Last week, I watched a client launch into an excited monologue about their new AI implementation strategy. Twenty minutes later, I finally managed to ask: “What specific problems are your AI implementation plans trying to solve?” The room fell silent. Nobody knew.

With big tech stocks taking a beating and businesses scaling back their AI implementation budgets, the Financial Times recently reported that AI adoption is facing a significant slowdown.

This shouldn’t surprise anyone who’s developed technology implementation strategies as long as I have. Every AI implementation strategy follows the same pattern: wild enthusiasm, followed by a reality check when deployment proves harder than expected.

This isn’t an AI implementation technology problem. It’s an AI implementation strategy problem.

The missing foundation in AI implementation strategy

The fundamental flaw in most AI implementation strategies is starting with the technology rather than the business problems it’s meant to solve.

When organisations approach AI implementation without clear user needs, they inevitably waste resources creating capabilities nobody asked for while ignoring the simple improvements users actually want.

I’ve seen this AI implementation strategy mistake play out repeatedly throughout my career.

When I was editor-in-chief for a complex insurance portal, our head of IT insisted on using SharePoint as the core system without considering user requirements.

The firm negotiated the implementation cost down by around 75 percent, which seemed like a win – until we discovered that everything we subsequently needed was “out of scope” and required expensive separate development work.

The banks eventually pulled the plug because the project was perceived as an unstable financial risk.

This technology-first AI implementation strategy manifests in several common patterns:

1. The executive mandate: “We need an AI implementation strategy” becomes a directive without any clear articulation of what problems the AI implementation should solve.
2. The hype-driven AI investment: Organisations commit significant AI implementation resources based on FOMO rather than identified opportunities for improvement.
3. The solution transplant: Companies attempt to replicate another organisation’s AI implementation strategy without understanding the specific user needs in their own context.

The recent market correction reflects a growing recognition that this AI implementation approach doesn’t work.

As the Financial Times reported, many companies “have not found a use case for AI yet”, despite the considerable hype surrounding AI implementation.

User needs as the foundation for successful AI implementation strategy

As someone who’s launched multiple B2B publications and spent years developing technical content strategies, I’ve learned that successful AI implementation strategy always begins with understanding user needs.

During my time at Ten Alps Publishing, launching B2B news platforms for Sir Bob Geldof’s media firm, our success came not from having the flashiest technology but from identifying specific information gaps that weren’t being filled by established outlets – like providing 7:30am email newsletters with quick summaries of the latest corporate finance developments.

A user-needs approach to AI implementation follows the same principle: understanding the actual problems your users face, whether they’re customers, employees or other stakeholders. This means conducting rigorous research before making AI implementation decisions.

Effective user research for AI implementation strategy includes:

1. Identifying friction points: Where do users struggle with current processes or systems that AI implementation could address?
2. Understanding information needs: What questions do users need answered that current systems fail to address efficiently but AI implementation might solve?
3. Recognising cognitive overload: Which tasks require significant mental effort that could be reduced through AI implementation?
4. Mapping repetitive tasks: What work do users perform repeatedly that adds little value but consumes significant time and could benefit from AI implementation?

Only after understanding these needs should you consider whether and how your AI implementation strategy might address them.

AI implementation strategy and content strategy connection

Content strategists have a crucial role to play in bridging the gap between user needs and AI implementation capabilities. Our understanding of both user requirements and content systems positions us perfectly to identify where AI implementation can genuinely add value.

When I led the restructuring and migration of 5,000 pages of content for Cancer Research UK, we had to thoroughly understand user pathways before implementing any technology solutions. The same principle applies to developing an effective AI implementation strategy.

For example, a content strategist developing an AI implementation strategy might identify that:

• Users struggle to find specific information in complex documentation, suggesting an opportunity for AI-enhanced search and summarisation as part of the AI implementation plan
• Customer service agents repeatedly craft similar responses to common queries, indicating a use case for AI-assisted response templates in the AI implementation roadmap
• Content creators spend excessive time on routine editing tasks, pointing to potential AI tools for grammar and style checking as a high-value AI implementation priority

In each case, the starting point for AI implementation is the user need, not the technology.

