Customer experience (CX) as a competitive differentiator, especially for AI focused organizations:
As we are seeing, models are being commoditized at lightning speed. Underneath, all LLMs,
- whether it be ChatGPT or Google Bard - are just transformers. The primary differentiating factors being:
- the data it’s trained on
- the layers on top such as the alignment and fine tuning layers based on reinforcement learning from human feedback (RLHF)
- and finally the consumption layer (front-end)
Since most consumer facing models have similar architectures and are trained on more or less the same datasets, they are converging in terms of performance - so performance is barely a key differentiator and any differences will eventually erode.
If performance is not a sustainable competitive differentiator, then what is?
Well, we are just left with the alignment/fine tuning and the consumption layers - when you don’t have many vectors for differentiation, the few ones you do have become paramount.
Alignment/fine turning attributes to CX and perceived sense of performance. The front-end is obviously the biggest contributor to CX.
In order for AI organizations to succeed, when everyone is using the same models - more or less - providing the best CX will act as a key sustainable competitive differentiator.
An analogy: When every lemonade stand uses the same premix and water, you obviously go to the stand where they treat you the best. It’s not rocket science.
What does this mean for organizations, particularly AI focused orgs:
1. Need to make CX a top priority as opposed to an afterthought (for orgs that simply use a foundational model such as Jesper AI - CX by default should be THE top priority).
2. Build a moat with CX at the heart.
3. Discover and build lock-in mechanisms geared around CX.
