A deeper companion to Zach’s Guide to AI Prompt Engineering.
Better prompting is the starting point. Better AI workflows are the next step. This guide helps teams move from asking better questions to using AI tools more strategically, responsibly, and consistently at work.
This companion builds on Zach’s quick-start guide, which introduces the core habits of better prompting: defining the output, adding context, increasing clarity, and using feedback to improve results.
Use the guide for:
Output, Context, Clarity, and Feedback are the foundation of stronger prompting. Click each card for a practical example.
Be clear about what you want the AI to create.
Try: “Draft a one-page memo with a recommendation, rationale, risks, and next steps.”
Give the AI the background it needs to respond well.
Try: “The audience is senior leaders at a 500-person healthcare organization.”
Remove ambiguity. Say what to include, avoid, prioritize, and format.
Try: “Use plain language. Avoid jargon. Include three options ranked by ease of implementation.”
Use the first answer as the beginning of the process.
Try: “Revise this for a more skeptical audience and identify assumptions I should verify.”
Use the 4 pillars to turn a vague request into a stronger starting prompt.
Copy this prompt into your AI tool, then paste in the starter prompt you built above.
The best AI results rarely come from a single perfect prompt. They come from choosing the right tool, giving it useful context, setting clear instructions, and reviewing the output with human judgment.
Prompting focuses on the immediate request: the wording, the task, the tone, the format, and the next revision.
Workflow thinking considers the tool, source materials, reusable instructions, privacy boundaries, review process, and how the output will be used.
What are we asking?
Which AI tool fits the task?
What information should ground the answer?
What should stay consistent?
How will humans check it?
Personalization helps move AI chat from generic to useful by giving the tool recurring context about your role, preferences, audience, and work.
Reusable preferences for how the tool should respond, such as tone, format, audience, or level of detail.
Facts the tool may remember across chats. Review them periodically and remove anything that is not useful.
Past conversations that may influence future responses, depending on your tool and settings.
Chat is useful for quick, one-off work. Projects, notebooks, and connected files become more useful when you need persistent context or repeated reference materials.
Chat is often the simplest option for quick ideas, drafts, summaries, and one-off questions.
Retrieval-augmented generation, often called RAG, helps AI answer from a specific knowledge base or set of documents. It can make outputs more useful, but it does not remove the need for human review.
The user asks something.
The tool searches trusted materials.
Relevant context is pulled in.
The AI creates an answer.
A human checks the result.
Projects and notebooks are especially helpful when teams need AI to reference the same approved documents, templates, policies, or examples across sessions.
Store approved policy language, benefits information, and templates so outputs stay consistent across roles and requests.
Reference competitor briefs, customer profiles, and battle cards for pre-call prep and account planning.
Load budgets, prior reports, and variance notes so teams can ask better questions across periods.
Use SOPs, runbooks, workflow specs, and process documents as a persistent reference set.
Compare new materials against standard templates, required language, or review checklists.
Reference runbooks, escalation guides, known issues, and service desk knowledge bases.
AI can help generate formulas, clean data, produce visualizations, and run analysis. Clear instructions make the work easier to verify.
AI can accelerate thinking, drafting, analysis, and decision support. It can also create risks when outputs are accepted too quickly or used without review.
Use this quick check to identify where your team may need clearer guidance, shared standards, or live training.
Northeastern’s AI education programs can be customized around your team’s tools, roles, policies, and real work.
Northeastern’s AI learning programs help teams build practical AI skills through instructor-led, hands-on learning shaped around your organization’s roles, tools, goals, and workflows.
Designed for real-world application.
Customized to your context.
Built to keep momentum going.