Starting small: the pilot project approach to AI implementation strategy

Rather than implementing AI broadly across an organisation, successful AI implementation strategy typically begins with small, focused pilot projects that:

1. Address a clearly defined user need with your AI implementation
2. Have measurable AI implementation success criteria
3. Involve minimal disruption to existing workflows during AI implementation
4. Require relatively modest AI implementation investment
5. Can deliver AI implementation results within 3-6 months

This approach allows organisations to validate both the technology and the user need before committing to larger-scale AI implementation.

I’ve seen implementation strategy work brilliantly and fail spectacularly. At DeskDemon.com, where I led a team of 12 for a £1.5 million portal development, we succeeded by starting with small, focused improvements that delivered immediate value rather than attempting a complete transformation at once. This is a lesson directly applicable to AI implementation.

The critical role of human oversight in AI implementation strategy

Every successful AI implementation strategy maintains appropriate human oversight. This isn’t just about preventing errors, but about ensuring the AI implementation genuinely serves user needs rather than creating new problems.

During my time at Towergate Insurance, where I managed a team of 8 writers, content managers and designers in a regulated environment, I learned that automation without proper oversight is a recipe for disaster. Financial services content requires meticulous human review to ensure compliance and accuracy, something that remains true with today’s AI implementation projects.

Effective human oversight in AI implementation strategy includes:

1. Clear guidelines for when AI should and shouldn’t be used within your implementation
2. Regular auditing of AI outputs throughout the implementation lifecycle
3. Accessible pathways for users to report issues with the AI implementation
4. Continuous evaluation of whether the AI implementation is actually solving the intended problem

This oversight should be designed into the AI implementation strategy from the start, not added as an afterthought.

Signs your AI implementation strategy is missing the user needs mark

How can you tell if your AI implementation strategy has lost sight of user needs? Watch for these warning signs:

1. Users create workarounds to avoid the AI system
2. The same questions keep coming up in AI implementation training sessions
3. Key metrics don’t improve despite the new AI implementation
4. Staff express frustration about having to “feed” the AI system
5. The focus shifts from outcomes to the AI implementation technology itself

I witnessed all of these red flags during my work on multiple digital transformation projects, including for the Rural Payments Agency. Users created elaborate workarounds because the system was fundamentally unsuited to their needs, yet management kept focusing on the technology rather than the underlying problems. The same pattern is  repeated in many failed AI implementation projects.

If you observe these patterns in your AI implementation strategy, it’s time to step back and re-evaluate whether your AI implementation is actually addressing genuine user needs.

Practical next steps for user-centered AI implementation strategy

If you’re considering implementing AI in your organisation, start with these practical AI implementation strategy steps:

1. Document existing pain points: Before even thinking about AI implementation, thoroughly document where your current systems or processes fail to meet user needs.
2. Quantify the impact: Estimate the cost (in time, money, or missed opportunities) of these pain points to prioritise which to address first in your AI implementation strategy.
3. Explore multiple solutions: Consider whether AI implementation is actually the most appropriate solution for each pain point, or whether simpler alternatives might work better.
4. Start small and focused: Choose a single, well-defined problem for your initial AI implementation project.
5. Measure what matters: Define AI implementation success metrics based on user outcomes, not technical capabilities.

The current slowdown in AI implementation adoption isn’t a sign that the technology lacks potential. Rather, it’s a necessary correction as organisations recognise that successful AI implementation strategy requires a fundamental shift in approach that puts user needs first and technology second.

Those who make this shift in their AI implementation strategy will be positioned to realise genuine value from AI, while those who continue with technology-first approaches will likely join the growing ranks of expensive AI implementation failures, just as I witnessed when the banks pulled the plug on our insurance portal.