Skip to content
AI Your AI Skills Lab
Newsletter
🔵 Practitioner

Workshop Google Workspace + AI — field notes from 9 classes

Companion to the SkilLab Google Workspace + Gemini workshop. What works, what surprises, what fails, and the critical question before contracting.

· 8 min · workshop-companion

What this post is

Companion to the Google Workspace + AI Workshop by SkilLab. Field notes distilled from 9 classes run between 2024-2026, primarily in startups, agencies and professional studios (Google-first audience).

Shorter than the Copilot post because classes have smaller scale (Google-first in mid-market is more “lean SMB” than enterprise).

What works in 100% of classes

1. NotebookLM is where participants “lose their minds”

Demonstrate NotebookLM with 10-15 real company documents (internal manual, past cases, specifications), and generate question + answer with exact source is the “wow moment” of most.

100% of classes react with “I need this for my team NOW”. Next question: how do I get my colleagues to use this?

2. Gemini in Docs + company voice

Train Gemini to write in company tone (by loading reference documents + few-shot examples) has consistent positive reaction. Different from Copilot, Gemini integrates Docs workflow more naturally.

3. Help Me Schedule

Simple feature — “help me find a time for meeting with X, Y, Z next week”. Accelerates every team intensive in meeting scheduling (sales, customer success). 100% of sales/CS participants adopt immediately.

4. Audio Overview in NotebookLM

When we demonstrate NotebookLM generating “podcast summary” of the material in target language — reaction is mixed but all laugh/are surprised. Real corporate use is limited, but becomes workshop icon (“was when I saw AI talking to me”).

What surprises 70% of classes

1. AI Mode in Drive searches data inside documents

Most didn’t know Drive already has deep AI search inside files. When they discover they can ask “which is the report I mentioned about client X in January 2024?” and get response with the right doc, becomes immediate daily tool.

2. Real quota limit in Gemini for Workspace

In 5 of 9 classes, team discovered mid-workshop they were hitting Gemini limit on the plan they had. Workshop needs to prepare this — check beforehand which plan (Business Standard, Plus, Enterprise) and corresponding quotas.

3. Gemini Code Assist is NOT included in Workspace AI Add-on

Devs on the team get excited about the idea. Discover it’s a separate license. Frustration.

4. Audit log in Gemini Enterprise is real differentiator

For teams in regulated sector (legal, accounting, health), Gemini Enterprise’s granular audit log is what justifies the upgrade. Workshop needs to expose this explicitly.

What fails in 20% of classes

1. Hybrid company (Microsoft + Google)

3 of 9 classes came from companies using 50/50 Office 365 + Google Workspace. Workshop becomes half-half — doesn’t accelerate Workspace because team doesn’t use full Workspace. Mitigation: prioritize Workspace or Copilot, don’t try both.

2. Technical team that prefers CLI/API

Engineering teams sometimes prefer to use Gemini API direct via Python/CLI instead of Gemini in-app. In-app workshop becomes less interesting for that profile. Mitigation: redirect to Claude Cowork workshop or customize.

3. NotebookLM hesitation due to data protection

Regulated-sector company’s DPO/legal sometimes halts NotebookLM rollout. NotebookLM Enterprise solves but requires upgrade. Workshop needs to prepare this path.

The critical question before contracting

“Does your company want depth in Gemini in-app OR in NotebookLM enterprise?”

  • Gemini in-app: 4h workshop covering Docs/Sheets/Gmail/Meet. Adoption in 2-4 weeks post.
  • NotebookLM enterprise: 4-6h workshop focused on designing knowledge workspaces for 3-5 departments. Adoption can take 6-12 weeks because depends on content curation.

Both are valuable, but different time and investment.

Where to go deeper between workshops

Next steps

By Ivan Prado · SkilLab AI · May 2026. Translated and adapted from the PT-BR original.