A writing assistant that drafts ticket replies, KB articles, announcements and email follow-ups, grounded in real customer context, your tone settings and your knowledge base. It auto-attaches to every textarea in your OpsIQ admin and the ticket UI. Your team reviews, tweaks and ships.
Grounded in your dataOne-click tone rewritesHuman reviews & sendsSee the exact prompt
Drafts from the customer record
Warmer · Shorter · More formal
Live
The writing layer
One writer. Every surface.
80%less time drafting replies
4one-click tone rewrites
Livecustomer record as context
Everytextarea, auto-attached
🖊️ The drafting engine
Draft, restyle and ground every reply.
Three things every great reply needs: a first draft, the right tone, and real facts behind it. OpsIQ Writing runs the whole flow before an agent touches the keyboard.
01Reply drafting
Drafts every ticket & email reply.
Inbound email lands in the Email drafts queue with an AI reply already written. In a native ticket the composer offers Draft AI reply. Nothing is sent until an agent presses Send.
Email drafts queue. Inbound mail gets a queued reply you edit, regenerate, then send.
Inbox Copilot in the ticket composer: Draft reply plus one-click rewrites.
Human-in-the-loop. The draft sits in a textarea until an agent reviews and sends.
Email draftsDraft AI replyRegenerate
Live
02Tone control
Pick a tone, and the draft restyles.
The same facts, dressed for the moment. The one-click rewrites are Warmer, Shorter, More formal and Apologetic, and the prompt carries your standing voice and the phrases you never use.
Real composer rewrites: Warmer, Shorter, More formal, Apologetic, not made-up labels.
Standing voice set once in AI Instructions applies to every draft.
Length and formality shift the wording; the underlying facts stay intact.
WarmerShorterMore formalApologetic
Live
03Context-grounded
It writes from the record, not thin air.
Before a word is drafted, OpsIQ pulls the contact, their tier and billing status, the open ticket, prior threads and matching KB entries into the prompt as {variables}, so agents tweak rather than rewrite.
Live variables: {customer_name}, {customer_account}, history and KB content substituted at runtime.
Per-customer prompts: VIP, Growth and free-tier templates swap by record.
Knowledge base aware. Answers cite your own help content instead of inventing policy.
{customer_name}{tier}{kb_match}
Live
🪄 Writes anywhere
A "Write with AI" button on every textarea.
No separate panel, no copy-paste. The assistant attaches itself to the surfaces your team already uses (ticket replies, the KB editor, the announcement composer, outreach templates and contact notes) so drafting happens right where you type.
Auto-attach: appears on ticket replies, KB articles, announcements, outreach and notes.
Inline popover. Write, drop in the draft, edit, send, all in the same field.
Per-role gating. Disable per surface, restrict who can draft with AI.
Live
📄 Beyond replies
One writer for every document.
The same assistant doesn't stop at replies. It drafts knowledge-base articles from an outline, fills the Announcements composer for maintenance, releases and incident notices, and writes Outreach templates for sales, follow-up, onboarding and win-back, all in your standing voice.
KB articlesMaintenanceReleasesWin-backOnboarding
ComposerArticle
Standing voicecarried into every type
01
KB articles
Outline to draft to published help content, voice intact.
02
Announcements
Maintenance, release and incident banners, drafted in the composer.
03
Outreach
Sales, follow-up, onboarding and win-back templates.
You own the prompt
No black box.
A full administrator edits the OpsIQ Writing AI slot in plain language, dropping in {variables} from the side panel. Hit Preview rendered prompt and you see exactly what the AI will receive next turn: variables substituted with live values, registered actions and the obedience frame appended. A dedicated Writing AI slot sits alongside Admin AI, Customer Chat, Inbox Copilot and Triage, and only full administrators can change how the writer behaves.
Dedicated Writing AI slot alongside Admin AI, Customer Chat, Inbox Copilot and Triage.
Live preview shows the real, substituted system prompt: transparency by design.
Owner-only. Only full administrators can change how the writer behaves.
How aggressively it cites your knowledge base, how outreach merge fields behave, and how every generation stays metered, budgeted and permission-gated.
Knowledge reply modes
Your help content flows into the reply.
