democratize personalization?
• 16:38 – How do you measure the ROI of personalization as a CX outcome?
• 18:33 – Proactive CX strategy or reactive problem-solving?
• 20:07 – Is your content clear enough to pass the AI litmus test?
• 21:40 – Conclusion

Transcript

[00:00:01] Robert Zirk: Right now, a customer is looking for an answer on your website. If they find it easily, great. You've anticipated a need and delivered a low-effort experience. If they don't, it likely means a call to your support team or a newly opened ticket.

[00:00:18] This episode is all about the unsung role your website is playing in your support experience.

[00:00:23] Andy Kaiser: If the experience is clear, relevant, easy to navigate, the customer moves towards their conviction, towards their goal. And if it's confusing, generic or inconsistent, that's a friction and they carry that friction to every next interaction.

[00:00:36] Robert Zirk: Some customers won't call at all. They'll take their questions straight to an AI chatbot, like ChatGPT or Gemini, and ask it to explain what your own website couldn't.

[00:00:47] Harry McIntosh: When you start quizzing AI about a product because it's not clear enough on somebody's website, you start to expose the problems pretty quick. Customers shouldn't have to do the work of going to an AI bot to understand a set of products. They should be able to find the content they need clearly on the website.

[00:01:03] Robert Zirk: So today on Questions for now, I'm joined by Andy Kaiser, vice president of product at Contentful, and Harry McIntosh, vice president of software engineering at TELUS Digital, as we ask: Is your website a help or a hindrance to your customer service?

[00:01:27] Welcome to Questions for now, a podcast from TELUS Digital where we ask today's big questions in digital customer experience. I'm Robert Zirk.

[00:01:36]

[00:01:39] Robert Zirk: A website is a digital front door to your company, and it serves many functions. An introduction to your products and services. A way to buy. A list of career opportunities on your team. For a customer with a question, it's often the first — and sometimes only — place they look for an answer.

[00:01:59] Gartner surveyed nearly 6,000 customers and found that only 14% of their service issues got fully resolved through self-service.

[00:02:08] Self-service involves a range of tools and resources to help customers resolve issues independently. While self-service isn't limited to a website, many of its common forms — like chatbots, knowledge bases and FAQs — are accessed through the web.

[00:02:24] That same Gartner survey found that 43% of customers cite not being able to find the content to address their problem as the reason self-service fails.

[00:02:35] Andy Kaiser is vice president of product at Contentful, a digital experience platform that helps companies deliver tailored content at scale. There, he leads the personalization and analytics side of the business — work that traces back to Ninetailed, the composable personalization company he co-founded before Contentful acquired it.

[00:02:55] Andy Kaiser: At Contentful, really my focus is on helping brands to create digital experiences that are more relevant and more measurable and easier to operate at scale. So a big part of my job is thinking about how teams move from static digital experiences to experiences that can actually adapt to customer context and, more importantly, prove business impact.

[00:03:19] Robert Zirk: Andy built Ninetailed around a specific idea: personalization needed to be API-first, composable and built around structured content.

[00:03:29] Andy Kaiser: What felt broken was that personalization was often treated like a thin layer on top of the website. It wasn't something foundational. What that means it's that it's not infrastructure.

[00:03:40] That also means that a lot of teams were trying to personalize with tools that were very channel-specific, so for example, only the website and also tightly coupled to a single CMS or to a single page or to a single content. This creates very fragile front-end workarounds.

[00:03:58] And at the end, this made it very hard to scale, hard to govern and even harder to extend beyond this one or two moments in the journey of the audience, in the journey of the users.

[00:04:11] Robert Zirk: Harry McIntosh is vice president of software engineering at TELUS Digital, where he supports a large group of engineers building apps and websites using martech platforms like Contentful, helping clients implement the same kind of composable architecture Andy just described.

[00:04:27] Historically, he's seen a similar tension play out among clients between the teams that run contact centers and the teams that build digital experiences.

[00:04:37] Harry McIntosh: It felt like it was a tug of war about who's supporting the customer. Of course, that's all internal politics, and that's not what customers want. They want the ability to get what they need to be done easily, whatever channel that makes sense in.

[00:04:48] Robert Zirk: Harry says that tension is starting to resolve because agentic AI is giving both teams a shared middle layer to work with.

[00:04:55] Harry McIntosh: Now, with the agentic world, I see all of these worlds coming together and a very harmonious outcome for both contact center CX groups and digital self-service groups being able to build complementary experiences where, actually, you're driving very rich conversations when you need them with a human.

[00:05:14] So a simple task could be done digitally. You've now got the agentic in the middle that can handle some of those "explain to me what's going on" type scenarios.

