The Mediator

The problem

Jira was the shared tool. “Ready for work” was not a shared definition.

• Tickets arrived underspecified, missing impact, examples, or acceptance criteria.
• Engineers chased context in comments instead of building systems.
• Leadership had low confidence the backlog reflected reality.

The solution

A structured intake + coaching layer that enforces Definition-of-Ready and makes clarity measurable.

• Normalizes noisy requests into consistent Bug/Story structures.
• Scores clarity with Malcolm’s Chaos Score (1–10) and pairs it with actionable feedback.
• Uses guardrails to prevent unsafe or unintended Jira updates.

Operational context

• Stories without “As a / I want / So that” structure.
• Bugs missing steps to reproduce, URLs, or impact.
• Tasks with unclear ownership or missing metadata.
• Duplicates scattered across projects.

The cost wasn’t just time, it was backlog trust and predictable delivery.

Objectives

• Turn vague requests into tickets that meet a clear Definition of Ready.
• Reduce back-and-forth clarification cycles before work can start.
• Standardize how “chaos” is assessed across the backlog.
• Coach requesters without adding heavy process or a new tool.

Mode A · The Auditor

Review existing tickets against Definition of Ready.

• Detects Bug vs. Story and checks only required fields for that type.
• Assigns Chaos Score (1 = clear, 10 = pure chaos).
• Drafts proposed updates only when explicitly requested.

Mode B · The Architect

Create new tickets from informal descriptions.

• Infers system, type, department, and default epic.
• Runs a short intake interview (max ~5 questions).
• Locks in AC for Stories; steps/examples for Bugs.

Mode C · The Evaluator

Automated triage + internal coaching on new tickets.

• Flags missing AC or weak descriptions for Stories/Tasks.
• Flags missing examples/steps for Bugs.
• Posts growth-minded feedback tied to Chaos Score.

Mode D · The Commenter

Comment-only, safely.

• Requires an explicit ticket key; no Jira search in this mode.
• Defaults to internal comments.
• Never posts Chaos Scores publicly.

Mode E · The Standup

Sprint overview for a given assignee.

• Builds deterministic JQL for assignee + open sprints.
• Returns a standardized list: link, summary, status, points.
• Avoids extra analysis to keep standups fast.

Mode F · The Librarian

Keyword discovery + history search.

• Normalizes phrases into reproducible keyword sets.
• Searches Jira and ranks matches by similarity.
• Returns short summaries + confidence scores.

Standard ticket template

• Metadata: system, type, priority, summary, reporter, assignee, epic.
• Bugs: expected vs actual, impact, steps, URL/example record.
• Stories: As a / I want / So that, notes, acceptance criteria.
• Other types: concise, fact-based technical context.

Hypotheses and “maybe” statements get separated into Additional Notes to keep facts and guesses distinct.

Definition of Ready checklist

• Clear ask: the request is explicit (what you want done/reviewed/changed).
• System context: app/workflow/environment is named (or link provided).
• Concrete identifiers: ticket/record IDs, URLs, examples, or screenshots/logs.
• Expected outcome: what “done” looks like (success criteria).
• Repro & impact (if broken): steps to reproduce + severity/business impact.
• Constraints: deadlines, access limits, dependencies, or known blockers.

If any core inputs are missing, the next step is coaching: ask targeted questions before starting work.

The coaching flow (3 steps)

1. Frame the intent: Restate what you believe the user wants in one sentence.
2. Sharpen the inputs: Ask 2–4 questions that improve precision (system, IDs, success criteria, impact).
3. Build the habit: Offer a reusable “next time” template so future requests are instantly actionable.

This isn’t about catching mistakes, it’s about helping everyone communicate in a way that makes work easier tomorrow.

Question prompts (copy/paste)

• System + scope: “Which system/app is this in, and what part of it are we touching?”
• Identifiers: “Can you share the ticket/record ID, link, or an example record?”
• Outcome: “What does success look like; review, fix, recommendation, or confirmation?”
• Impact: “Who is impacted and how urgent is this?”
• Constraints: “Any deadline, dependencies, or access considerations?”

Tip: If you only add one thing, add system + link/ID it unlocks speed.

Example coaching response (growth mindset)

Use this to move faster now and make the next request easier.

“Next time” request template

• System: Workato / Jira / Salesforce / NetSuite / …
• Link or ID: ticket, record, recipe, job, run ID, etc.
• Goal: what you want (review / troubleshoot / update / build)
• Success criteria: what “done” looks like
• Impact + urgency: who is blocked and by when

The goal is shared clarity with less back-and-forth, faster delivery, and cleaner handoffs.

Mapped standards

• System mapping: strict system strings (e.g., Salesforce, Workato, NetSuite).
• Priority mapping: normalized levels from P0 to P4.
• Department mapping: consistent departments for reporting and routing.
• Default epics: based on system (e.g., NetSuite → ERP Operational Support).

Safety & integrity

• Never updates fields that were not explicitly requested.
• Always drafts proposed changes before any update action.
• Defaults to internal comments; no Chaos Scores in public comments.
• Normalizes names/identity strings for predictable parsing.

Reduced ticket thrash

• Targeted questions once, up front allowing for fewer scattered comment threads.
• More time on design/delivery, less time chasing context.

More truthful backlog

• Consistent readiness standard makes prioritization faster and more reliable.
• Backlog becomes scannable by system, impact, and readiness.

Embedded coaching loop

• Feedback tied to Chaos Score helps requesters improve over time.
• No new tool required; no heavy training program.

Reusable beyond Jira

The pattern Definition of Ready, scoring, and coaching comments; adapts cleanly to other tools and teams.

Intake Mode (consistent data)

• Standardized fields: system, category, IDs/links, goal, success criteria, priority, impact.
• Cleaner handoffs: fewer follow-ups and faster triage.
• Comparable threads: every conversation becomes analyzable and reportable.

This is the turning point: the same inputs, every time, means the data is trustworthy enough to power strategy.

Strategist Mode (what it unlocks)

• Sprint awareness: watch active work against capacity and deadlines.
• Epic-level outcomes: track progress across the whole initiative, not just tickets.
• Proactive alerts: route signals to the right area before risks become blockers.

Once intake is structured, the next phase can focus on prevention: capacity, deadlines, and outcomes at scale.