This episode of the Time for AI Podcast features Jelena Miodragović, a freelance software developer and one of the organizers of the Init Conference in Banja Luka, Bosnia and Herzegovina. Jelena's conference presentation, "Power Up with Microsoft Power Platform to Build Smarter Business Solutions," focused on introducing the core concepts and demonstrating a real-world application of the platform, that she presented at the NTK conference in Portorož

Jelena Miodragović joins the podcast to discuss her work as a developer and conference organizer, alongside a deep dive into the capabilities of the Microsoft Power Platform.

• The Microsoft Power Platform: Jelena explained that the Power Platform is a powerful suite of low-code/no-code tools that allows even non-developers to quickly turn ideas into functional business solutions.

• Core Components Explained: She clarified the primary components of the platform and how they work together:

  • Power Apps: Used to build applications, such as the admin app in her demo for inputting event data.

  • Power Pages: Used to display data (like a website) where employees can view events and register.

  • Power Automate: A tool for automating various processes within the platform, such as sending notifications and registration confirmations.

  • Dataverse: The secure, central database for storing data within the platform.

• Copilot Studio and AI: Jelena expressed particular excitement about Copilot Studio, which allows users to easily create and embed interactive chatbots (including voice chatbots) into websites or Microsoft Teams. This integrates directly with the rest of the Power Platform services.

• Security and Ease of Use: A key benefit highlighted is the platform's security model, which allows for detailed, individual permission settings for tables and other assets. She also emphasized its ease of use, suggesting that many employees could build basic solutions with minimal training.

Filip Macukić, the founder and CEO of Mediartem AI, joined the podcast to discuss his presentation from the AI Weekend conference. His presentation was titled "The Unfair Advantage: Multiply Your Impact, Not Your Hours," which focuses on the distinction between human and machine productivity and the proper use of AI tools.

Filip Macukić is the founder and CEO of Mediartem AI, a company that provides customized training and services for organizations on how to use generative AI tools in daily business. Their training is tailored based on:

1. The team's current knowledge level of AI tools.

2. The preferred tool (e.g., ChatGPT, Copilot, Gemini).

3. The business need (e.g., sales, board, full company).

The company also offers a more comprehensive, long-term service called AI Academy, which aims for a complete AI transformation in an organization. This service is divided into four phases: awareness, upskilling, process transformation, and strategy/policy finalization.

Filip expressed that the AI Weekend conference in Rovinj, Croatia, was his first time attending. He appreciated the holistic approach of the event, which covered a wide range of topics from programming and technical aspects to ethics and everyday business.

The core of Filip's presentation was the difference between machine efficiency and human effectiveness.

• Machine Productivity is about efficiency: faster work with fewer resources to produce a bigger outcome.

• Human Productivity is about effectiveness: knowing how to prioritize and identifying the right thing to do at a specific moment.

AI tools are presented with two analogies:

1. They are cognitive amplifiers or partners that help people use more of their cognitive powers for a bigger impact.

2. They should be viewed as an assistant chef, where the main chef (the human) is always 100% responsible for the final outcome, even if the assistant (AI) helps with parts of the process.

Filip introduced a framework to help determine when to automate (outsource entirely to AI) versus when to augment (collaborate with AI). The framework uses two axes:

1. Level of Task Complexity.

2. Level of Human Involvement (relational vs. in-person).

• Automation (Robotic Tasks): These are low-complexity, in-person tasks that can be completely outsourced to AI (e.g., saving a document in multiple folders). The mistake people make is trying to automate highly relational tasks (e.g., a LinkedIn message to a connection).

• Augmentation (Leading Tasks): These are high-complexity, highly relational tasks that should be done in collaboration with AI, but never fully automated (e.g., agreeing on a strategic partnership or creating a company-wide strategy).

Filip noted that the initial push to implement AI often comes from the CEO or the HR/Learning segments of the company, or the IT department.

He stressed that AI transformation is a holistic process that involves:

1. Culture.

2. Upskilling of employees.

3. Process transformation.

Companies often go wrong by only focusing on one segment, such as simply buying licenses and expecting HR or IT to "push" the AI without a clear strategy or KPIs.

