How New Team Members Get Up to Speed in Days, Not Months

🧭 THIS WEEK AT AI SECOND ACT

Howdy, this week’s we’re talking about project onboarding. Resource churn is a huge negative weight on any organisation or project. AI can really help that process by giving instant answers to newer resources and reducing load on the current teams.

Growth of this newsletter has slowed a lot unfortunately. I need to get out a bit more on the socials and promote it. Meantime, would love it if could share it, thanks!

My goal is to make this as valuable and practical as possible as we navigate the new AI era. 🚀

🧰 AI NEWS + LEARNING

Here are a few things I found recently:

• Pretty incredible animation about the data flows of an AI query.

• Learn Your Way - Google tool transforming content into learning experiences. Google has some great learning tools, also like NotebookLM.

• My work in progress ‘About Me’ page - Next, add live updates for Strava, Beehiiv and integrate into my main site.

  • I’ve been going crazy with AI led (vibe) coding lately - that is, building stuff just using natural language, tools like Gemini CLI, Claude Code, OpenAI Codex. My GitHub profile is here showing the recent activity. Some ongoing work -

    • Moving my Wordpress site to faster, more secure static site based on Hugo.

    • Automated quality template to test/fix issues before ‘putting?’ in Github.

    • Crew AI based ‘builder crew’ that just builds stuff if you ask it - definitely work in progress/draft!

I’m filtering out a lot of AI noise, most ‘news’ is not useful or interesting in daily, weekly life. That means, a lot less ‘news. Moving forward, I’ll add more interesting and useful tips & tactics into this section. Here’s one -

• Try ChatGPT, Claude or Gemini to learn new things. Some ideas -

  • Running: "What's one technique elite marathoners use that weekend runners should try?"

  • Space: "Explain the James Webb telescope's most surprising discovery this year in simple terms."

  • Nature: "What are 3 emerging patterns scientists are seeing in how climate change affects Great Lakes ecosystems?"

🗺️ FEATURED INSIGHT

How New Team Members Get Up to Speed in Days, Not Weeks

I've been experimenting with using AI to onboard new project team members. You know the usual drill—they spend 2-3-6 months reading documents, asking questions, and slowly figuring out who does what.

Instead, I tried uploading all our key project docs to ChatGPT and letting new hires chat with it.

The results have been interesting. Not revolutionary, but useful enough that I'm expanding the approach.

The Traditional Onboarding Reality

New team member joins your project. You give them the "document dump"—charter, specs, org charts, process docs.

They spend weeks and months reading and asking questions like "Who handles approvals?" and "Why did we choose this architecture?"

The real issue: Your existing team spends hours answering the same onboarding questions over and over. Meanwhile, the new person feels stupid asking basic questions in meetings.

What I'm Testing Instead

Step 1: Create the Project Knowledge Base (2-4 hours setup)

I upload key documents to ChatGPT:

General Project Chat:

• Project charter and scope documents

• Org charts and stakeholder maps

• Process documents and workflows

• Meeting notes from key decisions

Technical Project Chat:

• Technical architecture and design docs

• Development environment setup guides (for R&D projects)

• Customer background and requirements

Then I create a system prompt that sets the context:

You're helping new team members learn about [PROJECT NAME]. 
Key context: [brief project summary, timeline, main objectives]
Key people: [PM, tech lead, customer contacts, stakeholders]
When someone asks questions, provide specific answers based on the uploaded documents. 
If you don't know something, say so and suggest who they should ask.

Step 2: New Team Member Walkthrough (1-2 hours)

Instead of handing over a reading list, I show them how to query the AI:

"Ask it anything you'd normally ask me or the team lead. Start with basics like 'What are we building and why?' and 'Who are the key stakeholders?' Then dig into your specific area."

Sample questions that work well:

• "What's the project timeline and major milestones?"

• "Who do I need to know and what do they do?"

• "What decisions have been made and why?"

• "How does our development/testing process work?"

• "What are the main technical constraints?"

• "Who handles customer communication?"

Step 3: Follow-up and Refinement (ongoing)

After a few days, I check what questions the AI couldn't answer well and either:

• Add more documents to fill gaps / or less to reduce the scope!

• Update the system prompt with better context

• Connect them with the right person for things AI can't handle

Seems better to :

• Define a narrower scope - don’t create a chat that tries to do everything, it’ll get confused and not provide the best answers

The output: New team members get answers to basic questions immediately and come to meetings with better, more specific questions.

What Actually Works (and What Doesn't)

AI is excellent at:

• Explaining project context and background quickly

• Identifying key people and their roles

• Summarizing past decisions and rationale

• Answering "where do I find..." questions

• Providing consistent information to everyone

AI is terrible at:

• Understanding office politics and personalities

• Knowing which processes are actually followed vs. documented

• Assessing project health or real risks

• Replacing human relationship building

The sweet spot: AI handles factual onboarding questions so your team can focus on strategic context and relationship building.

Time saved: New team members ask 70% fewer basic questions. When they do ask questions, they're better and more specific.

The Reality Check

This isn't magic. It's just better information access.

You still need face-to-face time to build relationships. You still need to explain the unwritten rules and company culture. You still need human mentorship.

But if you can eliminate the basic "what is this project and how does it work" questions, your team can focus on the strategic onboarding conversations that actually matter.

The key is having good documentation in the first place. AI can only work with what you give it. Garbage in, garbage out.

Weekly AI strategies for operating executives
— Brett

👉 Hit “Reply” and share your experience — I read every one!

_{Picture by Meghan Holmes on Unsplash.}