Part 1: Website Renewal in the AI Era — Before Rebuilding, Ask Which Business Challenge You’re Solving

When we receive inquiries about a website renewal, the conversation usually starts from a desire to change or refresh the current site’s design, functions, content, and so on. That’s a very natural place to begin. In other words, it’s an attempt to put into slightly more concrete words the gap between the current website and the ideal role it should play within the business.

Because a website is one of the touchpoints with customers and stakeholders, any change to it is, ultimately, one of the means for solving some kind of management challenge.

And now, AI is beginning to change the meaning of this “website as a means” from the ground up. In this article, we’d like to talk—across two installments, Part 1 and Part 2—about a way of thinking about renewals that does not start from “rebuilding.” In Part 1, we’ll go as far as the point where we recognize that even the way we frame the question itself has changed.

Don’t start from “rebuilding”

When you consider a renewal, the first question you should place on the table is not “What kind of website should we make?” It’s “Which management challenge do we want to solve?” First comes the management challenge; then there are multiple possible means to address it; and among them, the website is one touchpoint. If you skip that order and jump straight into “Let’s rebuild it anyway,” the means ends up hijacking the goal.

In many companies, renewal is discussed as a binary: do we do it or not? But we want to reframe it as: “For this management challenge, which touchpoint should we invest in, and how much?” That’s why, if a different touchpoint is the one that will work best, even the conclusion “We won’t invest in the website” can be the right answer. Why not start by looking from one level higher?

What AI changed: the baseline for production cost

So why is this rethinking necessary now? Because AI has lowered the “baseline for production cost.” Still, we should choose our words carefully here.

For a level of production where “it’s enough to have something” or “we just need a site,” the cost has effectively fallen close to zero. In the past, there was always a minimum production cost that served as the base; that base is no longer a base. But we must not misunderstand: the cost of “building a site that moves management challenges” has not become zero. What has become zero is the commodity portion—parts that anyone can make, that become standardized and therefore don’t create differentiation.

If so, the conclusion actually points in the opposite direction. Being able to build something is no longer value in itself. Precisely because anyone can have “something that merely exists” for free, the value of the judgment that decides “what to build” rises sharply. Not “So we can build it cheaply,” but “What used to require money has become free—therefore we should invest in deciding what to build.” That reversal, we think, should be the starting point for renewals going forward.

The game isn’t “building cheaper,” it’s designing the return

Let’s talk a little arithmetic. ROI is Return (R) divided by Investment (I). With AI, the “just build it” portion of I is dropping out of the baseline. If you look only at the division, ROI can improve without limit.

But there’s a pitfall. Making I smaller is not, by itself, the purpose. No matter how close you push it to zero, if the absolute amount of R is small, it means nothing. You fall into a “poor optimum” that makes the numbers look nice by shrinking the denominator while moving management not even a millimeter. That’s why the rules change—from “How do we save on I?” to “How do we design R?”

Here, one important note: this R is not “the number of conversions or visits on the site.” The origin of R is, fundamentally, the amount of management challenge solved. Website metrics are only proxy signposts toward that goal. Until now, production cost consumed a large share of the budget, so there was a limit to how much could be allocated upstream—persona design, customer journey work, role definition, and so on. Now that production cost has fallen out of the baseline, we can allocate boldly upstream. But if you get lured by the lower cost and start economizing on upstream thinking as well, R will not grow. Where should we place our bets, and how much? We think we’re entering an era where that discernment is what gets tested.

Summary, and toward Part 2

In summary: a website is one means for solving a management challenge, and the first question in a renewal is not “What kind of site should we make?” but “Which management challenge are we solving?” What AI lowered is the production baseline for the “good enough to have” level—it did not erase the cost of building a site that moves management challenges. Therefore, the battleground has shifted from making I cheaper to designing R.

So how do we design that R? The key lies in what role we assign to each touchpoint. In Part 2, we’ll redefine that role as “moving someone from one state to another,” and we’ll talk about the idea of focusing all the way on the bottleneck—along with what an appropriate allocation looks like, and how to proceed. If you’d like, please join us for Part 2 as well.

Next preview: Part 2: Website Renewal in the AI Era — Defining the Role of Digital Touchpoints and Focusing on the Bottleneck

In Part 2, we’ll walk through five steps for designing R. We’ll translate management challenges into the language of “states,” identify where the journey is getting stuck, and explain why we sometimes decide not to build at all. Coming soon.