Fix the foundations first

 AI has very quickly become the answer to almost every single CRM challenge. Poor engagement? Add AI. Low conversion? Layer in some AI decisioning. Struggling to deliver the right message at the right time? There’s an AI platform for that. But for many marketers, the results haven’t quite lived up to the promise. Not because AI doesn’t work, but because it’s being layered onto CRM systems that weren’t in great shape to begin with. 

AI amplifies bad CRM. If customer data is fragmented, outdated or inconsistent, it processes those issues faster and at scale. Instead of improving customer experience, it can end up accelerating the delivery of irrelevant messages and disconnected journeys. AI is often positioned as a short cut, but it can only work effectively if the basics are in place.  

CRM isn’t simple. Organisations are often dealing with disconnected systems, duplicated records and teams that define the customer in different ways. Marketing, sales and service often operate in parallel, rather than in sync. AI enters this environment with a lot of promise. It can surface patterns humans miss, automate decisions and help brands respond in real time. But it doesn’t clean the input. 

If the underlying data is messy, AI can’t correct it. Unfortunately, it learns and then it acts on it. This is where the gap starts to appear. Customers begin to receive messages that feel a little ‘off’. Timing is wrong. Recommendations don’t land. Channels are inconsistent, rather than connected. 

Research published in Heliyon (ScienceDirect, 2024) highlights this risk, finding that unplanned AI adoption can negatively impact organisational performance and can increase employee resistance. The issue isn’t the technology itself, but the lack of strategy before implementation. 

The companies seeing real value from AI in CRM share a similar approach in that they start with data. Which means getting the fundamentals right: clean, structured customer data. Consistent definitions across marketing, sales and service. Clear governance over how the data is captured and used. And enough trust in the underlying system that teams rely on it. 

AI promises speed and efficiency, while governance demands discipline and patience. It’s no surprise teams are often drawn to the speed first, but the businesses getting this right understand that governance isn’t slowing AI down – it’s what makes it usable in the first place. This aligns with The Drum’s reporting on AI in CRM, which highlights that the strongest performers treat AI as an enhancement to solid CRM foundations, not a replacement for them. 

There’s a growing tendency to position AI as something that will take over decision-making in CRM. But in practice, the most effective use cases are far more grounded. AI is useful for prioritisation. It can reduce manual workload and speed up execution, but it doesn’t replace judgement. CRM is still about relationships, which rely on context and nuance, which AI can support but not fully comprehend in isolation.  

The strongest teams are the ones using AI to enhance decision-making, not remove it. When that balance is right, AI is incredibly powerful. But when it isn’t, you risk automation without relevance. Where messages are technically accurate but emotionally off.  

One of the most overlooked challenges in AI adoption is human, not technical. If teams don’t trust the system or understand its decisions, they’ll work around it, reducing both data quality and performance.  

This is why it’s crucial to involve users early, during design, not just rollout. CRM, sales and service teams need to shape how AI fits their workflows, because adoption is about trust. If people don’t trust outputs, they won’t use them, and system integrity erodes. 

AI presents huge opportunity when it comes to CRM, but it can only be unlocked once the fundamentals are in place. Customers expect brands to feel relevant and connected, and AI can help deliver that. but only if it’s built on reliable data, aligned teams and trusted systems. 

The organisations seeing the greatest success aren’t using AI as a shortcut or replacement for good practice. They fix the foundations first: improving data quality, strengthening governance and creating a clearer view of the customer. AI then becomes an accelerator.