To properly automate a task, employees should have at least basic knowledge of the tool and know the task well themselves.

Filip is currently excited about and experimenting with AI agents (like Manus), which demonstrate how many complex tasks can be handled by these tools. He's also focusing on using tools like Zapier to automate internal, low-complexity, non-human processes within Mediartem AI.

In this episode, I host Gael, the VP of Strategic Marketing at Shalion, a company specializing in marketing intelligence that helps brands optimize their e-commerce strategy. We discussed his recent presentation at the ShopTalk Europe conference in Barcelona.

Key talking points:

• The Digital Shelf: We started by defining the "digital shelf," which is the digital version of a physical store shelf. It's how online retailers present products to their customers, with an emphasis on a vertical, one-dimensional layout and search, unlike in the physical world.

• The Generative AI Revolution: The main topic of discussion was how generative AI and tools like ChatGPT are changing the way we shop online. Users are no longer just searching with keywords but are asking complex questions in natural language. New AI shopping assistants like Rufus from Amazon and Sparky from Walmart are emerging and will play a key role in the future.

• From SEO to GEO: This shift creates the need for a new type of optimization called GEO (Generative Engine Optimization). Success in traditional Search Engine Optimization (SEO) no longer guarantees visibility in AI-generated answers. Brands must adapt, as AI engines use different sources and criteria.

• Practical Advice for Brands: Gael offered three concrete steps that companies can take today:

  1. Experiment with New Tools: Start using and analyzing the performance of AI assistants like Amazon Rufus to understand how they recommend products.

  2. Understand AI's Sources: Research which sources of information generative models use. It often turns out that in addition to e-commerce sites and specialized media, platforms like Reddit and YouTube carry significant weight. Presence and opinions on these channels are becoming crucial.

  3. Technically Adapt Websites: Make your websites accessible and readable for AI agents, which may in the future make purchases independently on behalf of users. This includes implementing protocols like MCP (Model Context Protocol) and clearly structuring product information in a question-and-answer format.

• Future Challenges: We also touched on the challenge this development poses for smaller, less-established brands. Since AI engines often recommend only a small number of verified and well-reviewed products, breaking through to the top will become even more difficult.

In this episode, I welcomee Philipp Krenn, Director of Developer Relations at Elastic, to discuss the evolution of search and information retrieval in the age of AI.

Philipp delves into Elastic's pioneering tools like Elasticsearch and Kibana, explaining how they address search challenges in logs, observability data, and security through effective information retrieval techniques. While the AI buzz brings new attention and investment to search technologies, Philip emphasizes the importance of leveraging traditional keyword search and newer vector search methods to enhance retrieval processes.

Join us as we explore the dynamic role AI plays in search, the potential challenges in adopting AI indiscriminately, and practical ways businesses can integrate Elastic's solutions into their workflows. Discover how Elastic continues to innovate and future-proof search solutions in a rapidly evolving tech landscape.

You can find Philipp at xeraa.net or on Linkedin.

Watch the whole workshop on YouTube.

In this episode, we talk to Fatima Sumroo (Linkedin), a data engineering and analytics consultant at Argusa, about the impact of generative AI, particularly LLMs, on data-related jobs. Fatima shares insights from a systematic study conducted at Argusa, revealing surprising findings about which tasks are most likely to be augmented or replaced by GenAI. The discussion touches upon the importance of integration, the hype cycle surrounding new technologies, and practical examples of using AI tools like Grammarly and Copilot.

Check out Fatimas presentation:

Other interesting things mentioned:

Tobias Zwingmann - 7 Deadly Sins of AI - Presentation

This episode features Giulia Solinas (linkedin) (the Al a la Carte Newsletter), an AI and Data Science consultant at CROZ, a Croatian company providing business and technology consulting services. Julia discusses her work at CROZ, focusing on programming solutions that enable LLMs to work with business requests. She also talks about her transition from academia to business, highlighting the bridge between the two and how her academic experience has helped in her current consulting role. The main topic of the episode is Julia's presentation "Unlocking the Power of AI Assistants" at the AI Weekend 2024 Media Festival (part of the Rovinj Conference Weekend Media Festival 2024 ). Julia summarizes the presentation, emphasizing the challenges and considerations of implementing AI assistants in production for organizations. Key discussion points include:

• Topicality of AI Assistants: The widespread belief that AI will impact businesses and the potential for AI assistants like Copilot to significantly increase productivity.