Matching knowledge-base entries stream straight into the draft as the writer composes, and you decide how aggressively. Knowledge reply mode ranges from smart and balanced to strict and links-only, so the answer either paraphrases your docs, quotes them tightly, or simply points to the right article. The writer also adapts to the surface it is on, loading the right context for a ticket, an email, a KB article or an announcement.
Knowledge reply mode: smart, balanced, strict or links-only, set per workspace.
Mode-aware context loading. Detects ticket, email, KB or announcement and loads the right context.
Per-customer tier templates: VIP, Growth and free-tier prompts swap by record.
smartbalancedstrictlinks-only
Live
Outreach
Rewrite the subject and the body, separately.
Outreach templates carry merge variables with fallbacks like {{first_name|there}}, so a half-filled record never produces an awkward "Hi ,". And because the subject line and the body are independent, you can A/B a punchier subject without touching copy you already love, or restyle the body and keep the winning subject.
Subject and body rewriting. Rewrite each independently, keep what works.
Template variables with fallbacks. {{first_name|there}} never renders a blank.
Sales, follow-up, onboarding, win-back: categorised templates ready to drop into any send.
Live
Governed by design
Metered, budgeted & permission-gated.
Every generation is counted against your plan, so spend never surprises you. A configurable token budget caps the cost of any single draft, and access is controlled by the manage_opsiq_writing permission plus per-role gating, so you decide exactly who can press Write, on which surfaces. Instructions themselves stay owner-only.
Usage metering. Generations counted against plan limits, over-limit blocked.
Token budget enforcement. A configurable max-token cap controls per-draft spend.
Permission-based access: manage_opsiq_writing plus per-role and per-surface gating.
Draft AI reply from the composer, grounded in the thread and customer record.
02
Email reply queue
Inbound email gets a queued reply you edit, regenerate and send.
03
Tone rewrites
Warmer, Shorter, More formal, Apologetic: one click each.
KB articlesOutline to published help content in your brand voice.AnnouncementsMaintenance, release and incident banners from the composer.Outreach copySales, follow-up, onboarding and win-back templates.Context groundingLive {variables} from the contact, billing and history.Per-customer promptsVIP, Growth and free-tier templates swap by record.Live prompt previewSee the exact substituted system prompt before it runs.
🛡️ Grounded, not guessing
Built so it doesn't make things up.
The writer is grounded in your data, gated behind human review, and fully auditable. The risky parts of "AI writing" are designed out.
The risk is designed out.
Grounding, human review and an immutable audit trail aren't add-ons; they're how the writer works by default, on every surface.
🧱
Grounded in your dataDrafts are built from the customer record, ticket thread and knowledge base, not the model's imagination. Less to invent means less to get wrong.
🙋
Human in the loopDefault is review-then-send. The draft waits in a textarea; an agent reads, edits and presses Send. Auto-reply is opt-in with explicit rules.
🚫
Phrases you never useSet guardrails in AI Instructions (promises you can't make, words off-brand) and the writing prompt carries them every turn.
📜
Full audit trailAI History records every writing turn for owners: the context that went in, the draft that came out, and which template was used.
👁
See the exact promptThe live Preview shows the real, substituted system prompt before anything runs. No hidden instructions, no surprises.
🔒
Owner-controlled behaviourOnly full administrators can change how the writer talks. Per-role gating decides who can even press the Write button.
⚖️ Why it's different
OpsIQ Writing vs a generic AI writer.
Standalone tools like ChatGPT or Jasper write well, but they don't know your customer, your ticket, or your knowledge base, and they live in another tab.
Capability
Generic AI writer / Jasper
OpsIQ Writing
Knows the customer & their account
You paste context manually
✓ pulls the live record
Grounded in your knowledge base
General web knowledge
✓ cites your help content
Lives where you work
Separate app / tab
✓ on every textarea
Sends the reply
Copy-paste back out
✓ review & send in place
One-click tone rewrites
Re-prompt each time
✓ Warmer / Shorter / Formal
You see the exact system prompt
Opaque
✓ live preview
Audit trail of every draft
None
✓ owner AI History
Part of one platform (CRM, tickets, KB)
Bolt-on
✓ built in
Full feature list
Everything in OpsIQ Writing.
Every capability, grouped. ★ marks a stand-out.
Feature
What it does
Reply drafting
Email Drafts Queue ★
Inbound mail lands as a queued reply you edit, regenerate and send.