[00:05:23] And then for those really complex things, we've got humans being able to do the rich experience with our customers to help them get through really serious problems. And they're not waiting on hold as long because those more simple things are being handled across the other tiers.

[00:05:37] Robert Zirk: For Harry, that harmony depends on a strong operational model and the discipline to maintain a healthy system as part of the ongoing process.

[00:05:46] Harry McIntosh: The biggest gap between what clients look for doesn't happen in the initial build. It happens over time. Managing content models effectively and staying disciplined on a lot of heavy research upfront to identify IA...

[00:06:00] Robert Zirk: That's Information Architecture...

[00:06:02] Harry McIntosh: ...and content models and things like that to drive a very healthy system. And over time, these things start to decay because there's new products in the launch and so on.

[00:06:09] So I think learning how to get that discipline on keeping that content model rigorous rather than "can we just add a button here?" or whatever, you start to drift quite far away from the strategy. So I'd say it's about sustaining things and finding that operational model is some of the more challenging things for a digital operation.

[00:06:27] Robert Zirk: Andy's seen a version of that same discipline problem earlier in the process before a personalization program even gets off the ground. He pointed to two common mistakes:

[00:06:38] Andy Kaiser: I always say personalization is the holy grail of martech or marketing. And that's also why it has had a credibility problem. It's in my opinion, very attached to the most common failure mode on this and it's usually teams always start with the tooling instead of the problem. So they usually can, through their technology, for example, deliver the variants but they focus less on "Are we really creating meaningful customer value — how we measure it, how we operationalize this across content and teams?"

[00:07:10] And another common mistake is that usually organizations try to approach this as one big project that has to solve everything at once. This makes the program too heavy, too slow and sometimes very abstract. And what I have experienced in practice is that teams that really succeed are the ones that start with a clear use case, learn quickly, prove value, and build from there. So it's really about a lot of velocity learnings and agility. around this.

[00:07:39] Personalization is a capability you develop through iterations and not just a single transformation that is completed in one step. So the technology works, but the system around it doesn't and this is also why the content isn't differentiated enough. The audience logic usually also isn't tied to a customer need and sometimes also the measurement framework can't isolate the impact. So personalization only earns credibility when it really produces outcomes people can see and trust.

[00:08:08] Robert Zirk: To get to those outcomes, Harry says personalization has to start with matching business intent to customer intent.

[00:08:16] Harry McIntosh: I find when people lead with just, "I want to personalize a website," but don't have a clarity of goal, we can meander down all sorts of paths and not have focus. The strongest opportunities in personalization are where you have high intent as a business, and that matches the clear intent from a customer and you're marrying the two together.

[00:08:35] Robert Zirk: For Harry, that clarity comes long before he ever brings up the technical side of composable architecture with a client.

[00:08:42] Harry McIntosh: I think you have to work your way towards that over time by starting with client needs and the user base. What do we know about them? What are their needs? Why are they coming to a digital experience? Or why are they seeking service in general? 'Cause it could be about the contact center as well, about when they're asking there.

[00:08:59] So understanding everything from a needs-based perspective, prioritizing the content, going through IA exercises, and that's how you start to break the problem apart, and then you can get into content hierarchy and that leads you finally to content models. I like to let the business operations talk about the customer first and all that lead the way, and then we only bring in the technical pieces to the people that really need to know those components.

[00:09:24] Robert Zirk: I asked Andy what it costs a support team when the website gets it wrong.

[00:09:28] Andy Kaiser: I don't have a number on this. I would say it's a meaningful share. And of course, not every contact is caused by content interactions, but what we're seeing is that it's a surprising amount of avoidable volume.

[00:09:40] We see it even in our own support processes, et cetera, that everything starts really with unclear messaging as I said, also fragmented journeys and digital flows that break the customer confidence. And at the end, the support organization ends up absorbing all that friction that was introduced much, much earlier.

[00:09:58] This cost or this impact is sometimes very undercounted in a lot of decisions. If we look, for example, teams often evaluate web and content choices through the lens of technology. So, for example, how fast they can launch new features, also about ownership boundaries. But they don't always account for the impact on customer trust, service demand and experiences quality.

[00:10:21] Robert Zirk: Andy says the fix starts with ensuring CX leadership is in the room when those decisions get made.

[00:10:26] Andy Kaiser: It's really about having the seat on the table, but also understanding how, at the end, we are speaking about customer experiences and customer experiences happen in so many different touch points. It's not just from one part of the business responsible for it.

[00:10:40] Robert Zirk: Andy's right that CX spans a lot of ground, but for the support team specifically, that seat at the table is what keeps friction from landing entirely on their desk.

[00:10:50] Harry McIntosh: We've seen especially the way that visits to websites are collapsing. That also means that there's cost pressures on people that are owning websites and their web experiences. That means they need highly efficient operations to get things done and workflows for that.

