AI has become an indispensable tool for increasing productivity and effectiveness in the workplace. This presentation explores practical ways individuals can harness AI models to transform their daily workflows, focusing on real-world applications and actionable lessons.

We begin by examining the use of internet-based GPT models for code generation and validation. From writing efficient scripts to debugging existing code, these models can simplify tasks significantly. A key takeaway: for many scenarios, code is rarely intellectual property, allowing safe and effective use of online AI tools.

Next, we discuss how GPT models can facilitate learning and conceptual discussions. Whether you’re brainstorming ideas or sense-checking complex topics, AI-powered discussions can provide valuable insights and enhance your understanding.

For tasks involving sensitive or proprietary data, we explore the advantages of local AI models. We will discuss considerations like GPU versus CPU-based processing, model size, and methods like embeddings and retrieval-augmented generation (RAG). The key lesson here is determining what you actually need from an AI model to maximise efficiency without overcomplicating your setup.

Finally, we dive into specific use cases that highlight the versatility of AI in personal workflows, including:

Key Takeaways:

Internet-Based GPT Models for Code Generation and Validation

AI can streamline code generation and debugging processes for languages like ABAP, Python, PowerShell, and Bash.

Takeaway: Code is rarely intellectual property, often making it safe to use internet-based AI tools for many scenarios.

Internet-Based GPT Models for Discussions and Learning

GPT models can facilitate conceptual discussions, sense-check ideas, and enhance learning.

Takeaway: Use GPTs to discuss and refine ideas, ensuring better understanding and decision-making.

Local AI Models for Sensitive Data

Local models are ideal for tasks involving sensitive or proprietary data, with considerations for GPU vs. CPU, model size, and retrieval-augmented generation (RAG).

Takeaway: Determine your actual requirements to balance efficiency and complexity when choosing AI tools.

AI Use Cases in Workflows

Practical applications of AI include:

SAP Performance Tuning and Parameter Checking

Writing clear and concise emails

Generating and validating code across multiple languages

Streamlining error checking and debugging processes

Takeaway: AI tools can enhance effectiveness across diverse tasks, unlocking new levels of productivity.

Deploying AI in a support organisation can be transformative, but the journey is rarely straightforward. In this session, we share our lessons learned—the challenges we faced, the mistakes we made, and how others can avoid similar pitfalls. This is not a theoretical guide, but a candid exploration of our experience in bringing AI to life within our organisation.

We start by discussing how we defined what we wanted to achieve with AI. This process revealed hidden opportunities and forced us to confront organisational assumptions. Our key takeaway: identifying where AI can genuinely add value is the first critical step.

Next, we reflect on selecting a platform. Navigating data constraints and access challenges shaped our decisions and taught us that no platform is perfect. Attendees will gain insight into the compromises and considerations required to make this decision work for them.

A chatbot use case then highlights one of our biggest lessons: the absolute importance of data. Without the right data, even the best AI struggles. We’ll share what went wrong in our approach and how better data preparation could have avoided issues.

Finally, we examine the deployment of basis agents—AI tools capable of taking action. Here, we learned the hard way about the risks of overreach and the need for careful control over automation levels.

This session is ideal for those embarking on their own AI journeys, offering real-world insights into the opportunities, risks, and rewards of AI adoption. Join us to learn from our missteps and to gain a better understanding of the complexities of deploying AI in a support organisation.

Key Learnings

Defining Opportunities

The importance of identifying clear and actionable AI opportunities within your organisation.

Understand where AI can genuinely add value and challenge existing assumptions.

Platform Selection

Be aware of constraints related to data availability, governance, and access when choosing a platform.

Recognise that compromises are inevitable, and no platform is a perfect fit.

Chatbot Use Case

Data quality and availability are critical to AI success—poor data undermines even the best AI models.

Invest in data preparation to avoid costly missteps.

Deploying Basis Agents

Carefully control the level of automation to avoid overreach and unintended consequences.

Establish clear boundaries and safeguards for AI-driven actions.