Propel Ventures

We all know the thrill of a great idea—and the terror of watching it stumble at the finish line 😫💔 That's why the next Propel Ventures + iterate meet-up is taking aim at the treacherous chasm between product dreams and operational realities. We're asking: When the rubber hits the road, how do we ensure our product doesn't blow up? 🛞🔥 Huddle around, friends, as we zoom in on the oft-overlooked human element of product delivery with an all-star lineup: 👉 Craig Rees: Global engineering leader and launch expert 👉 Shannon Tresider: Ops strategist bridging products and customer experience 👉 Simon Raik-Allen: Tech innovator and former CTO moderating the discussion And they're all set to tackle the engineering-ops ecosystem head-on. They'll be unveiling strategies for seamless launches, happy customers, and products that don't just survive—they thrive in the wild. So lock in your spot at the link in comments 👇🔗, because it's all happening at Propel Base Station next month. The countdown begins now! 📆 Thu 22 August, 4.30-7pm 📍 Propel Base Station: 7 Hill St, Cremorne #continuousimprovement #engineeringmeetup #productstrategy Michael Boyd Ryan Lynch Kyla Boyd Tatiana Ashby Hannah Campbell Stephanie Futerman Sam Bowring

We had an incredible time at the Sydney Tech Leaders Meetup! 🚀  A huge thank you to industry leaders Simran Ghumman, Mouna Neelakanta, 👋🏻 Melissa K. and Grant Burrow for sharing their insights on the dynamic relationship between Tech & Product.  Special thanks to Propel Ventures for partnering with us on this fantastic event! #TechLeaders #Innovation #TechAndProduct 

🚀 Exploring the Next Challenges with LLMs 🧠 At Propel Ventures, we constantly strive to push the boundaries of what’s possible with large language models (LLMs). We’re currently tackling two key challenges that many in our field might find familiar: 1. Fine-Tuning for Better Outputs: Out-of-the-box, ‘off-the-shelf’ LLMs are fantastic at delivering a solid 7/10 solution using Retrieval-Augmented Generation (RAG) techniques. However, achieving a 10/10 output often requires Reinforcement Learning from Human Feedback (RLHF). Unfortunately, RLHF is both expensive and time-consuming. Has anyone discovered ways to reduce the cost and effort of integrating RLHF into their LLM training process? 2. Reducing the Cost of LLMs: Scaling off-the-shelf LLMs can be costly. It seems more efficient to leverage free or open-source LLMs from platforms like Hugging Face and fine-tune them—ideally using RLHF—to achieve that coveted 10/10 performance at a lower cost. I’m keen to hear from others about their experiences fine-tuning LLMs efficiently. Are there any shortcuts or valuable insights you can share? There’s a growing need for comprehensive LLM training, fine-tuning, and human evaluation solutions to assess performance, such as: • Human Rating: Involving human judges to rate the quality of LLM outputs based on criteria like relevance, coherence, and fluency. • Red Teaming: Employing a team to actively find flaws, biases, or vulnerabilities in LLM outputs, often crafting adversarial inputs to probe the model’s behaviour. • A/B Testing: Comparing outputs from different LLM systems (or versions) through human judgments to determine which performs better for a given task. At Propel Ventures, we have some thoughts on how to tackle these challenges and will share them shortly, as we see this as an increasingly large problem that needs a solution. We’d love to hear from the community about their experiences and strategies. Let’s collaborate and drive innovation together! Stay tuned for an announcement on how we can help solve these issues.

🎤Sydney Tech Leaders Meet Up Event 🚀 Ever wondered how tech and product teams can be both best friends and arch-rivals? 🤔 Get ready to find out! 🏆 The best companies know the secret: cross-functional product teams that blend product, tech, and design talent are the key to success. 🔥 Expect a spirited conversation filled with real-life stories, challenges, and strategies for working together. Our experts will share their experiences—the highs and lows, the harmony and the friction—giving you insights on maximizing these essential partnerships. 📢 Whether you're a tech leader, product manager, or curious stakeholder, bring your burning questions and get ready to learn valuable lessons. 🗓️ When: Thursday, July 18, 2024, from 5-7pm 📍 Where: Stone & Chalk Scale Up Hub, 477 Pitt Street #Level 1, Haymarket, NSW 2000 Register here: https://hubs.la/Q02C8pS_0 #TechAndProduct #PanelDiscussion #Innovation #Leadership #CrossFunctionalTeams

🚀 Empowering Software Engineers with Generative AI at Scale! 🚀 I'm excited to share our latest blog post, which dives into how leading companies are leveraging generative AI to power internal developer platforms, making their software engineers and data workers more efficient and effective. In "Empowering Software Engineers with Generative AI: The Rise of Internal Developer Platforms," we explore Intuit's innovative GenOS. This proprietary operating system integrates generative AI across their product teams, streamlining development and institutionalising organisational knowledge. GenOS addresses key challenges like accelerating product development, integrating domain-specific knowledge, and ensuring responsible AI and data governance. It provides a comprehensive framework that teams can build upon without starting from scratch every time. At Propel Ventures, we specialise in developing these AI operating systems and their components, helping our clients harness the full potential of generative AI safely and efficiently. Read the full blog post here https://hubs.la/Q02yXzGC0 and discover how you can leverage generative AI to transform your software development processes. I'd love to hear your thoughts and experiences in the comments! More about the Intuit platform is here https://hubs.la/Q02yXzpK0

