The PM's New “Basic”: Automating Workflows with Conversation Memory
Hey Product People!
How often do you feel like you're drowning in a sea of repetitive tasks? Constantly juggling priorities, drafting endless documents, and summarizing countless meetings? It's easy to get lost in the busywork instead of focusing on the true product strategy you signed up for. This blog is where we'll explore practical ways to shift that dynamic, together.
Before we dive deep into advanced AI integrations and complex automation setups (and trust me, we will!), I want to take a step back. This blog is about mastering the basics. Because the truth is, the most powerful productivity gains for PMs often come from leveraging the most basic tools you already have access to, right now. That's why we're starting with the foundational tools: ChatGPT, Claude, Gemini, Grok etc. These conversational AIs are your untapped Productivity Multipliers.
But here’s the secret: the real magic isn't in asking a single question. It's in teaching these AIs to remember who you are, what you do, and how you like things done. This is the PM's new "basic" skill—mastering Conversation Memory.
Think about having a great new colleague. On their first day, you explain your role, your team's goals, the preferred format for reports, and the key stakeholders. You wouldn't repeat all of that context every single time you asked them for help, right? Once they "get it," they remember and apply that understanding to all future tasks. Conversation Memory in AI is exactly like that.
This capability is the foundation of automating PM workflows without complex coding. It means you can define your "rules of engagement" just once, freeing you up to focus on the actual variables of your daily work – the new feature idea, the raw user feedback, the specific meeting notes.
Exercise: Your AI-Powered Discovery Agent in Action
Imagine you've just kicked off a new product initiative – say, building a Money Management app. Before even thinking about features, you, as the PM, manually need to establish a ton of context like :
Who are your target users (e.g., "young professionals seeking financial peace")?
What's your primary goal (e.g., "help users budget proactively")?
What's your preferred prioritization framework (e.g., "Reach, Impact, Confidence, and Effort (RICE) scoring, prioritizing low-effort features")?
How do you want the output structured (e.g., "a table with problem, solution, RICE, and summary")?
Traditionally, you'd hold brainstorming sessions, document these constraints, and then manually apply them to every idea or piece of feedback. It's a critical but time-consuming setup.
Instead of all that manual re-contextualization, we can leverage AI's Conversation Memory to automate this entire setup. To unlock this power, we create a dedicated chat thread—your "PM Agent"—that automatically stores all your repetitive context and formatting rules. This is your Anchor Prompt.
Step 1: Set Your Anchor Prompt (The One-Time Setup)
In a new, dedicated chat thread (label or pin it for future use!), you input a detailed "Anchor Prompt." This is where you permanently establish the AI's Role, Output Format, and Constraints for that specific conversation. Think of it as onboarding your new, highly efficient AI assistant.
Example Anchor Prompt:
"From now on, you are a Product Manager at a B2C FinTech startup. Your target users are young professionals (25-35) who feel financially anxious and need simple tools for proactive money management.
Your primary task in this chat will be feature discovery and prioritization. For every problem I give you, you must generate a table with four columns: 1. User Problem, 2. Feature Idea, 3. Estimated RICE Score (R: 1-10, I: 1-10, C: 1-10, E: 1-10), and 4. 2-Sentence Executive Summary.
All feature suggestions must prioritize low development effort (RICE Effort must not exceed 6) and focus on automation to reduce user effort."
Once the AI acknowledges this, that chat thread is now your custom "Feature Discovery Agent."
Step 2: Input Your Single Variable (The Daily Task)
With your AI agent trained, you simply return to this specific chat thread and provide only the single variable—the core problem or idea you want to explore. The AI will then automatically apply all the predefined rules from your Anchor Prompt, every time.
PM's Single Variable Input Prompt:
"Our users have told us they constantly overspend right after payday and struggle by the end of the month."
Output:
The AI quickly creates a detailed, structured output, turning feedback into prioritized opportunities in under 30 seconds—boosting productivity 4x to 10x. Imagine this impact all week!
Leveraging Anchor Prompts eliminates the daily correction cycle: previously requiring 3-4 turns to remind the AI of its PM persona and rules, the Anchor Prompt now delivers perfect, structured output in just 1 turn, compounding productivity daily.
| Comparison Metric | Scenario WITHOUT Conversation Memory (Hypothetical) | Scenario WITH Conversation Memory (Actual Process) |
|---|---|---|
| Number of Turns Needed | 3-4 Turns (Initial prompt + 1-2 corrective prompts). | 1 Turn (The user simply poses the problem). |
| Agent Initial Output Quality | Incorrect/Incomplete: Forgot PM persona, table format, and violated the Effort (≤ 6) constraint. | Perfect: Delivered the complete, correctly formatted table, adhering to the PM persona and all RICE constraints. |
| User Effort & Anxiety | High: User must stop, remind the agent of the rules, and repeat context. | Zero: User focuses only on the next task/problem, treating the agent as a reliable tool. |
| Process Efficiency | Low. Time is wasted on formatting and correction cycles. | High. Rapid, consistent, and structured output allows for quick feature iteration. |
Beyond Discovery: Automating Other Repetitive PM Tasks
The Anchor Prompt method isn't limited to feature discovery within your money management app. You can apply this exact "set it and forget it" principle to dramatically increase efficiency for other common PM tasks:
| PM Task | Anchor Prompt (One-Time Setup) | Single Input (Daily Use) | Automated Productivity Boost |
|---|---|---|---|
| User Story & Acceptance Criteria Generation | "From now on, generate User Stories in the 'As a [User], I want [Goal] so that [Benefit]' format. Include Acceptance Criteria using Gherkin syntax (Given/When/Then). Keep stories concise and relevant to a money management app." | "Feature: Automated bill payment reminders for recurring expenses." | Instant Story Drafts: Rapidly convert high-level ideas into ready-to-refine user stories and acceptance criteria, saving hours of manual drafting. |
| Stakeholder Communication Drafting | "You are my Communications Assistant for MoneyApp Inc. Draft executive summaries for our leadership team. Tone: confident, concise. Include: Highlights, Key Decisions, Blockers, Next Steps. All updates should be FinTech-focused." | "Raw notes: Launched new budgeting feature. Met 80% of Q1 KPIs. Identified bug in savings goal display. Need decision on Q3 roadmap funding for credit score integration." | Professional Comms in Seconds: Transform messy notes into polished, structured stakeholder updates, ensuring clarity and consistency every time. |
| Competitive Analysis Summaries | "Generate competitive analysis for FinTech budgeting apps. Output a table: Competitor Name, Core Differentiator in Budgeting, SWOT Analysis (3 points each), Strategic Recommendation for MoneyApp. Focus on apps like Mint, YNAB, and Personal Capital." | "Analyze 'Mint' and 'YNAB' focusing on their latest premium features." | Structured Insights on Demand: Get instant, organized competitive breakdowns, allowing you to quickly identify opportunities and threats without manual formatting. |
What's one repetitive task you're currently doing that you're most excited to automate with Conversation Memory? Share your ideas in the comments below!
