Analyze business activity against stated priorities, then build a roadmap from the claims that survive source checks.
Source episode · 07:03
How I AI: My First Impressions of Claude Opus 4.8 – Coding, Strategy, and Where It Shineswith Claire Vo
07:03 to 09:31: Claire gives two Claude models the same three months of business context, compares their time and priority analysis, and challenges an unverified roadmap.
A source-linked priority analysis and a practical roadmap based on verified business conditions.
Anthropic AI assistant
Anthropic AI assistant
Anthropic AI assistant
Step by step
Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.
5 steps
In Claude Cowork, grant the model access to your relevant business context from the last three months. This can include files, messages, and other data sources that reflect your work patterns and priorities.
Load the same business records, date range, priorities, and metric definitions for each run.
Prompt the model to analyze how you are spending your time versus where your priorities should be to achieve significant business growth. The goal is to get an initial assessment of the alignment between your activities and strategic goals.
Calculate time spent by category and compare it with the stated priorities, citing the records behind each estimate.
Based on the initial time vs. priority analysis, ask the model to synthesize its findings into a written strategy and a concrete roadmap. The output should be a specific, actionable plan.
Propose a strategy and roadmap using only verified findings, with tradeoffs and measures.
Run the same prompts with both Claude Opus 4.7 and Opus 4.8 to compare their performance. The author found that Opus 4.7 produced a more structured, data-anchored analysis, while 4.8 was more vague and prone to 'narrow vision'.
Run the identical prompt in the comparison model and tabulate material differences in evidence and conclusions.
Critically review the model's output. Challenge its conclusions by asking for the specific data it used in its analysis, as the newer model may hallucinate or fail to perform necessary checks.
Open the cited records, recalculate key numbers, and remove any unsupported roadmap item.
Turn an idea into a PRD, user stories, and a plan.
After the steps
How to recover when the loop fails and where human judgment helps.
Recover
Keep building
Use a hybrid AI workflow to bridge design and development. Generate high-level design concepts and tokens with Claude, then use Codex for robust implementation of the technical components in your codebase.
Implement a QA loop for AI-generated code using Devin. This playbook combines automated code reviews with an agent-recorded video walkthrough for visual verification, ensuring high-quality merges at scale.
Build a reusable 'Watchdog' playbook in Devin to automatically monitor customer accounts. The agent proactively checks for errors, synthesizes a prioritized list of critical issues, and saves you time.
Join 100,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.