[00:11:07] The more that we can empower the entire organization, the specialists to bring their content in from other departments into a system, but then bring web expertise to content optimization and end-to-end flows, then you're winning because you're driving a more efficient operation.

[00:11:24] Robert Zirk: Andy sees that same efficiency case when it comes to return on investment.

[00:11:28] Andy Kaiser: I would say there is also like a people lens to the ROI. So, when systems are really less tightly coupled, teams can work more clearly with their areas of expertise without constantly waiting for another to duplicate the effort. At the end, as content teams can move with more autonomy, engineers can focus on higher value work and CX and service teams, they can really rely on more reusable, trusted content across the journey they support. And this, at the end, does not only improve the speed. It's also about alignment, accountability and organizational ability to scale change.

[00:12:00] Robert Zirk: Harry views the marketing versus service split as a decision shaped by the kind of business you're running.

[00:12:06] Harry McIntosh: I think it depends on the maturity of the company. When you have a company that has a lot of contract-based work or engagements with customers, service is much more at the front. Our customers stay with us much longer than it takes to sign them up. Therefore, you see the content being prioritized towards that.

[00:12:25] In an organization where it's kind of one-time SKU purchases and things like that, it's gonna veer the other way. So it really depends on the client type.

[00:12:35] The other thing as well, though, for those ones that are more service-orientated and they do have that, this is when personalization plays a key role: being able to balance. Give people, if it's unauthenticated or If it's a first-time visitor, then you can prioritize and personalize that experience for marketing functions. But if we know they're returning visitors, we know they're authenticated or they've been before, then we can prioritize service-based messages for them.

[00:13:00] So really, again, context matters, and so personalization is what you really need to do to get away from the tension and just start investing in knowing your customer better.

[00:13:09] Robert Zirk: Andy made a similar case for why the same piece of content has to behave differently, depending on where it shows up.

[00:13:16] Andy Kaiser: When we speak about personalization, we think about the very complex omni-channel personalizations. But I also like to put examples like easy personalizations, which means, for example, localization. Because if you think about personalization and messaging, it's always about how we adapt the message to a specific audience.

[00:13:32] And if we have, for example, a specific audience which is from a specific country, we localize that content or we translate that content, which by the way, translation and localization is not the same. But the important part is about understanding the intent of that audience.

[00:13:45] And that's what really is about translating the message to that specific audience, to that specific need of that audience. So you could have some things like, I don't know, discovery of something and that's the different experience than somebody who's really looking into acquiring a product.

[00:14:02] Robert Zirk: Harry referenced offers as an example of how modular content can become a competitive advantage for personalizing the customer experience.

[00:14:10] Harry McIntosh: Sometimes you have even the same headline or the same name for the offer, but each channel needs a little bit of extra content.

[00:14:16] If you nail your content model, you present both. You've got one piece just for all the content, say the offer, and you've got your key hero message, your main banner, et cetera. But then you've got those more detailed messages that your internal team members might have and you haven't had to go through a duplication process for the rest of the pieces.

[00:14:34] Robert Zirk: Another example of where modular content can be beneficial is the fine print — the kind of detail that can quietly create a compliance problem if it drifts out of sync across channels.

[00:14:45] Harry McIntosh: How we treat that mouse type on the website is completely different how we treat it in an agent experience and present it, so using the same content. And further, that makes life a lot easier for your internal operations, talking to legal, getting it right once. You can have one content team deliver multi-channel as well.

[00:15:04] Robert Zirk: Once the content itself can flex across channels, Harry says the next question for CX leaders to ask who the content is for.

[00:15:13] Harry McIntosh: To me, the most effective thing is being able to drive people into an authenticated state because then you have first-party data available to drive at that. However, you don't need that. You can look at signals in what they're searching on, what they're clicking on, so that you have this concept of a first-time visit.

[00:15:30] And depending on where things go, you can start in their second visit, to prioritize different content. You're capable of doing that today. It'll improve conversion, but ultimately you want to drive them into that authenticated state because then you really know them deeply and that's when you get really high lifts on conversion.

[00:15:48] Robert Zirk: Earlier, Harry mentioned how agentic AI can act as the shared layer bringing contact centers and digital teams together. Andy zooms in on what this means for personalization.

[00:15:58] Andy Kaiser: It's like this continuous process, continuous iterations. It's all about adapting to the reality of the business. No business set up something and then succeeded for the next 30 years. It's always about adapting to a market, adapting to new business opportunities.

[00:16:12] And let's be fair — personalization can be complex, but that's also where the right technology makes this process easier. Now with agentic support, it becomes easier. And that also means that you don't need these super specialized teams. You need specialized agents that help you to succeed. And, at the end, that's also a democratization of personalization that hasn't happened and it's happening right now.

