Founder and CX leader says customer-facing AI fails when companies ignore the journey around it
Jaco van den Heever
CEO of Sand Dollar Design
Jaco van den Heever explains why AI improves CX when it reduces journey friction, strengthens employee workflows, and leaves judgment to humans.
Key Takeaways
Core View: Effective CX uses AI to reduce effort and improve journeys, not merely add technology to polished interfaces.
Human Judgment: AI can analyze patterns and prepare context, but people must own empathy, ethics, prioritization, and sensitive decisions.
Earlier Testing: AI-enabled prototyping lets teams test realistic product experiences sooner, improving feedback before expensive development begins.
Context First: Strong context, structured information, and clear constraints improve AI output more reliably than prompt wording alone.
Employee Experience: AI-augmented CX improves when internal workflows help employees access context, reduce administration, and serve customers faster.
Jaco van den Heever is a serial founder with a background in digital customer experience. He is currently CEO of Sand Dollar Design, a remote UX strategy, product design, and AI development consultancy.
We sat down with Jaco to learn how he’s using AI to augment CX workflows — and how applying AI to EX can improve CX.
CX is about more than front-end polish
I’m Jaco van den Heever, founder and CEO of Sand Dollar Design. I’ve spent more than 20 years working at the intersection of digital customer experience, product strategy, UX, and technology. Before starting Sand Dollar Design, I led digital and omnichannel product work across mobile apps, websites, and service platforms. This shaped my view of CX: It is not only about front-end polish, but about whether the product, service, and operating model work for users.
At Sand Dollar Design, we help startups, SMEs, and enterprise teams improve digital products and customer journeys through UX strategy, UX research, product design, and increasingly, AI-enabled MVP and product development.
Our channels include websites, mobile apps, customer portals, B2B platforms, internal tools, dashboards, and workflow systems. Our clients are typically founders, product leaders, innovation teams, digital transformation teams, and enterprise leaders who need to reduce friction, improve adoption, and make better product decisions before committing heavily to building.
We work globally, with experience across South Africa, the USA, the UK, and Europe, and have delivered work in complex sectors including financial services, healthcare, telecoms, B2B commerce, and enterprise software.
More Articles
How AI can be used to ease the customer journey
I'm moving AI much earlier into the customer journey, especially during onboarding and product discovery.
A good example is GigExecs, a platform I cofounded for experienced professionals looking for fractional, interim, and project-based work. We saw a friction point: senior professionals often have deep experience, but turning it into a clear, searchable, compelling profile takes effort. People either over-explain, undersell themselves, or delay completing the onboarding flow because the blank page feels like work. That affected their experience.
So, we designed the experience around AI-assisted profile creation. Instead of expecting users to manually write everything from scratch, the AI-assisted flow interprets a CV, LinkedIn profile, or conversation-style input, then turns it into clearer positioning, profile content, and structured data.
The important change was not simply “adding AI.” It was redesigning the journey around a lower-effort path to value.
This changed how I think about AI in CX more broadly. I'm increasingly looking for points in the journey where users get stuck, feel unsure, become overloaded, or must translate their needs into a system’s language. These are often the best places for AI to help.
The result is a more practical view of AI in customer experience. I am less interested in AI as a novelty layer and more interested in it as a way to reduce effort, improve relevance, and help people move through a digital journey with more confidence.
Which CX tasks should be done by AI
I now rely on AI to inform or power several CX activities:
- Synthesizing user research
- Identifying recurring friction points
- Prioritizing issues from support or feedback data
- Generating first-draft content
- Speeding up prototyping
- Creating more personalized journey options
User research synthesis stands out. In a typical research project, we collect large amounts of qualitative data: interview notes, transcripts, survey responses, usability findings, stakeholder input, and journey observations. In the past, synthesis was a manual process. We still do human analysis, but AI now helps us find patterns much faster. It helps us cluster themes, spot recurring friction points, compare feedback across user types, and identify signals for further investigation. This means we spend less time organizing data and more time asking harder questions: what does this mean, what should the client do next, and which product decisions need to change?
Which CX tasks should be done by humans
The activities I keep explicitly human are the ones involving judgment, trust, and accountability.
For example, journey design still needs human judgment because it involves trade-offs between customer needs, business goals, operational reality, and technical constraints. Service recovery should also remain human-led when there is frustration, financial risk, medical sensitivity, reputational impact, or emotional context. AI can surface the issue and prepare the context, but a person should decide how to respond.
The same applies to product strategy and prioritization. AI can help us see patterns faster, but it should not decide what matters most. That still requires understanding the customer, the business model, the brand promise, and the consequences of getting it wrong.
So my view is simple: AI should reduce effort, improve context, and speed up learning. Humans should own the decisions where empathy, ethics, trust, and commercial judgment matter most.
