Artificial Intelligence(AI)¶

Summary of the state of AI!

Some people argue that AI is merely advanced autocomplete, lacking true intelligence. But on the other hand I have been able to do things and collaborate with AI that I would not have been able to do without it.

In a few years time no one will care how many people logged into ChatGPT! People will care about how it has transformed their work. Those in charge of organizations understand the difference between more use and better use going to outperform their past selves and their competitors.

In order to make the best use of AI we need to better understand what types of tasks AI is suited and which are not good candidates for AI automation.

The great AI measurement delusion: Organizations are tracking usage metrics while missing the only thing that matters-impact.

The question isn't "How many people are using AI?" It's "How much increased has organisational Throughput or productivity increased and how much real time and money are we saving?".

How many critical workflows have you improved? What percentage of your regular responsibilities are AI-augmented?

Adoption techniques:¶

• Creating space for experimentation.
• Treat AI like a conversation partner.
• Start an "AI Innovation Channel" in Slack/Teams
  • Create a dedicated space where people share successful prompts and lessons learned.
• Audit one high-value process per department​
  • Where's the biggest ROl potential for a 45-minute AI fix?
• Track actual cost savings or new revenue​
  • Gather real data on how much time or budget each AI-augmented pilot saves

We are moving to a world were Agents will do work on our behalf¶

Context is essential for you AI Work¶

Loading your AI agent with information for the work is essential.

References to incorporate - WIP¶

See interview recommendations

• Problem Formulation: Understanding which problems are suitable for AI solutions is crucial. AI should be applied to problems that are descriptive rather than prescriptive.
  • Are there sub-problems that are suited to AI?
  • 1. Do you have enough data that you can use for training? Either locally or on the Internet.
  • 1. Is there a probabilistic nature to the data? ie when you plot variables of the data you can see that there is a relationship. You could draw a line that follows the dots in chart. For example the sales price of property and relationship to size of property. There need to be common dimensions, every house has a size.
• Conditional Probability: Over-reliance on AI can lead to compounding errors. Deterministic solutions should be employed where possible to ensure reliability.