[00:16:38] Robert Zirk: I asked Andy how you know if your website's content and personalization are improving the support experience.

[00:16:44] Andy Kaiser: The clearest signal is reduced customer effort. But I would frame this also through the lens of content intent. At the end, it's really about understanding, for example, is the customer being met with the right content for what they're trying to do in that moment? Are they finding the right answers faster? So completing the tasks with less confusion and they are moving forward without needing to translate the experience for themselves. So if your people need assisted support for something that should have been resolved digitally or during the journey of the user, that's a strong sign that the content and personalization layer is doing real work for also the CX part.

[00:17:23] And from there, if you connect it to the service metrics, it's like this chain of cause and effect. So if content intent is aligned with customer intent, but the digital experience has become clearer, more relevant and easier to act on, that should lead to fewer avoidable contacts, more efficient resolutions when contact has happened and overall better satisfaction because customers feel understood early in the journey.

[00:17:48] Robert Zirk: In other words, don't measure your content and personalization efforts as distinct marketing metrics. Measure them as service outcomes in terms of how much friction your website takes out of a customer's day.

[00:18:01] Andy Kaiser: The question is really about how well does our website understand and serve the customer intent and where is it creating extra work for the customer instead? This question changes how you look at content architecture and personalization.

[00:18:15] So it's less about seeing the website as just a brand surface and it's starting to see it as a system. And this system should help customers to what they came to, and with as little friction as possible. So, content architecture and personalization, it's not just a brand surface, it's this operating model around the customer.

[00:18:33] Robert Zirk: When content is aligned with what customers want or need, the customer experience as delivered through the website builds trust and reduces customer effort.

[00:18:43] Harry says sustaining lower service demand more broadly means resisting the pull to just react to whatever's driving calls in a given week.

[00:18:52] Harry McIntosh: I find often organizations will come forward and say, "Look, this is what's driving the demand. These are the calls I'm receiving. This is what everybody's saying." They're not always the same. It's not a one-to-one mapping of just solve that and then you change things. That is like a Whac-A-Mole approach.

[00:19:08] I see high intent across the world already in this. What it is is about being able to understand that if you just play the Whac-A-Mole game, you won't get very far very quick. You have to go with that more broad thinking of let's understand our users better about their intent so that it doesn't happen in the first place.

[00:19:25] Robert Zirk: He used billing cycles for a hypothetical company as an example.

[00:19:29] Harry McIntosh: Customers start contacting them all the time saying, "I don't understand my bill." If they don't understand their bill, we can quickly start developing solutions that are like, "Okay, let me explain your bill online," and so on. But we've actually missed the point. We're too late in the process at that point.

[00:19:42] There's actions that we can do much more proactively up front with the effect of digital messaging, with the right digital communications and personalization that gets people ready for that bill. And it's finding that line to get that strategy beyond just "I'm seeing this signal in my system that's saying I have a customer problem," to looking at it more broadly and rather than just "how do I help them in that moment?", it's "how do I get to a proactive state of helping them earlier?"

[00:20:06] Robert Zirk: To wrap up, Harry offers a litmus test for whether your website's content is actually doing its job.

[00:20:13] Harry McIntosh: I see so many times where — and it's so painstakingly obvious nowadays with LLMs and all that stuff, that when you start quizzing AI about a product because it's not clear enough on somebody's website, you start to expose the problems pretty quick that, customers shouldn't have to do the work of going to an AI bot to understand a set of products. They should be able to find the content they need clearly on the website.

[00:20:37] So I think that's the question to ask yourself: is my content clean and clear enough on the website that somebody doesn't need to go to ChatGPT or the likes?

[00:20:48] Robert Zirk: If you're looking to get to a confident answer, TELUS Digital helps brands build the content architecture on Contentful that gets them there, whether it's building composable experiences from scratch or scaling existing instances.

[00:21:01] As a Gold Solutions Partner that holds more than 20 certifications across technology, content management, studio and personalization, we've helped dozens of Fortune 500 companies implement Contentful and maximize their use of the platform. For one e-commerce client, a rebuild that used Contentful as part of the stack led to a 185% revenue increase within six months.

[00:21:27] You can learn more about TELUS Digital's Contentful practice by visiting the link in our show notes.

[00:21:40] Thank you so much to Andy Kaiser and Harry McIntosh for joining me and sharing their insights today. And thank you for listening to Questions for now, a TELUS Digital podcast.

[00:21:51] If you enjoyed today's conversation, subscribe to Questions for now on your podcast player of choice to get the newest episodes as soon as they're released.

[00:22:00] I'm Robert Zirk, and until next time, that's all... for now.