🚀 Balancing AI Investment: How to Stay Ahead Without Overcommitting! 🚀 I'm excited to share our latest blog post, which tackles the crucial decision of how much to invest in AI for software product businesses. In "Balancing AI Investment: How to Stay Ahead Without Overcommitting," we discuss the importance of going all in on AI to offer maximum benefits to users and stay competitive. However, this comes with the challenge of rapid AI advancements that can quickly render today's investments outdated. We explore Intuit's investment in their GenOS platform as a prime example. Despite the heavy investment in custom LLMs and tooling, similar capabilities will become available from major AI providers like OpenAI, Google, and AWS. This case highlights the need for flexible and incremental AI development. At Propel Ventures, we faced a similar challenge with our brand voice agent. We designed it to be easily upgradeable, allowing us to stay current with the latest AI models and technologies. This approach ensures we remain competitive without overcommitting to technology that may soon change. Read the full blog post here https://hubs.la/Q02yXCW60 to learn more about making smart AI investments that keep you ahead of the curve. I’d love to hear your thoughts and experiences in the comments!

Coming up on Thursday June 20th, Trent Conway from Propel Ventures helps us manage that shiny Ai thing that's been trying to take over your roadmap. https://lnkd.in/grK9rMtb

🚀 Balancing Automation and User Agency in AI Product Design - A Crucial Insight..... about eggs! 🚀 I'm thrilled to share our latest blog post that delves into a fundamental aspect of AI product design: the balance between automation and user agency. In "When it comes to AI product design, make sure we allow people to add the egg," we explore the fascinating story of General Mills's Betty Crocker instant cake mix from the 1950s. This historical example reveals how a simple tweak—allowing users to add an egg—transformed the product's success by making users feel more engaged and involved. At Propel Ventures, we believe that while AI can significantly enhance efficiency, it's crucial to ensure users still play a meaningful role in the process. Our latest blog post offers valuable insights and practical tips for product designers to strike this balance effectively. We are committed to helping companies create AI-driven products that not only meet market needs but also foster user engagement and satisfaction. By integrating user input into AI workflows, we can design products that users love and trust. Read the full blog post here https://hubs.la/Q02yXqTd0 and share your thoughts! How are you balancing automation with user engagement in your projects?

Last Thursday, Propel was honoured to host another engaging session in the Propel Ventures Tech Talk Series, alongside the talented team from Propel: Paul Greenwell, Trent Conway, Wilson Lilburne, Milan Modi, and Abdi Daud. Doubling Productivity with AI: Event Highlights AI's Impact on General Productivity: ------- Trent Conway delved into how AI technologies like Microsoft Copilot are revolutionising workplace productivity. By automating routine tasks such as meeting summaries and email management, employees are freed up to engage in more strategic pursuits. Select the Right AI Tool for the Job: ------- Milan Modi discussed the advantages of using Perplexity over ChatGPT and Google for exploratory research, highlighting its unique capabilities and practical applications in our projects. Innovative Applications of AI in Software Development: ------- Wilson Lilburne shared insights into doubling productivity in software projects through AI. He showcased how a custom-trained GPT for creating user stories not only enhances productivity but also improves quality, alongside generating UI components code. An Architecture Thought Partner: ------- Abdi Daud introduced his custom GPT that acts as a thought partner in defining software architecture, evolving solutions and generating C4 diagrams directly using PlantUML. Engaging Marketing: ------- At Propel, we utilise AI beyond development. Paul demonstrated our proprietary image model that produces the iconic images of Propel Spaceman for our blog and social media. Our integrated use of Zapier workflows with OpenAI and Slack maximises the reach of our posts. The aim of this event was to expand your perspectives on the diverse applications of generative AI in your current work. We also shared insights into what didn’t work, underscoring the importance of continual experimentation in leveraging AI. You can read more about the event and view a recording of the session at https://hubs.la/Q02wVbLB0 Stay tuned for more insights and feel free to share your thoughts or questions!

🚀 Learn About The Product Operating Model From The Experts - Propel Ventures! 🚀 In our latest blog post, we reflect on Propel's recent events which have led the thinking in Australia on Product Operating Models. At Propel Ventures, we're not just advocates of the product operating model; we use the tools to both implement and coach Product Operating Model for companies, and for our own business when we deliver software. Our commitment to advancing this approach is demonstrated through our series of live Product Operating Model events featuring global thought leaders like Marty Cagan, Radhika Dutt, Joshua Seiden, John Cutler, Janna Bastow, and Ant Murphy. These events have established Propel as the thought leader and the leader of implementing these techniques in practice. At Propel, we also practice what we preach. We don’t just teach companies how to operate like a product business; we ourselves develop software using the very principles of the product operating model. This approach, often referred to as 'eating our own dog food,' allows us to continuously refine and perfect our methods, and apply these strategies directly to our own software development methodology, ensuring every project meets market needs on time and within budget. Take a look at our books and our speakers in our latest blog post https://hubs.la/Q02w6ksP0