
Sign up to save your podcasts
Or


As organizations have worked to leverage the power of AI in interacting with large language models, they've invested in prompt engineering to generate better results. But agents shift the need manage the full context of not only the prompt, but also the data that's being presented. Analysts Jean Atelsek and Alex Johnston return to the podcast to look at the new discipline of context engineering and how it's being put to work in AI environments with host Eric Hanselman. The process of context engineering looks at ensuring that the right data context is in place for agents to act on. It requires a shift from thinking that more data is necessarily better and understanding to getting the right data is the best insurance against agents picking up bad habits. We've come full circle in approaches to data and organizations need to raise the level of abstraction at which they address data need for agentic applications.
We've been working through waves of capability in the march to agentic operations. Organizations have access to the same models, but how they're used is where differentiation is possible. Agentic approaches demand greater sophistication and understanding around the context with which data is presented to applications. There has to be more careful curation, to get reasonable results.
More S&P Global Content:
For S&P Global Subscribers:
Credits:
By S&P Global Market Intelligence4.9
2828 ratings
As organizations have worked to leverage the power of AI in interacting with large language models, they've invested in prompt engineering to generate better results. But agents shift the need manage the full context of not only the prompt, but also the data that's being presented. Analysts Jean Atelsek and Alex Johnston return to the podcast to look at the new discipline of context engineering and how it's being put to work in AI environments with host Eric Hanselman. The process of context engineering looks at ensuring that the right data context is in place for agents to act on. It requires a shift from thinking that more data is necessarily better and understanding to getting the right data is the best insurance against agents picking up bad habits. We've come full circle in approaches to data and organizations need to raise the level of abstraction at which they address data need for agentic applications.
We've been working through waves of capability in the march to agentic operations. Organizations have access to the same models, but how they're used is where differentiation is possible. Agentic approaches demand greater sophistication and understanding around the context with which data is presented to applications. There has to be more careful curation, to get reasonable results.
More S&P Global Content:
For S&P Global Subscribers:
Credits:

1,955 Listeners

2,707 Listeners

1,651 Listeners

1,095 Listeners

170 Listeners

6 Listeners

1,448 Listeners

40 Listeners

9 Listeners

6 Listeners

638 Listeners

228 Listeners

27 Listeners

28 Listeners

9 Listeners

4 Listeners

59 Listeners

29 Listeners

11 Listeners

10,019 Listeners

4 Listeners

5,528 Listeners

204 Listeners

1 Listeners

63 Listeners

186 Listeners

149 Listeners

6 Listeners

3 Listeners

0 Listeners

6 Listeners

5 Listeners

5 Listeners

59 Listeners