Jaco States
…my view is simple: AI should reduce effort, improve context, and speed up learning. Humans should own the decisions where empathy, ethics, trust, and commercial judgment matter most.
How AI prototyping improves the quality of user feedback
One big advantage of AI can be seen in prototyping. Previously, a typical process moved from wireframes to high-fidelity designs, then clickable prototypes, then developer handover. Now, through AI-enabled product development, we can often put a working prototype in front of users much earlier. This prototype may include real interactions, realistic data, and enough functional logic for users to respond to it as something closer to the finished product.
This improves the quality of the feedback. People behave differently when using something that feels real. And that means we can better understand real product behavior before completing a full build.
But the downside is that AI can also create false confidence. Teams can build faster than they can think. Code still requires review. Product decisions still require evidence. Research synthesis still requires human judgment. A polished prototype can make a weak idea look stronger than it really is.
So the best result has not been “AI replaced the team.” Instead, AI has compressed the distance between insight, prototype, testing, and iteration.
Why customer-facing chatbots are missing the mark
My clearest disappointment is customer-facing chatbots in large corporate environments, especially in financial services, insurance, banking, and fintech.
Today's AI capability far exceeds what many customers experience.
In theory, customers should have a natural conversation with a digital assistant. It should understand intent, remember context, ask useful follow-up questions, access the right information, explain options clearly, and hand over to a human when an issue becomes sensitive or complex.
In practice, many corporate chatbots still feel like old decision trees with a new label. The bot forces customers into a narrow set of choices, it fails as soon as a question becomes slightly unusual, and users often try to escape the bot rather than get help from it.
The issue is usually not the model itself, but the surrounding experience. Data is fragmented. Integrations are incomplete. Escalation rules are weak. Companies have not designed the bot around the real customer journey. And in regulated industries, risk and compliance concerns often lead companies to deploy something safer, but much less useful.
This creates a strange gap. Customers hear a lot about AI, but then interact with systems that do not feel intelligent at all.
Tools every CX team should consider using
CX teams can benefit from the following tools:
- ChatGPT and Codex for analysis and planning
- Claude Cowork for complex task support and building knowledge repositories
- Cursor for AI product development and automation
- Qualtrics for customer feedback
Why CX leaders must invest in context first
The quality of AI's output depends as much on the context you give the AI as on the model itself.
Initially, it's easy to think the skill is simply “writing a good prompt.” That matters, but it's only part of it. I've learned that AI performs much better when you give it structured context: the goal, the customer problem, the business rules, the tone, the constraints, the examples, the data sources, and what a good answer should look like.
That changed how I now set up AI work. I use more structured prompts, markdown files, project folders, and reusable context documents so the AI doesn't start from scratch every time. This gives the work more consistency and makes the output much more useful.
My advice is to treat your first AI pilot less like a one-off experiment and more like the beginning of a new operating system for how your team works. The better you structure the context, the better the AI becomes as a real partner in the workflow.
Why AI-augmented EX benefits CX
More leaders should redesign employee experience, especially the internal systems and workflows staff rely on to serve customers.
Many CX problems do not start at the customer-facing touchpoint. They start behind the scenes, where employees try to deliver a good experience while dealing with fragmented systems, manual processes, missing data, and poorly designed workflows.
AI creates an opportunity to fix that.
If AI can help employees pull together information from different systems, understand customer context faster, summarize history, recommend next steps, and reduce admin work, the customer experience improves almost immediately.
That is why I believe EX is one of the most important frontiers for AI-augmented CX.
Why the goal should never just be to add AI to CX workflows
My advice to CX leaders is to embrace AI — but don't abandon the fundamentals of good customer experience.
AI is a powerful new layer of technology. It can help teams move faster, analyze more data, personalize journeys, prototype sooner, and support customers in previously impractical ways. But the core process of solving problems in a user-centered way remains unchanged.
You still need to understand the customer. You still need to define the right problem. You still need to align business goals, technology, and human needs. You still need to test assumptions before scaling a solution.
My advice to CX leaders is to embrace AI — but don’t abandon the fundamentals of good customer experience…The goal should not be to add AI everywhere, but to make the journey feel easier, clearer, and more useful for the customer. That takes design, research, data, and human judgment, not just technology.
The danger is that teams use AI to build faster without first asking if they are building the right thing. That can create more noise, more automation, and more complexity for customers.
The goal should not be to add AI everywhere, but to make the journey feel easier, clearer, and more useful for the customer. That takes design, research, data, and human judgment, not just technology.
Follow Along
You can follow Jaco van den Heever's work on LinkedIn, X, and his personal website. Or check out Sand Dollar Design on its website and LinkedIn:
More expert interviews to come on The CX Lead!
