Latest / Anaiya Algorithm / Veridian: Episode 1_The Agentic Shift — From Chatting to Executing
Transcript
- 0:16SPEAKER_00: Welcome to the Anaya algorithm.
- 0:19SPEAKER_00: I'm Magdalene Amekashti, CEO of Anaya Group.
- 0:25SPEAKER_00: Over the next six special episodes, I want to have a very honest conversation about something that is happening right now in boardrooms and several rooms across the UK and Europe.
- 0:42SPEAKER_00: The shift from AI as a tool we use to AI as an agent we deploy and what that actually means for governance, accountability, and trust.
- 0:58SPEAKER_00: So I'll give a couple of examples.
- 1:01SPEAKER_00: Example one a logistics company deploys an AI agent to manage supplier contracts.
- 1:11SPEAKER_00: The agent identifies a cheaper supplier, terminates the existing contract, and initiates a new one, all within its defined parameters.
- 1:25SPEAKER_00: It wasn't just a cheaper contract, it was a decade long relationship terminated by an algorithm that didn't know how to value loyalty.
- 1:38SPEAKER_00: No human signed off.
- 1:41SPEAKER_00: Was that the right decision?
- 1:45SPEAKER_00: Nobody knows, because nobody was watching.
- 1:50SPEAKER_00: Example two.
- 1:53SPEAKER_00: A financial services firm uses an AI agent for customer onboarding.
- 2:00SPEAKER_00: The agent starts declining applications from setting postcodes at a higher rate.
- 2:09SPEAKER_00: It's not breaking any explicit rule, but it is creating a pattern that looks a lot like indirect discrimination under the Equality Act.
- 2:22SPEAKER_00: When did it start?
- 2:24SPEAKER_00: Why?
- 2:26SPEAKER_00: Nobody knows.
- 2:29SPEAKER_00: Now we go into the third example.
- 2:33SPEAKER_00: Example three.
- 2:34SPEAKER_00: A retailer deploys a pricing agent.
- 2:39SPEAKER_00: The agent adjusts prices based on demand signals.
- 2:45SPEAKER_00: During a national emergency, it raises prices on essential goods by three hundred percent.
- 2:53SPEAKER_00: The algorithm was doing its job, but the company faces regulatory action and a public relations crisis.
- 3:04SPEAKER_00: Here is the question I want every leader listening to this to sit with.
- 3:11SPEAKER_00: If one of your AI agents made a significant error tomorrow, say a financial error, a discrimination claim, a regulatory breach, could you reconstruct exactly what it did?
- 3:28SPEAKER_00: Why it did it?
- 3:29SPEAKER_00: And when it started going wrong?
- 3:33SPEAKER_00: If the answer is no, you have a governance problem.
- 3:40SPEAKER_00: At Anaya Group, we have been working with organizations navigating exactly this challenge.
- 3:48SPEAKER_00: And what we kept seeing was the same gap.
- 3:52SPEAKER_00: Incredible investment in AI capability, but almost no investment in AI governance.
- 4:01SPEAKER_00: So you find that the agents were live, however, the oversight was not.
- 4:10SPEAKER_00: So this is exactly why we built Viridian, not as a compliance checkbox, but as a command center.
- 4:21SPEAKER_00: Every AI agent your organization deploys should be registered, classified, monitored, and governed.
- 4:31SPEAKER_00: You should know its risk tier, you should know when its drifts, you should be able to generate audit evidence when things go wrong or prove they did not.
- 4:46SPEAKER_00: That is what active EI governance looks like.
- 4:51SPEAKER_00: In episode three, I will show you exactly how it works.
- 4:56SPEAKER_00: Next week I want to talk about what happens when the agent does go wrong.
- 5:04SPEAKER_00: Because it will.org.
- 5:47SPEAKER_00: And if you want to explore Meridian, it's live today on Microsoft Azure Marketplace.
- 5:54SPEAKER_00: Link in the show notes.
- 5:57SPEAKER_00: If you are a leader committed to building your own intentional equitable technology strategy, you can learn more about our frameworks at Anana Group by visiting ananya.org.
- 6:10SPEAKER_00: Until next time, keep leading intentionally.
- 6:14SPEAKER_00: Thank you.