Why the orchestration engine business model is ripe for AI innovation


Orchestration is thought of as an evolution of workflows. Workflows are historically designed to mimic single-threaded, synchronous, processes. Orchestration introduces different self-contained workers that don’t block the orchestration engine – the brain. A simple analogy could be an ant colony. Individual workers are coordinated through an orchestration brain – the queen. They fulfill the task and role they are delegated to do independently and don’t interfere with other workers. The queen knows what to do next depending on the outputs of the workers. Remember the delegation part for later.

Phase 1: Early orchestration

A vendor recognizes a customerโ€™s business problem is best solved through orchestration. At this stage, a conscious decision needs to be made about what scenarios are in or out of scope for the orchestration. Slicing happens. How big a piece of cake do you cut off to eat, that you are able to eat, to deliver on? It can be mitigated with extensions and integrations to an extent, but it’s a conscious decision. The main demarcation of the slices will probably be determined by what useful inputs are needed in the orchestration engine.

Disruption happens when the vendor is first to market and no other vendors compete in terms of orchestration to address this slice of the customer business problems. A discussion can be opened here around horizontal and vertical coverage of solutions, as in focusing on a small slice and being the best vs. addressing a bigger slice of the customer’s business problems and reliance and focus on integrations.

The challenge in this phase is to communicate why the innovation is disruptive. What value does orchestration bring to the customerโ€™s business problem? Does the customer attribute value to this? Will the market reward the vendor for this approach? 

No alt text provided for this image

Phase 2: First competitor

A first competitor appears. Communicating the value of the disruption in the given slice of solving a customer’s business problems takes a backseat.

Vendors that decide to compete in solving the same customer problem in the orchestration game will have the challenge to communicate why their approach to orchestration is better.

In the past, you could say slicing happened based on the source of the truth and separation of concerns for entities important to the customers’ business, for example, content got clearly separated out into a headless CMS. Some focused on the UX advantage, and some on the technical capabilities and integrations.

In short, different vendors focused on combining slices, whilst some focused on deepening their capability in the selected slice.

No alt text provided for this image

Phase 3: Cooperation and alliances

There are many slices of the cake. Vendors will agree on how to slice it and which alliances are beneficial to not eating each otherโ€™s pieces.

Phase 4: A new disruption

A new cycle begins and orchestration alone might be deemed irrelevant by the next disruptive innovation. Orchestration still requires a lot of domain knowledge. Orchestration is important. However, what about when it can be automated? Orchestration will always be the engine, but the current brain is still a human one.

We are witnessing a big leap in AI processing, achieved by outsourcing, and delegating smaller tasks to dedicated workers. The main AI model is responsible for evaluating the output of individual workers and improving the next iteration of outputs by adjusting the input for the next iteration.

Winners in the orchestration game will be vendors that are able to create a measurable feedback loop; training an AI model. This AI model will then evaluate iterations of the different versions of a customer’s orchestration possibilities. Is the next iteration better or worse? The AI model re-adjusts inputs, continuously improving future iterations. 

A great example of this can be the online marketing space. Here AI models can be rewarded by tracking click-throughs and purchases; punished based on the bounce rate. This feedback loop then allows the AI to re-adjust the messaging and ad placement in marketing campaigns. 

Orchestration gives nearly endless possibilities on how to achieve something. AI will help us to figure out how to do it as effectively as possible. 

Let the AI orchestration games commence. A new, bigger cake appears, ready to be sliced.

No alt text provided for this image

Related post