Keynote / Engineering & Delivery

Agentic Coding: From Prompt to Software Factory.

Generating code is one step. Turning it into reliably delivered software is a system challenge. The talk combines lessons from practice with a clear look at task breakdown, context, tests and reviews, and shows what a software factory means beyond the buzzword.

45–60 minGerman / EnglishFee from €6,500 plus VAT and travel
From brief to software: a harness connects implementation, tests and review. Keynote: context and perspectives.

Who it’s for

Who this keynote is for.

For engineering leads, CTOs, product leaders and development teams who want to understand agentic coding as a way of working.

What your audience takes away.

  • A clear target picture of a repeatable, AI-assisted development process.
  • An understanding of the harness: the environment of tools, rules and feedback in which an agent works.
  • Criteria for which tasks can be delegated and where human decisions remain necessary.

Content framework

Three perspectives on the topic.

We tailor the focus and technical depth to your question.

01

Why a single prompt is not enough

Treat the brief, context and acceptance criteria as prerequisites for verifiable development work.

02

Design the harness

Combine tool access, tests, feedback and boundaries into a controllable working environment.

03

From code to shippable software

Lessons from practice on reviews, integration and failures, without handing responsibility over to automation.

How I work

Delivery needs tests, boundaries and approvals.

The explorer shows the overarching architectural framework, not a live coding environment. In the bootcamp, we apply it to tasks, harness, tests and review.

Enterprise AI · architecture principleSelect a layer
01 · Strategy02 · Architecture03 · Delivery04 · Operations

03 · Orchestration

Individual capabilities become a controlled workflow.

Models, tools and people take on defined tasks. LangGraph lets you model stateful workflows, including handoffs, approvals and clear limits on autonomous actions.

  • Workflows · LangGraph
  • Tool integration
  • Human-in-the-loop

ResultA traceable workflow from input to result.

Related: AI agents workshop →

Schematic architecture example. The specific choice of technologies and building blocks follows the use case.

Before you inquire

Frequently asked questions.

How long is a keynote?

The keynote format runs 45–60 minutes. Alternatives within the offering are a 20-minute talk, a 90-minute deep dive with Q&A, or a panel by arrangement.

What does the keynote cost?

Fee from €6,500 plus VAT and travel. International €8,500–12,000 plus VAT. All fees plus VAT.

In which languages can you give the talk?

I speak fluent English. All offerings and intro calls are available in German and English.

Can the keynote be delivered remotely?

Yes. Keynotes are available on-site and remotely; just mention your preferred setting in your inquiry.

How technical is the talk?

The focus depends on the audience. Leadership teams get a strategic perspective; technical teams can go deeper into retrieval, graphs and agentic workflows.

Do we need Neo4j or LangGraph in-house?

No. The talks explain architectural principles. Neo4j and LangGraph serve as concrete examples; you do not need to use these tools already.

What do you need for an inquiry?

Helpful details are your preferred date, venue or setting, audience and size, topic, and the language you want. Availability is checked case by case.

Portrait of Hendrik Henze.

Hendrik Henze

Enterprise AI · strategy, architecture and delivery.

More about how I work

Your next step

Agree on date and focus.

We align technical depth and examples with your audience. The focus is on transferable engineering principles, not on the promise of fully autonomous development.

Tell me about your audience or team, your question and a preferred date. On that basis, we clarify format and scope.

See the workshop on this topic
← All keynotes