
Sabrina Ramonov makes the first Grok Bot a chief of staff that interviews you, suggests the bots you need, then creates them and walks you through connecting apps.
Gemini Decision 1 answers yes/no, choice and score questions about text, images, audio, PDFs and video with a confidence score and zero output tokens, so your code can route, gate and audit automatically.
Prompt Engineering · YouTube · published Nov 13, 2026 · Automation & Agents
▶ Watch on YouTubeUsing the Gemini Decisions API (model gemini-decision-1), check this customer message with three questions: did the item arrive damaged (yes/no), which support queue fits (choice, include an 'other' option), and how frustrated is the customer (score). Only act automatically if confidence is above my cutoff; otherwise escalate to a person.
Gemini Decision 1 is a decision model: it never writes text, it answers yes/no, choice or score questions about text, images, audio, PDFs and video, and returns a probability plus a confidence score. Your code compares that to a cutoff and acts, or escalates to a human or a bigger model.
The problem: Using a full LLM for small yes/no decisions means paying for thinking and output tokens, waiting for them, and then parsing the text that comes back.
The free tier of Google AI Studio is enough for every example in the video, and the code is in a Colab linked in the description.
Examples built in the video: a browser agent that completes a checkout, an expense-claims auditor, and a voice agent.
0:00What Gemini Decision 1 does0:56Demos: checkout agent and expense auditor1:43How decision models differ from LLMs3:11Yes/no, choice and score questions4:36Best use cases6:14Cost versus a Flash model7:39First API call9:11Images, audio, PDF and video inputs11:00Mistake 1: no way out12:16Mistake 2: too much context14:11Mistake 3: big questions16:03Voice agent and when to use an LLMDistilled from the video’s captions

Sabrina Ramonov makes the first Grok Bot a chief of staff that interviews you, suggests the bots you need, then creates them and walks you through connecting apps.

Tim clones his voice in AssemblyAI, then builds a Python agent that listens, thinks with an LLM and answers back in his cloned voice in under 100 lines of code.

Jeff Su shows six durable habits: convert PDFs to markdown, give AI a folder index, save prompt shortcuts, run an intent check before big tasks, use an API key as a spare entrance, and keep one rulebook file per connected tool.

Oli Morris builds an n8n workflow that fetches Berlin's weather at 8 a.m. every day, turns it into a one-line report and saves it to a data table, first by prompting the n8n Assistant, then node by node by hand.