Strategic Anchoring in the AI Sea

Methodology Over Tools: The Real Innovation in AI

If you have been working with AI tools for any length of time, you must have asked “Which LLM model is best?” The answer keeps on changing. Focussing only on the tools distracts from where the real revolution is happening; it’s in the methods we discover for using them. Methodologies currently drive innovation more than tools alone. A new way of working emerges, and a whole ecosystem of tools follows. The “how” of working with AI is what truly shapes the future of human-AI interaction.

The Evolution of Interaction Methods

Consider the journey of interaction. Each step wasn’t just a new feature; it was a new methodology for getting value from this technology.

1. The Chat Box: Initially, interaction was simple: a conversational Q&A partner.
2. The Prompt Engineer: Then came the realisation that input quality dictates output quality, birthing the discipline of prompt engineering.
3. The Grounded Model (RAG): The methodology shifted to grounding models in external data to improve accuracy.
4. The Tool-User (Function Calling): With function calling, the methodology evolved to directing a tool-user that could interact with the world.
5. The Autonomous Agent: More recently, the methodology has become one of delegation to autonomous agents.
6. … By the time you read this, something new is bound to be happening.

Each time a new methodology appeared, the community raced to build tools that enabled it. The methodology was the catalyst.

Strategic Anchoring: The Value of Temporary Stability

In the presence of this much change, the more effective approach is strategic commitment, which balances stability and flexibility. You drop an anchor long enough to gain the benefit of expertise and build something robust. But you keep your hand on the chain, ready to pull it up when the tide truly turns. This requires the wisdom to know when to commit and the courage to know when to move on.

While strategic anchoring provides the stability to build expertise, the accelerating pace of methodological change demands that we complement this commitment with the agility to pivot when the tide turns, making rapid methodology learning not just valuable, but essential.

Building When Your Tools Keep On Changing

Here is an example, that is sure to be outdated soon: Any project that started in 2024 may not have considered using agentic AI tools or MCPs. Any project that starts in 2025 has to make sensible choices around these. It changes who has to be involved in a project, when and where. Spec Coding changes the artefacts that go into and out of a sprint. Your AGENTS.md file and memory bank now should go into your repository too.

We have moved from picking the team IDE, to picking the team coding agent, and this changes not only what we do, but how we do it. It is a different tool for a different methodology.

And it will keep on changing. In the past this type of shift happened over years. It is now a matter of months and probably shorter than the life of the project.

The New Premium: Rapid Methodology Learning

This environment creates a new premium on a specific human trait. Adaptability is the key to thriving in fast-changing AI landscapes. The advantage no longer belongs to the person who has mastered a single methodology and its tools, but to the person who can quickly grasp the essence and benefits of new perspectives and methodologies. The rapid learning of new methodologies is the primary skill of the AI era.

Building Adaptive Organizations

This premium on adaptability reshapes workforce priorities and leadership. The pace at which methodologies change now often outpaces the tempo at which projects are delivered. A project’s core assumptions can become obsolete before it even launches.

Therefore, leadership must prioritize learning cultures over rigid planning. The challenge is no longer just to manage a project, but to build resilient teams and flexible processes capable of navigating a constantly changing methodological sea.