← Adopt SpecDD on existing projects guides
How to introduce SpecDD to a skeptical team
This guide shows you how to introduce spec-driven development to a skeptical team without asking them to accept a new process on faith.
Skepticism is useful. It forces the rollout to prove that specs reduce real ambiguity instead of becoming another documentation obligation. The best answer is a small example tied to a real pain point.
Short answer
Start with the team’s concern, not a framework pitch. Explain SpecDD as small source-adjacent contracts that preserve intent, boundaries, tasks, and completion criteria. Show one local example, run one real change, and measure whether it reduced rework, wrong-file edits, review ambiguity, or onboarding friction.
When to use this guide
Use this guide when teammates say:
- “We already have tests.”
- “We already have tickets.”
- “This sounds like more docs.”
- “Agents should infer context from the code.”
- “We do not have time to spec everything.”
- “Architecture docs already exist.”
- “Specs will get stale.”
Steps
1. Start from the team’s concern
Ask what problem would make SpecDD worth trying:
- repeated review comments
- unclear module ownership
- wrong-file edits by agents
- stale architecture docs
- tickets that lose durable behavior after merge
- onboarding questions that repeat
- tests that prove behavior but do not explain intent
Tie the first example to that problem.
2. Explain what SpecDD does not replace
Use simple distinctions:
- tickets still handle planning, priority, and delivery coordination
- tests still provide evidence
- docs still explain broad concepts, users, and rationale
- code still implements behavior
- specs preserve local intent, authority, boundaries, tasks, and done criteria
SpecDD should make existing work more reviewable, not replace every artifact.
3. Show one local example
Use a short spec:
Spec: Itinerary
Owns:
./itinerary.js
./itinerary.test.js
Must:
Reject itinerary items without a place name.
Must not:
Change destination search behavior.
Tasks:
[ ] Add missing-place validation.
Done when:
Missing-place behavior is covered by a check.Point out the concrete benefits:
- reviewers know which files are in scope
- tests map to
Done when - agents know what not to touch
- the behavior remains visible after the ticket closes
4. Address common objections
For “This is more docs”:
The first spec should replace repeated explanation, not add a parallel essay. If it does not help the next change, it is too broad or unnecessary.
For “We already have tests”:
Tests prove behavior. Specs explain intended behavior, ownership, boundaries, and what work is still open.
For “We already have tickets”:
Tickets coordinate delivery. Specs keep durable implementation rules near the code after the ticket closes.
For “Agents should infer from code”:
Inference is exactly where wrong-file edits and accidental behavior happen. Specs make authority explicit.
For “Specs will get stale”:
Behavior-changing pull requests should update specs in the same changeset, and spec changes should be reviewed like code.
For “We do not have time”:
Do not spec everything. Try one active area where rework already costs time.
5. Run a small pilot
Choose:
- one team
- one repository or subproject
- one active module or workflow
- one to three changes
- one review checklist
Use a real task:
Plan the Itinerary validation change.
Then review the plan and diff against the local spec.
6. Use evidence to decide
After the pilot, ask:
- Did the spec prevent wrong-file edits?
- Did review comments become more concrete?
- Did the agent need fewer corrections?
- Did
Done whenclarify testing? - Did the spec stay short enough to maintain?
- Did anyone use the spec during review?
If the answer is no, adjust or stop. If the answer is yes, expand to the next active area.
Common mistakes
- Pitching SpecDD as an agent-only tool.
- Starting with a full-repository documentation plan.
- Arguing abstract benefits instead of showing one local workflow.
- Claiming specs replace tests or tickets.
- Ignoring the team’s real delivery pressure.
- Treating skepticism as resistance instead of useful review input.
How to verify the introduction
The team introduction worked when:
- teammates understand what specs do and do not replace
- the first pilot area is small and real
- objections are answered with concrete workflow changes
- success criteria are written before expansion
- the team agrees on whether to continue based on evidence
Related how-tos
- How to explain SpecDD to your team
- How to run a SpecDD pilot
- How to introduce SpecDD without slowing delivery
- How to decide what belongs in a ticket vs a spec