There are benchmarks for which AI models write the best Python code or score highest on the LSAT. But there hasn't been one to help business users pick the best AI tool for real-world work. Until now. 🗞️ Read about it in the latest edition of our #AskMoreOfAI newsletter.
So in conclusion, in a space and time where new AI models outperform each other every week, we now need to be less worried about ensuring a future proof LLM architecture? Sounds counter intuitive to me. Some model from Open AI may be better at certain use cases today, and next week not anymore. Besides if you’re looking at developing and deploying LLM use cases enterprise wide, you will need access to models that may be from completely different providers. Hence you need to be able to access all model providers and be able to switch from one model to the other as the landscape evolves and more importantly as your business needs evolve in terms of use cases.
Namrata Pal
4w
We need to define clear metrics to evaluate which AI model is best for a task. 1. For example, if we use AI-generated content to onboard new hires into a company's culture, we could gather feedback from new hires to know if the AI-generated content was really effective. 2. On the other hand, if we use AI to extract information from unstructured data / text and load it into database, we could have another automation that validates the accuracy of data inserted AND we should also compare the performance of AI model vs using traditional automation. To sum up, the choice of AI model should depend on the use case and those who are responsible for the use case should define correct metrics to evaluate different solutions.
It is clear that, if AI will result in reducing CAPEX, or reduction in lead time in logistics, improvement of efficiency’s or with all that make your Competition more Revenue and profit. No company will be able to not implement these models if they don’t want to be out of business. Companies who are in AI since already a long time claim that how AI will change the world the next 5 years , we haven’t seen or experienced anything like that the last 50 years. And if one looks deeper into into what these companies can do already today it looks like their outlook is right! Interesting times and impressive if one diggs deeper than just the surface of it.
Acredito que, assim como nos tornamos especialistas em executar tarefas específicas, as IAs seguirão esse mesmo caminho. Haverá inúmeras IAs especializadas em determinados assuntos que evoluirão continuamente, aprendendo mais sobre esses temas. Isso permitirá que as equipes se concentrem em aspectos mais estratégicos de suas funções, ajudando a construir relacionamentos mais duradouros e lucrativos.
Interested
Farooq Omer
3w
Useful tips
Olga Gibbs
4w
Good point!
I agree! ❤️🫠
For those of you interested by getting guidance on which LLMs to use for your CRM, check our latest LLM benchmark for CRM hereby....
Secretary @ LiTrichy | IIM Trichy | MBA,HR
4wEvaluating AI usability in enterprise settings, such as CRM systems, is more practical than current LSAT techniques. It is amazing that Salesforce is making it industry-specific as well, the curation of LLM benchmarks with a human touch at their core is a significant advantage, enabling more effective assessment and refinement of AI capabilities.