THE FDE PROCESS / FIELD TO CAPABILITY
Forward-deployed engineering brings delivery into real work: understand the judgments, connect methods and tools, evaluate on real tasks, and retain what works.
Download the earlier process reference PDFBEFORE BUILDING / DEFINE THE GOAL
Choose a primary problem in time, errors, waiting or service quality. A goal brief should identify who uses the result, what should improve and the evidence needed to decide.
TEACHING EXAMPLE / WEEKLY REPORT ASSISTANCE
This is a goal-setting example, not a completed client engagement or a promised improvement rate. Resolve missing inputs, measures or ownership, or narrow the scope.
Define who uses the result, how the work is done today and why improving it matters.
Map the goal, resource constraints and baseline. Select a valuable use case with usable inputs.
A business owner explains the goal and provides current examples.
Follow a task through its inputs, judgments, tools and output. Find where expertise makes the difference.
Identify sources, missing information, business definitions, exclusion rules and recurring errors.
A practitioner demonstrates the task and explains key decisions.
Make expert judgment callable, challengeable and testable.
Specify inputs, decision rules, templates, counterexamples and sources; package reusable methods as Skills where useful.
An expert reviews the method and contributes difficult examples.
Choose a fixed workflow or an agent, then define each step’s inputs, permissions, actions and human approval points.
Connect approved data and tools, record key steps and handle missing inputs, failures and duplicate requests.
Provide the agreed data and tool access, and review outputs.
Compare quality, time, rework and cost on the same material. A demo is not a business evaluation.
Cover normal, edge and failure tasks, including held-out examples; have business reviewers check quality, total effort and actual use.
Business users judge usefulness and record actual follow-through.
Retain corrections, feedback and adaptations so the next task starts from a better place.
Hand over capabilities, versions, dependencies, examples and ownership. Separate shared methods from client-specific settings.
Confirm maintenance ownership and the next improvements.
ACCEPTANCE / DECIDE WHAT IS READY
A successful demo is a starting point. System readiness, adoption and business improvement need separate evidence, pass criteria, ownership and responses when they fail.
Check facts, sources and actions on normal, edge and failure tasks, including held-out examples. Test permissions, timeouts, duplicate requests, takeover and rollback.
Evidence: review examples, test and runtime records, recovery drills and operating guidance.
Have users complete agreed tasks. Record adoption, completed actions and rework. After handoff, the team should know how to operate the workflow and where to get help.
Evidence: usage, adoption and rejection records, user feedback and handoff confirmation.
During an agreed observation period, compare time including review, errors, rework and tool costs, then check the agreed outcome measures. Explain concurrent changes in processes, people or markets.
Evidence: baseline and pilot comparisons, user feedback and the owner’s conclusion.
If a gate is missed, narrow the scope, retain human review, retest or stop. Recheck critical tasks when models or data change.
Evaluation references: Anthropic · Agent evaluations · 2026-01-09 · AWS · Agentic AI Lens
WORKING TOGETHER / YOUR TEAM’S PART
These are responsibilities, not mandatory new job titles. One person may cover several roles, but each decision needs an owner. Agree participation and time commitments during scoping.
Set priorities and available resources, resolve scope tradeoffs and decide whether results justify continuing.
Bring: the task, current process, decision makers and constraints.
Demonstrate real work, explain exceptions, review samples and record errors, rework and adoption obstacles.
Bring: typical inputs, good and bad examples, rules and feedback.
Confirm permitted data, interfaces, access and approval points; review technical readiness before release.
Bring: authorization, system limits, integration and release requirements.
Join handoff and recovery practice; own issues, versions and regular checks, with a route to engineering support.
Bring: maintenance arrangements, escalation contacts and improvement priorities.
The FDE connects requirements, implementation and evaluation evidence and makes limits explicit. Agree business decisions, data authorization and release ownership together before work begins.
WHAT FIELDWORK TEACHES US
Matching names may hide different entities. Cumulative data may not be additive. Resolve identity, grain, time and definitions before automating analysis.
Retain: definition cards and test casesHigh potential is not high confidence. Retain evidence, disagreement and unknowns so the business knows what to pursue and what to verify.
Retain: judgment rules and an evidence indexGeneration does not prove sending, adoption or resolution. Include human corrections, waiting, tool costs and actual use.
Retain: comparisons and improvementsTHE COMPOUNDING OUTPUT
Briefs, shortlists, drafts, analysis and options
Skills, templates, counterexamples and test cases
Role Agents, tool configurations and entry points
Sources, dependencies, versions and ownership
Manage shared methods separately from client-specific data, configuration and commercial information. Reuse does not mean taking or publishing client material.
START WITH ONE REAL PROBLEM
You do not need a complete AI strategy. Start with a recurring problem, one existing example and someone who will actually use the result.