What should we call efforts to sway large language models and generative AI?
Here are some options…
LLMO – Large Language Model Optimization
Pros
- There’s a reasonable degree of familiarity with the term LLM.
- LLMO could be fairly unambiguous.
Cons
- We may move away from the term ‘LLM’ as the most widely-used AI models become more ‘multimodal’ (i.e. they deal with images, audio and more as well as just text.)
- LLMO is quite a mouthful to say (unless it starts being pronounced ‘Elmo..?’)
GAIO – Generative AI Optimization
Pros
- ‘Generative AI’ has become a fairly popular and widely-understood term.
- It should remain just as relevant as models become more multimodal.
- It could perhaps be pronounced ‘guy-oh’ which would be more convenient to say than L-L-M-O.
Cons
- People might not know how to spell it.
- It might be hard for people to guess what it means.
SGEO – Search Generative Experience Optimization
Pros
- Google use of the term ‘Search Generative Experience‘ (SGE) has popularized that terms within the SEO industry.
- Its similarity to SEO could help people understand what it is.
Cons
- There’s little awareness of the SGE term outside of the SEO community.
- It’s a mouthful to say.
- ‘SGE’ has slight Google-specific connotations.
- It suggests it’s about search whereas it may be important to influence AI models in non-search contexts, too.
GEO – Generative Engine Optimization
Pros
- It’s convenient to pronounce.
- It’s easy to spell.
- When pronounced ‘G-E-O’, the similarity to ‘SEO’ would helpfully suggest a relatedness between the two.
Cons
- ‘Geo’ has what would be confusing connotations of being related to the Earth, ground or geography.
SEO – Search Engine Optimization
Pros
- It’s a very widely-understood term. Why invent something new?
- Traditional SEO is likely to be a key element of influencing AI models.
Cons
- It suggests it’s about search whereas it may be important to influence AI models in non-search contexts, too.
AEO – Answer Engine Optimization
Pros
- The similarity to SEO could help it be easily understood.
Cons
- The ‘Answer Engine’ term seems a neat one but doesn’t seem to have caught on.
Conclusions
There are a number of terms that are candidates to describe this new practice, all with their pros and cons. For me, at the moment (February 2024) there’s no clear ‘best’ term. I went with LLMO at the start of 2024. But I’m now thinking GAIO has a slight edge.
Will one become dominant? Or will a new term emerge? I think it’s hard to predict.
Watch this space!