Forward-Deployed Engineers for AI Agents

Scope, run and govern an engagement where vendor engineers take an AI agent from one workflow to production, and keep the result

A enterprise course from ibl.ai (self-paced, 6 to 8 hours, free). A self-paced workshop for enterprise sponsors of AI agent deployments: what forward-deployed engineers do, how to choose and map a first workflow, connect systems with least privilege, set an evaluation bar before production, govern the live agent with the NIST AI RMF, and write an engagement charter that leaves your teams owning the agent.

Last updated:

Course description

Forward-Deployed Engineers for AI Agents teaches CIOs, CTOs, heads of platform and operations leaders how to scope, run and govern an engagement in which a vendor's engineers work inside their teams to take one AI agent from a scoped workflow to production. You will compare forward-deployed engineering with consulting and SaaS onboarding, define the roles on both sides, choose a first workflow with a named business owner and a success metric, map it with the people who run it, connect systems with least-privilege access using OWASP's Excessive Agency guidance, NIST CSF 2.0 and the Model Context Protocol's security rules, build an evaluation set and pass bar with adversarial cases from the OWASP Top 10 for LLM Applications 2025, govern the live agent with the NIST AI RMF and its Generative AI Profile, and plan a handover that leaves your teams owning the agent while the partnership continues. Six labs build one engagement charter, which the capstone assesses. Vendor-neutral; last verified 18 November 2026.

About this course

Forward-deployed engineers are vendor engineers who work inside your teams to take an AI agent from one scoped workflow to production: mapping the process, connecting systems, building evaluations and handing over operations. This course teaches the people who sponsor and run that work how to scope it, govern it and keep the result. It is vendor-neutral and built from published frameworks: the NIST AI Risk Management Framework and its Generative AI Profile (NIST AI 600-1), the OWASP Top 10 for LLM Applications 2025, the Model Context Protocol specification and security best practices, and the NIST Cybersecurity Framework 2.0. Where it shows one vendor's published approach, it uses ibl.ai's pages and says so. Across six labs you build one engagement charter for a first production agent on a real workflow.

What you will be able to do

  1. Distinguish forward-deployed engineering from consulting and SaaS onboarding, and define the roles, time and access on both sides of an engagement.
  2. Select a first workflow that can be measured and shipped, with a named business owner and a success metric, by scoring candidates on volume, measurability, data access, owner commitment and risk.
  3. Map a process with the people who run it into steps, decisions, exceptions and the approvals that must stay with people.
  4. Connect an agent to enterprise systems with least-privilege access, choosing integration routes and applying OWASP's Excessive Agency guidance, NIST CSF 2.0 and the Model Context Protocol's security rules.
  5. Set an evaluation set built from real cases and a pass bar the agent must meet before production, including adversarial tests from the OWASP Top 10 for LLM Applications 2025, and govern the live agent with the NIST AI RMF.
  6. Write an engagement charter and handover plan that leave your teams owning and operating the agent while the partnership continues with updates and the next agents.

Who this course is for

Roles: CIOs, CTOs, heads of platform and of AI, automation or digital transformation, and the operations leaders who sponsor an AI agent deployment and work beside a vendor's forward-deployed engineers; business process owners, enterprise architects and procurement or vendor management staff.

Organizations: enterprises in any sector that are taking a first AI agent, or their next one, from demo to production with a vendor's engineers.

Before you start: have a candidate workflow in mind with a business owner who could be named. No programming is needed. Recommended first: Agentic AI for the Enterprise (ENT-1).

Format

Time: 6 to 8 hours, self-paced. Level: intermediate. Hands-on: six labs that each produce a part of your engagement charter (workflow selection, process map, system access register, evaluation set and pass bar, risk register and monitoring plan, handover checklist), with an AI tutor that coaches each part from the drafts you paste.

Structure: a Start here section, four content sections of readings, worked examples, templates and labs, each ending in an AI assessment, and a capstone in which you assemble and defend the engagement charter, followed by a reading list.

Grading: the four section AI assessments count for 60 percent of the grade and the capstone assessment for 40 percent. The pass mark is 70 percent.

Currency

Last verified 18 November 2026 against NIST AI 100-1, NIST AI 600-1, the OWASP Top 10 for LLM Applications 2025, the Model Context Protocol specification version 2026-07-28, NIST CSF 2.0, NIST SP 800-61 Revision 3, and ibl.ai's AI Transformation and Process pages.

These frameworks and vendor offerings are still changing: NIST, OWASP and the Model Context Protocol project revise their documents, and vendors change how they run engagements. Check the current sources before relying on any page or template. This course is training, not legal or procurement advice; your counsel, procurement and security teams decide what your contracts and controls must say.

Outline

  1. Section 0 · Start here

    • 0.1 Welcome and orientation
  2. Section 1 · Decide what the engagement is for and where it starts

    • 1.1 What forward-deployed engineers do
    • 1.2 Choosing the first workflow
    • 1.3 Section 1 assessment
  3. Section 2 · Map the work and connect the systems

    • 2.1 Mapping a process an agent can run
    • 2.2 Connecting the agent to your systems
    • 2.3 Section 2 assessment
  4. Section 3 · Set the ship bar and govern the live agent

    • 3.1 Proving the agent is good enough to ship
    • 3.2 Governing the agent once it is live
    • 3.3 Section 3 assessment
  5. Section 4 · Hand over and write the engagement charter

    • 4.1 Handing over without creating a dependency
    • 4.2 Writing the engagement charter
    • 4.3 Section 4 assessment
  6. Section 5 · Capstone and wrap-up

    • 5.1 Capstone: Engagement charter for a first production agent
    • 5.2 Capstone assessment
    • 5.3 Reading list and next steps