https://andrewsheves.com/categories/risktech/Recent content in RiskTech on Andrew ShevesHugoen-usTue, 22 Aug 2023 00:00:00 +0000https://andrewsheves.com/2023/08/22/why-and-how-to-build-custom-chatbots-/Tue, 22 Aug 2023 00:00:00 +0000https://andrewsheves.com/2023/08/22/why-and-how-to-build-custom-chatbots-/<p>I’ve been building custom knowledge agents (AKA chatbots) for a few weeks, and my friend Greg suggested I share some background on how you build these tools and why. The ‘how’ is covered in a short demo video towards the end of the piece, but before you start building nodes, it’s worth thinking about what these new chatbots are and what they can do for you.</p> <p>First, we should ask what chatbots are and why they weren’t so helpful before.</p>https://andrewsheves.com/2023/08/10/risk-and-security-managers-are-positive-about-ai/Thu, 10 Aug 2023 00:00:00 +0000https://andrewsheves.com/2023/08/10/risk-and-security-managers-are-positive-about-ai/<h2 id="risk-and-security-managers-positive-on-ai">Risk and Security Managers Positive on AI</h2> <p>Last week, I had a great session with ASIS at the AI for Risk and Security Managers webinar. As part of the session, I outlined the three questions organizations should ask themselves concerning AI and how it affects their business.</p> <ul> <li> <p>How could AI give us an <strong>advantage</strong>?</p> </li> <li> <p>How could AI <strong>challenge</strong> our business?</p> </li> <li> <p>How could AI <strong>threaten or kill</strong> our business?</p> </li> </ul> <p>(<a href="https://prepared.andrewsheves.com/p/businesses-can-plan-ai?utm_source=prepared.andrewsheves.com&amp;utm_medium=newsletter&amp;utm_campaign=risk-and-security-managers-are-positive-about-ai">Read the full article here to give you the context of the questions and some additional considerations.</a>)</p>https://andrewsheves.com/2023/03/15/crisis-management-exercise-generation-using-ai/Wed, 15 Mar 2023 00:00:00 +0000https://andrewsheves.com/2023/03/15/crisis-management-exercise-generation-using-ai/<p>To be successful, the process would have to:</p> <ul> <li>Be faster than if I did it ‘by hand’</li> <li>Produce realistic inputs</li> <li>Be easily repeatable</li> <li>Allow for customization to meet client needs</li> </ul> <p>The outputs had to include:</p> <ul> <li>Ideas for scenarios</li> <li>A short video news report</li> <li>Social media clips</li> <li>Text for a news story</li> </ul> <p>This left me with five challenges to solve:</p> <ol> <li>Get ChatGPT into my workspace.</li> <li>Generate scenario ideas.</li> <li>Generate supporting text material.</li> <li>Generate a video news clip input.</li> <li>Generate social media inputs.</li> </ol> <p>Here’s how the AI Crisis Management Exercise Build Hackathon* went down. (*Attendee one)</p>