Speaker Prep
Thank you for joining us and taking the stage!
We are grateful for your time.
Below is everything you need to know about the event, the audience, and how to prepare.
Email with questions: madeline@trueventures.com
Hands on Keyboards. Follow Along Building.
120+ Founders, CEOs, builders and start-up leaders.
This is our second run of the format. V1 was an AI Maximalist summit in April. V2 is bigger and, for the first time, we're expanding the room.
Check Your Calendar Invite
You should see a calendar invite for your specific demo time. Please arrive at least 15 minutes before your demo time slot.
We'll save 5 minutes at the end of your time for questions from the room.
From All Areas of Building
The room spans founders, CEOs, CTOs, PMs, designers, engineers, people leaders, marketers, and operators across the company-building stack.
Think of the room in two tracks. Track A has local tools open: API keys, terminals, Claude Code, Codex, Gemini CLI, repos, and commands. Track B is building through chat: Claude.ai, ChatGPT, Gemini, copy/paste, uploads, and browser workflows.
Same end state, different on-ramps. Make both tracks feel successful and in the right room.
By experience with AI, everyone is experimenting. Some are building at the frontier. Others use chat tools every day but have not built an agent yet. Everyone is here to learn, build, and go deeper.
The most important thing we learned in April: if your demo only lands for the experienced half, you've lost the room. You have to land for both.
One Workflow. Clear Enough to Follow.
We sent over your demo focus area via email. Pick one of your real workflows and make the pattern useful for the room.
Keep it small enough for people to build the first version with you. Even if they won't use the exact workflow, they should leave knowing how you built it and what to bring back to their team.
Send any pre-work for the audience to madeline@trueventures.com at least 1 week in advance of the summit. Include anything they need to sign up for, download, test, do deep research on etc. Think about what would set someone up for success so they can build alongside you without too much time prepping the workflow.
Focus on the workflow, not the content. Shared sample content is fine if it helps people learn the pattern before applying it to their own work.
Run your session in this order:
- Set context. Who you are, what the workflow is, and what pain it removes.
- Show the leverage. Run the real thing and show what AI makes faster, cheaper, repeatable, or possible.
- Build it slowly. Rewind, walk step by step, and pause often.
- Help people make it theirs. Show how the pattern translates to their role, company, or use case.
- Be honest. Share token cost, build time, what broke, what is not automated, and where a human still sits in the loop.
We'll save 5 minutes at the end of your time for questions from the room.
The bar for your session: attendees leave with a workflow running. Not "I learned something." Built. Running.
What They Want to Learn and Build
We asked what people want to build. Here is the room in one format.
What they want to build: Outbound, CRM cleanup, agent-driven selling, marketing automation, lead gen, web scraping, intent signals, competitive intel, and content systems.
Direct quotes:
- "Automating GTM research, automating CRM, building and updating internal/institutional knowledge repositories"
- "An agent selling a complex product in real-time, without 20 questions."
What they want to build: Agents managing real projects end to end, coordinating across tools, and handling security or operations with less babysitting.
Direct quotes:
- "Multi-agent orchestration"
- "Managing agents handling a semi-complex project end-to-end"
What they want to build: Plan, implement, test, deploy as one loop, from PM spec to shipped code, with CI validation and compounding workflows.
Direct quotes:
- "Agentic Engineering - Plan, Implement, Test, Deploy"
- "I’d like to see everyone’s PM spec -> agent build workflow"
What they want to build: Workforce planning, comp benchmarking, role playbooks, executive briefings, compliance monitoring, AI collaboration, and company-wide AI adoption.
Direct quotes:
- "Anything related to People Ops"
- "How to integrate AI into my company"
What they want to build: Human-out-of-the-loop or human-at-the-ends workflows. Only two asked for dark factories by name, but many are talking about the concept.
Direct quotes:
- "Human out of the loop agentic engineering"
- "Dark Factories"
What they want to build: Internal data pulls, RAG, agent context, research workflows, and turning customer or company knowledge into usable outputs.
Direct quotes:
- "Building context for agents across using information from different sources, searching corpus efficiently - to make agents smarter over time"
- "Meeting to Granola Notes to Use Cases / Agents to Demo Built & Outbound Created based off Industry / Use Cases discussed"
What they want to build: No-code, step-by-step workflows for Slack overwhelm, daily operations, and a clear starting point.
Direct quotes:
- "Step by step do it with me ANYTHING"
- "Getting out of the overwhelm of slack messages"
The quiet thread underneath all of this: how do we know it's right? People are asking about validation, eval loops, and human-at-the-edge architecture. Speakers who address validation explicitly are going to land harder than speakers who don't.
Paste This to Help You Prepare
Drop this into Claude Code, Codex, Gemini CLI, ChatGPT, or your model of choice.