When do I use M365 Copilot, Copilot Studio, or Advanced Data Science solutions?

It seems like everybody is interested in adopting M365 Copilot and surrounding AI capabilities. Part of this is the general interest in AI and another part is they perceive it as being closer to how they’ve treated other IT-like projects. The adoption of Copilot is not just another technology adoption. It’s part of a larger picture of leveraging AI as part of the mission of your business. The conversation on M365 Copilot inevitably moves into a broader conversation of use cases that would fit into semi-custom and fully custom solutions because of the nature of the work.

The first thing to realize when you start down the journey is that tools like M365 Copilot are part of a larger picture which includes Commodity solutions and Mission-Driven solutions in alignment with your business strategy. Your strategy will include solutions you adopt (like M365 Copilot) from diverse vendors that are embedding AI within their products. The other solutions you are building yourself, but you are building them not as a side-project, but because it directly impacts the strategy of your business.

The second thing to understand is that all solutions are part of a broader ecosystem. The modalities of AI all should be part of interconnected brains that can exchange information with each other to accomplish greater ends. For example, many companies have build purpose-built GPTs for their organization (like ConcurrencyGPT) that can surface specific information to answer questinos with precision and accuracy. This can be integrated into M365 Copilot or Copilot Studio as a form of skills or plug-ins that exchange information.

The third thing to think about is that there are right use cases for the right tools. Generally speaking, the more you need accuracy and precision that is specific to your business, the more you’ll be moving into an advanced AI solution. This sounds intuitive, but often when companies are envisioning Copilot solutions it’s not something they consider. It’s also critical as you look at the middle lane, which is semi-custom solutions. This lane is immensely powerful and scaling it can create huge dividends. It’s also a lane you need to govern, because often solutions built there need to move into management more-akin to advanced AI, due to their business criticality.

The most important things to remember when selecting a solution:

1. These are the “how” not the “why” and “what”. If you haven’t started with “why” and “what”… do not pass “go”. Start over, go back to basics and confirm what you are trying to achieve with AI in your business.
2. The adoption of M365 Copilot specifically will achieve increased success if you know specifically what different personas will use it for. Identify what you want pilot users to leverage it to achieve, such as HR using it to draft job description and specifically testing it to perform that function. What you should hear in this is…. “what is the job to be done”. Also remember that the skill of delegation is not always intuitive, so this is not just tool adoption, but skill adoption as well.
3. The use of Copilot Studio, similarly to Power Apps, has a dramatic capability to scale impact of near term semi-custom solutions in a business. As you can see from the examples above, each of these has relatively low stakes, but can provide scaled impact. It also has an opportunity to broaden the reach of AI skills across various business groups. Building here is a bit like building with pre-created lego bricks, enabling quick creation of mid-criticality solutions for the business. It also can facilitate interaction with both M365 Copilot and Advanced AI. The thing to watch for is that sometimes these solutions move into fully custom AI, or will interact with it, so it might move up in its need for security, control, governance, and stability through a mature MLOps pattern.
4. The use of custom AI solutions will be directly tied to increased need for Accuracy or Precision in the context of your business goals, as well as increased control over the operational stability of the solution. Many AI solutions built in this lane become the heart of the business and quality, reliability, security, management of output are critical elements.

We are in an exciting time where gaining ground in AI skills across the business, as well as the intersection with high quality AI/ML solutions are part of the greater strategy being created. I hope this gives you some initial perspective of how these all fit together and encourages you to think about your strategy as a whole, not just in one lane.