• Types of AI Assistants: The difference between general AI for everyone (like Copilot) and customized AI assistants tailored for specific company needs.

• Challenges of Implementation:

  • Identifying users and their needs.

  • Integrating the AI assistant with existing infrastructure (legacy and non-legacy systems).

  • Ensuring scalability to other departments and use cases.

  • Addressing governance and compliance, particularly GDPR and the EU AI Act. These considerations include data anonymization, masking, location, and usage.

• An Example Use Case: An AI assistant developed for a large bank's HR department to handle employee requests and save HR personnel time.

• Looking Ahead to 2025: The prediction that 2025 will be the year of agentic AI. Julia anticipates the increasing complexity of AI systems, with multiple models interacting and integrating with each other.

• AI Literacy in Organizations: Julia stresses the need for AI literacy across all levels of an organization, from employees to C-level decision-makers, suggesting training and leveraging online resources. This includes fostering trust in AI, facilitating familiarity with AI tools and resources, reducing fear of job displacement, and empowering informed decision-making about AI use cases.

• Ethical Considerations: The importance of understanding ethical issues related to AI, something that is often overlooked but crucial to consider from both technical and user perspectives.

• Julia's Current Pet Project: A GitHub repository focused on personal training on InstructLab, a methodology co-developed by Red Hat and IBM for fine-tuning smaller LLMs. This project stems from Julia's enthusiasm for open-source models and their potential for cost-effective and efficient AI solutions in production.

Key Takeaways:

• Implementing AI assistants is a complex process requiring careful planning and consideration of various factors.

• AI literacy is essential for organizations to maximize the benefits of AI and address potential challenges.

• Ethical considerations are crucial in AI development and deployment.

• Open-source models and fine-tuning offer promising avenues for cost-effective and efficient AI solutions.

Jason Schemmel, a marketing expert from Get Social Done, joins us to discuss how AI is revolutionizing the marketing landscape. He shares his insights from a recent AI conference presentation, focusing on practical AI tools and tactics businesses can implement to improve operations, increase sales, and decrease costs. Jason emphasizes the power of AI for research, ideation, and time-saving, highlighting Perplexity AI as a particularly valuable tool. He also discusses the importance of verifying AI-generated content and maintaining authenticity.

You can find the slides here.

Key Takeaways:

• AI is rapidly becoming essential for businesses, much like the internet was years ago.

• AI can be a massive time-saver for tasks like ideation, competitor research, and content creation.

• Perplexity AI excels as a research tool, sourcing and citing information from across the web.

• It's crucial to verify AI-generated content and maintain your brand's authentic voice.

• The future of large language models may involve a single, dominant platform like Google or Apple for AI.

Mentioned AI Tools:

• Perplexity AI: A powerful research tool that sources and cites information from the web. (https://www.perplexity.ai/)

• Claude: An AI chatbot known for conversational abilities and detailed responses. (https://www.anthropic.com/index/claude)

• ChatGPT: A versatile large language model used for various tasks, including content generation. (https://chat.openai.com/)

• Gemini: Google's large language model. (https://blog.google/products/gemini/)

• Copilot: An AI pair programmer that assists with coding tasks. (https://github.com/features/copilot)

Other Mentioned Resources:

• Get Social Done: Jason Schemmel's marketing agency. (https://getsocialdone.com/)

• Grow with AI Summit: The conference where Jason presented (https://growwithaisummit.com/free-clone-page)

• Claude Projects (now called Claude 2.1): A feature in Claude that allows users to upload documents and use them as a knowledge base.

• Perplexity Spaces: Similar to Claude Projects, this allows users to upload documents for Perplexity to use as a context. (https://www.perplexity.ai/spaces)

Connect with Jason Schemmel:

• Facebook

• LinkedIn