Draft AI Reply (Inbox Copilot) ★
One-click grounded draft straight from the ticket composer.
Tone Rewrites ★
Warmer, Shorter, More formal, Apologetic; facts stay intact.
Regenerate Draft NEW
Re-run the AI for a different variant of the same reply.
Context grounding
Ticket Reply Context ★
Loads the thread, customer, billing and prior tickets live.
Live Variables
{customer_name}, {tier}, {billing_status} and 50+ more substituted at runtime.
Knowledge Base Aware ★
Searches the KB and cites your real help content.
Per-Customer Tier Templates NEW
VIP, Growth and free-tier prompts swap by customer record.
Mode-Aware Context Loading NEW
Detects ticket, email, KB or announcement and loads the right context.
Knowledge Reply Mode NEW
smart, balanced, strict or links-only: control how the KB gets cited.
Integration
Writes Anywhere ★
Auto-attaches to every textarea: tickets, KB, announcements, outreach, notes.
Inline Popover ★
Draft above the textarea, drop it in, edit and send, no copy-paste.
Per-Role Gating NEW
Disable per surface and restrict who can press Write.
Permission-Based Access NEW
The manage_opsiq_writing permission controls who can draft.
Beyond replies
KB Articles (Outline → Draft)
AI drafts full help articles in your brand voice.
Announcements Composer
Drafts maintenance, release and incident copy in seconds.
Outreach Templates
Sales, follow-up, onboarding and win-back templates with variables.
Outreach
Email Subject + Body Rewriting NEW
Rewrite the subject and the body independently.
Outreach Template Variables NEW
{{first_name|fallback}} merge fields never render a blank.
Configuration
Writing AI Slot (Admin-Authored Prompt) ★
Admin edits the Writing AI prompt in plain language with {variables}.
Live Prompt Preview ★
See the exact substituted system prompt before it runs.
Standalone Writing AI Slot NEW
A dedicated slot, separate from Admin, Chat, Copilot and Triage.
Owner-Only Access NEW
Only full administrators can change the instructions.
Safety
Human-in-the-Loop by Default ★
The draft waits in a textarea; the agent reads, edits and sends.
Grounded in Customer Data ★
Built from the record, thread and KB, not the model's imagination.
Phrases You Never Use (Guardrails) NEW
Forbidden words and promises carried into every turn.
AI History Audit Trail ★
Every draft logged for owners: context in, draft out, template used.
Tone & brand voice
Standing Voice NEW
Set tone, formality and signature once; it applies to every reply.
Real Composer Tone Buttons NEW
Actual rewrites, not cosmetic labels.
Length and Formality Shifts NEW
Shorter or longer, formal or casual; the facts stay intact.
Draft in Your Voice on Every Surface NEW
The same voice carries across every place you write.
Administration
Usage Metering (Plan Limits) NEW
Generations counted against your plan; over-limit blocked.
Token Budget Enforcement NEW
A configurable max-token budget controls per-draft AI spend.
FAQ
Writing questions, answered.
What teams ask before they let AI draft on their behalf: control, grounding, brand voice and spend.
Only when you configure auto-reply with explicit rules. The default is human-in-the-loop: AI drafts, the agent reviews, the agent sends. The draft sits in a textarea until someone presses Send.
Yes. Set tone, formality, signature and "phrases we never use" once in AI Instructions. The writing prompt loads those every turn, and the same voice carries across tickets, KB articles, announcements and outreach.
Before a word is drafted it pulls the contact, their tier and billing status, the open ticket, prior threads and matching knowledge-base entries into the prompt as live {variables}, so the reply references what actually happened on the account.
Yes. Hit Preview rendered prompt and you see the real, substituted system prompt before anything runs: variables filled with live values, registered actions and the obedience frame appended. No black box.
On every textarea your team already uses (ticket replies, the KB editor, the announcement composer, outreach templates and contact notes) via an inline Write-with-AI popover. No separate app, no copy-paste, and you can disable it per surface.
Yes. AI History records every writing turn (owner-only) with the original context, the draft that came out, and which template was used.
Access is gated by the manage_opsiq_writing permission plus per-role and per-surface controls. Every generation is metered against your plan and a configurable token budget caps per-draft cost. Instructions themselves stay owner-only.
