Let me tell you two stories about the same prospective student in 2026.
Story One:
Thabo is researching computer science programmes in Johannesburg. He opens ChatGPT and asks: “What are the best computer science programmes in Johannesburg?”
ChatGPT gives him detailed information about three universities. Comprehensive descriptions. Entry requirements. Career outcomes. Everything he needs to make a decision.
But here’s what Thabo doesn’t know: Two of those universities’ information came from their competitors’ websites, because the AI couldn’t find clear, structured information on theirs.
The story ChatGPT told about their programmes? Generic. Could be any university. And not quite accurate.
Story Two:
Same year. Same student. Same question.
But this time, when ChatGPT searches, it finds your website. The information is structured, specific, authentic.
It cites your actual research breakthroughs. It mentions Professor Ndlovu by name and her work on AI for South African retail. Likewise, it describes the exact facilities in your Innovation Hub.
Thabo clicks through to your site. What he finds matches what the AI told him. But it’s richer—more human, more compelling. Real students in videos. Specific stories about real projects. Clear pathways to apply.
He starts his application that day.
The Dual Challenge Nobody’s Talking About
I opened with a higher education example because that’s where this experience is rooted, rebuilding a major university website that won the highest-scoring award in South African higher education in 2016. But this challenge isn’t unique to universities.
If you’re managing a website for a healthcare system, a large nonprofit, a government agency, or any organization with complex user journeys and multiple stakeholder groups, you’re facing the exact same problem:
Your website now serves two completely different audiences, each with fundamentally different needs.
The first audience is human—patients seeking care, donors wanting to support your mission, citizens accessing services, customers researching solutions. They need warm, authentic, helpful experiences that make them feel connected to your organization.
The second audience is algorithmic—Google’s search engine, ChatGPT, Perplexity, Claude, and every other AI system that’s crawling your site daily, trying to understand your content and deciding whether to cite you or send traffic elsewhere.
And here’s the problem: Most organizations are building for neither audience well.
They’re deploying AI chatbots without asking if they actually serve user needs. They’re generating generic AI content because “something needs to be there.” They’re making their websites less human in the name of being more “tech-forward.”
This is backwards.
Journey Mapping: The Foundation That Never Changes
Ten years ago, when I led the rebuild of a major university website, we started with a principle that won us the recognition of the best website award in South African higher education marketing and communications:
Understand user journeys before deploying technology.
That principle still works. In fact, it’s more critical now.
Here’s why: When you start with technology—whether that’s a new chatbot, an AI content generator, or a fancy interactive feature—you’re solving for the tool, not the user’s problem.
But when you start with the journey, you can ask the right questions:
- Where does this user start their interaction with us?
- What obstacles do they hit?
- Where do they give up or have to contact us for help?
- Would AI smooth this friction or add complexity?
The Framework: Three Questions Before Any AI Implementation
Before deploying any AI technology on your website, ask:
1. What specific user problem are we solving?
Not “we want to be innovative” or “everyone else has a chatbot.” What actual friction point in the user journey will this remove?
2. For human users: Will AI smooth this friction or add complexity?
Will it genuinely help them get to their goal faster, or will it create a new layer they have to navigate through?
3. For AI systems: Can they understand and surface this journey effectively?
Is your content structured so that when ChatGPT or Google scans your site, they can accurately represent what you offer?
Case Study: The Course Finder That Anticipated User Needs
In 2015, we faced a challenge when it came to rebuilding the Information Architecture for the Wits website: 800+ academic programmes across multiple faculties. Prospective students were overwhelmed.
In those days, we couldn’t have solved this with a chatbot. We couldn’t have used AI to generate descriptions. Instead, we mapped the actual user journey:
User Goal: “I’m interested in technology and business. What can I study at Wits?”
Friction Point: Traditional A-Z course lists made them read through hundreds of options, most irrelevant to their interests.
Solution: A Course Finder that let them search by interest area AND course type, narrowing 800 programmes down to 10-15 relevant options immediately.
Human Benefit: Found relevant programmes in 2 minutes instead of 20.
This is an example of how we used structured data that search engines could understand and surface accurately.
That’s the power of starting with the journey, not the technology.
The Same Principle, Different Context
This isn’t unique to universities. For example, a regional healthcare system faces a similar challenge: hundreds of clinics, specialists, and services. Patients are overwhelmed.
User Goal: “I need a cardiologist who accepts my insurance and is located near my work.”
Friction Point: Traditional provider directories made them scroll through alphabetical lists, click into individual profiles, check insurance separately, and we’re looking at 10+ minutes of frustration.
Solution: A provider finder that filters by speciality, location, insurance, and availability simultaneously. Five questions, 30 seconds, three relevant options.
