In the boardroom it often sounds like this: "We want to become more customer-focused, and AI helps us with that." Everyone nods. No one means the same thing. One person thinks of faster response times to service requests. Another thinks of personalized offers. Someone else thinks of a customer service that also answers at night. The phrase "more customer-focused" connects three ambitions, each of which presupposes a different organization.
Making that ambition concrete means laying it out in layers: a vision (why customer focus matters now), a vision state (what the customer relationship looks like then), a target (what changes this year) and a target state (which behavior is standard afterward). Only then does it become visible which capabilities are needed to get there, and which already exist.
Customer contact consists of tasks that behave differently once AI joins in. Part of it can be taken over: answering standard questions, summarizing a customer file before a conversation, flagging a customer at risk of leaving. Part of it happens with oversight: AI proposes an answer, an employee approves it or redirects it, with reason. And part of it remains human work: the conversation in which a customer is angry about something that doesn't fit a protocol, or the decision to make an exception to policy.
This division is not fixed. At one company, eighty percent of first-line questions are already handled automatically, at another nothing yet, even though both companies operate in the same sector and serve comparable customers. The difference rarely lies in the technology. It lies in whether the organization has broken down the customer query into tasks that can be assessed separately, whether it knows which data is needed for that and whether that data is accurate, and whether someone has the authority to reject an AI proposal without that becoming a problem.
Working more customer-focused with AI requires readiness across eight dimensions, not just in technology. A few that are often overlooked:
Process readiness. Has the customer process already been broken down into steps that can be assessed separately for "can AI do this," "can AI do this with oversight," or "does this remain human work"? Without that breakdown, the ambition remains an intention.
Data availability and data quality. An AI system that answers customer questions is only as good as the customer file it relies on. Scattered, outdated, or contradictory customer data produces answers that are more likely to harm the customer relationship than improve it.
Decision rights. Who is allowed to reject an AI proposal for a customer, and on what grounds? If that is not established, the habit arises that everyone accepts the proposal because deviating from it is hard to justify, or that no one accepts it because no one wants to bear the responsibility.
Skill to exercise oversight. An employee who assesses AI proposals needs a different skill than an employee who formulates the answer themselves: the skill of critical reading and well-founded rejection, not that of writing.
This readiness is tested through five confidence gates: is the ambition clear enough to test, does the organization agree internally on what success means, are the risks of an incorrect customer answer known, is there a way to roll back if something goes wrong, and is there someone monitoring progress. Without these gates, "becoming more customer-focused with AI" remains a wish.
The capability translation makes visible which roles are needed, regardless of how many people fill those roles. Someone who can break down customer processes into tasks. Someone who monitors the quality of customer data. Someone with the decision right to reject an AI proposal. And someone who measures whether customer satisfaction is actually changing, not just the response time.
Whether these roles belong to existing positions or require new ones is a question that differs per organization and carries its own legal requirements once it touches personnel decisions. That is up to the employer; what matters here is which work is shifting and which capability is needed for it, not who fulfills that capability.
It has succeeded if the leadership team means the same thing by "more customer-focused," if customer queries are demonstrably broken down into tasks with a clear takeover percentage, if there is an established procedure for when an employee rejects an AI proposal, and if the hours freed up in customer service are visibly deployed on the work that remains human: the difficult conversation, the exception, the customer who needs more than a protocol offers.
This ambition is connected to others. Growing without expanding customer service proportionally requires the same readiness, elaborated in which capabilities growing without extra staff requires from your organization. Customer focus also becomes visible in quality, which touches which capabilities improving quality requires when AI shares the work. And if marketing, sales, and service all say "more customer-focused" but pursue different targets, what to do when department goals work against each other helps surface that conflict.
The question of which customer work in your business can truly be taken over by AI is answered per task with the work scan from FTE TO AI.
The free readiness check consists of eight short questions, one per dimension, and provides a picture of where your organization is furthest along and where it lags behind. The full ambition test, with the four layers and the five confidence gates, is under construction.
Vertel wat u wilt bereiken, dan kijken we samen wat daarvoor moet staan.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.