Iteration Planning Techniques

From Sprints to Flow Cycles: Re-thinking Process for AI-Native Development Agile solved 2001 problems—high failure rates, limited access to know-how, manual coding. And it eventually turned into big "A" Agile with an entire industry built around it. 2025 problems are different. Why 𝘤𝘭𝘢𝘴𝘴𝘪𝘤 𝘴𝘱𝘳𝘪𝘯𝘵𝘴 𝘢𝘳𝘦 𝘣𝘳𝘰𝘬𝘦𝘯 with AI development 2001+ ▶️ Manual coding ▶️ 2 week feedback loops ▶️ Big teams 2025 ✅ Generate code & refactor real time with AI (this in-loop refactoring is key) ✅ Feedback loop of minutes or hours ✅ 1 or 2 humans orchestrate multiple LLM teammates Enter 𝗙𝗹𝗼𝘄 𝗖𝘆𝗰𝗹𝗲𝘀 (our process at momentiq.ai) 1. 𝗪𝗮𝘃𝗲 → flexible goals for the big iteration. Wave planning produces Objectives, Dev Plan, Playbook. 2. 𝗙𝗹𝗼𝘄 𝗖𝘆𝗰𝗹𝗲𝘀 → iterations within Waves, guided by the Dev Plan, runs hours to days: • Plan with Grok / ChatGPT • Code in Cursor + LLM fit for purpose (Gemin, o3, etc) • Test, reflect, update docs 3. 𝗣𝗹𝗮𝘆𝗯𝗼𝗼𝗸: Guard-rails → architecture rules & tech constraints. Critical input to LLM prompts What is gained - what we constantly strive for in software development 𝗦𝗽𝗲𝗲𝗱: Shipping usable product every day 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻: Space for AI-driven research within the build loop 𝗩𝗶𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆: A team of one or two humans with AI teammates that moves extremely fast, makes decisions locally, updates status in living docs alongside code It's easy with AI programming to ditch all process and simply "vibe" but I don't think that's the way. With AI teammates ready to do whatever we ask, devs can go on crazy tangents - I went on a few 😅. The iterative nature of software development still holds true - it's a creative endeavor. And it's still very true that plans are useless, planning is invaluable. 🚀 In a world where opportunities from AI innovation increase week by week, I believe process must prioritize 𝘧𝘭𝘰𝘸 & 𝘭𝘦𝘢𝘳𝘯𝘪𝘯𝘨 over rigid cadence. This use of Waves, Flow Cycles, and Playbooks has worked well for me, adapting to the new AI development paradigm. Question → Have you improved dev processes alongside the introduction of AI programming? Are you still sprinting or have you started experimenting with something like Flow Cycles?

90% of startups don’t fail because of: Bad marketing, a weak team, or even a poor product. They fail because they lack a repeatable decision-making process. Here’s the framework I use to make better, faster decisions in business. I call it “The Iteration Loop.” It’s a structured way to identify what’s working, what’s broken, and what to do next, without getting stuck in endless guesswork. It gives you a systematic way to eliminate bottlenecks, optimize execution, and scale with clarity. Here are the 6 phases: 1. Bottleneck Identification 2. Clarifying the Goal 3. Solution Brainstorming 4. Focused Execution 5. Performance Review 6. Iterate & Improve 1️⃣ Bottleneck Identification Before you can fix anything, you need to identify the real problem. Most entrepreneurs spin their wheels solving the wrong issues because they never dig deep enough. To get clarity, ask: + What's the biggest constraint stopping growth right now? + What metric, if doubled, would create the biggest impact? + What’s preventing us from getting there? If you don’t identify the root problem, every solution you apply will be wasted effort. 2️⃣ Clarifying the Goal Once you know the problem, define the exact outcome you’re solving for. I use a simple Three-Part Goal Formula: 1. What are we trying to achieve? 2. By when? 3. What constraints do we have? Vague goals lead to vague actions. Precision forces progress. 3️⃣ Solution Brainstorming Now, generate every possible solution—without filtering. Most people limit themselves to their existing knowledge, which is why they get stuck. Instead, ask: “If there were no rules, what would I do?” This opens up better, faster, and often simpler solutions you wouldn’t have otherwise considered. 4️⃣ Focused Execution Don’t test everything at once—test one variable at a time. Most teams waste months by making too many changes at once, leading to messy, inconclusive results. Instead, break it down: 1. Test one key assumption. 2. Measure one KPI that proves or disproves it. 3. Execute for a set period, then review. 4. Speed matters. Complexity kills momentum. 5️⃣ Performance Review Your data isn’t just numbers—it’s feedback on your decision-making process. Your job is to analyze: + Did the solution work? + Why or why not? + What does this tell us about our business? Every test refines your ability to make better future decisions. 6️⃣ Iterate & Improve Most companies don’t fail from making the wrong move—they fail from making no moves at all. The only way to win long-term is to keep iterating. Instead of fearing failure, build a culture that rewards learning. Failure + Reflection = Progress. If you aren’t improving your decision-making process, your business will eventually hit a ceiling. That’s why I built The Iteration Loop—so every problem becomes an opportunity for better, faster execution. P.S. If you want the scaling roadmap I used to scale 3 businesses to $100M and beyond, you can get it for free from the link in my profile.

This slide gets copied and stolen from me more than any other. It’s the blueprint for saving 4+ years and $4+ million on failed AI initiatives. Start with an iterative PMPV framework to avoid 4 expensive mistakes. Propose – Top-down and bottom-up opportunity discovery workshops. The business articulates its needs vs. being told what should be built. The opportunity is assessed. Does it require AI, or can a less expensive technology work? Measure – AI Product Managers work with stakeholders/customers to define the problem space and assess the opportunity size. They work with the data/AI team to assess feasibility and estimate costs. Prioritize – The 3 assessments allow the business to reach a consensus on a value-based prioritization without being dragged into technical solution complexity. The roadmap is updated. Validate – Did the initiative deliver the expected impact, revenue, margins, etc.? If not, why, and is it salvageable? If it did, can more value be delivered quickly? How much? The roadmap is updated/reprioritized. The roadmap can’t be static. New opportunities emerge, and some opportunities don’t pan out. Businesses need to take a pipeline approach with multiple opportunities on the roadmap. It can’t be opinion-driven or abandoned for every fire drill. Opportunity size estimation is critical, or the loss from constant reprioritization cannot be quantified. Loss allows AI Product Managers to push back. That’s it. Iterative PMPV is a lightweight product strategy framework that supports the unique needs of AI features and products. Remember, frameworks are only as good as the people who manage them. No AI Product Manager == No AI products, revenue, or cost savings…just a giant cost center. #ProductManagement #AIStrategy