Human Benefit: Find the right doctor in under a minute instead of giving up and calling the main number.
AI Benefit: Structured data that search engines could surface when patients searched “cardiologist near me accepts BlueCross.”
The technology differs. The principle is identical: Understand the journey before deploying the solution.
The Mistake I’m Seeing Everywhere
Organizations across sectors are adding AI chatbots to their homepages asking “How can I help you?”, without fixing the underlying information architecture.
If your chatbot is answering the same basic question 500 times a day, that’s not success, that’s a symptom that your information architecture is broken.
The chatbot becomes a band-aid over a structural problem. You’re forcing users to ask an AI for information that should be immediately visible and accessible.
Fix the IA first. Then, if there are remaining complex queries that genuinely benefit from conversational assistance, add the AI.
The Dual Audience Question
Here’s the litmus test for any website change you’re considering:
Does this serve both human visitors AND AI systems well?
Good AI use—serves both:
- Structured data (schema.org markup) on programme pages
- Human benefit: Clear, consistent information presentation
- AI benefit: Can accurately parse and cite your content
- AI-powered course recommender based on interests and desired qualifications
- Human benefit: Faster, more relevant results
- AI benefit: Structured queries and responses for search engines
Bad AI use—serves neither:
- Generic chatbot on homepage with no clear purpose
- Humans get frustrated with vague responses
- AI systems can’t index chatbot conversations
- AI-generated programme descriptions with no institutional specifics
- Humans recognize it as generic and distrust it
- AI systems rank it lower because it lacks expertise signals
If You Do One Thing This Week: Map ONE Journey
You don’t need to overhaul your entire website. Start small.
Step 1: Pick ONE primary user journey
Examples across sectors:
- Higher Ed: “Prospective student wants to find programmes matching their interests and qualifications”
- Healthcare: “Patient wants to book an appointment with a specialist who accepts their insurance”
- Nonprofit: “Potential donor wants to understand impact of contributions and how to give”
- Government: “Citizen needs to renew a license and understand requirements”
Step 2: Map the current journey honestly
- Where do they start? (Google search? Homepage? Social media?)
- What obstacles do they hit?
- Where do they give up or contact you for help?
- How many clicks to accomplish their goal?
Step 3: Ask the dual-audience question
- Would AI help humans get to their goal faster?
- Can AI systems understand and surface this journey accurately?
Step 4: Fix the information architecture FIRST
Before adding any new technology, make sure the information is logically organized, easily findable, and clearly presented.
The Path Forward
The organizations winning Story Two, where users find accurate, compelling information and take action, aren’t necessarily spending more money or deploying more AI tools.
They’re starting with strategy. With user journeys. With understanding what problems they’re actually solving.
AI is a powerful tool for enhancing these journeys, for transcribing interviews with subject-matter experts, for structuring content, for helping users navigate complex information. But it’s a tool in service of the journey, not a replacement for understanding it.
The organizations that thrive in the AI era will be those that use technology to amplify authentic human connection, not replace it.
Which story will your organization live?
Want to Go Deeper?
At ActiveLink and addGlue, we help organizations with complex user journeys develop AI-ready marketing strategies that keep the human in the loop. We work with higher education institutions, healthcare systems, nonprofits, government agencies, and enterprises. We start with journey mapping, move through content strategy, and ensure your website serves both human visitors and AI systems effectively.
Next week: I’ll share the 15-minute interview technique that extracts authentic stories from busy subject matter experts—no writing required from them.
This article is part of a multi-week series on building websites for the AI era, drawing on my presentation at the MACE 2025 Conference. Whether you’re in higher education, healthcare, nonprofit, government, or enterprise—if you serve multiple audiences with complex journeys, this series is for you.
Tags: #AIStrategy #WebsiteStrategy #UserExperience #JourneyMapping #DigitalMarketing #AIReadyMarketing #ContentStrategy
About Jonathan Taylor: Jonathan is the founder of addGlue, an AI-focused storytelling consultancy, and ActiveLink Media. He specializes in helping organizations with complex user journeys navigate the intersection of AI technology and authentic human storytelling. In 2016, he led the rebuild of the University of the Witwatersrand website, which won the MACE Platinum Award of Excellence and Chairperson’s Award—the highest-scoring entry across all divisions. He writes on marketing strategy, AI, and digital experience at activelinkmedia.co.za/hype.
The views expressed herein are personal and do not reflect the views of our clients or partners. Some content links may be affiliate links and are an endorsement of the product or service.
Photo by Hassan Pasha on Unsplash
Original article: https://www.activelinkmedia.co.za/why-your-website-is-playing-two-games-at-once-and-losing-both/