SRM is the Delaware Dept of Corporations support triage workflow. Paste an inbound email, SRM auto-creates the ICISTWO JIRA, asks the AI for the 6-field issue summary, assigns it to one of 55 developers, drafts the request-update + follow-up + verification messages, and logs every step. Five steps, one screen, full audit trail. Or skip the wizard entirely with One Shot — paste the email and get 8 AI-generated outputs in parallel, then create or update the JIRA with a single click.
Every support request moves through the same six stages — from JIRA ticket creation all the way through team reporting. Each stage is atomic, logged, and resumable from the history tab.
One screen, three columns: paste the email on the left, watch the AI fill in the 6-field summary in the middle, see assignments and verifications track on the right. A 5-step stepper at the top walks you through Create JIRA → Input → Analyze → Assign → Verify so you never lose your place. Stat tiles show analyzed today, assigned today, verified today, and currently open.
One Shot is a speed lane alongside the 5-step wizard. Paste one email, hit Process, and six parallel AI threads generate eight outputs simultaneously — JIRA summary, email summary, TL;DR, executive comment, assignment message, request update, follow-up, and verification draft. Each field shows a source tag (email-extracted, form override, system default, or user-edited) so you know exactly where every value came from. When you're ready, Create JIRA builds the ticket in one call; Update JIRA pushes all fields at once; Update JIRA Partial lets you selectively toggle which fields to push.
The Analytics tab pulls live counts straight from JIRA via JQL and renders two charts: a doughnut of Support Request tickets by status, and a horizontal bar of tickets per developer (top 15). Date range pills toggle Today / This Week / This Month / This Year so you can see what landed in the last 8 hours or what closed all year. The exact JQL is shown in the header for full transparency — no black box.
Every analyzed request lives in a filterable history with status badges, priority labels, assigned developer, JIRA number, and creation date. Filter by status, developer, or free-text search across the email content, summary, JIRA number, or issue title. CSV and JSON exports for compliance reporting; ShipLog entries for cross-app tracking.
JIRA tickets created outside the SRM workflow often miss the Delaware 6-field format. The Review Board fetches all Support Request tickets via JQL and lets you run the AI rewrite pipeline across hundreds of tickets in one session. Auto-Run mode with smart skip zips instantly past tickets with no MSG attachment or already-processed descriptions, firing the configured delay only after a successful rewrite. A live banner tracks progress (done / total — ok, skipped) and an error log captures failures without stopping the run.
Two full-featured performance reports built on live JIRA data. The Developer Report profiles a single assignee — ticket volume, resolution speed, priority distribution, status breakdown, and recent activity — with HTML and PDF export. The Team Report covers the whole roster across any custom date range: workload heatmap, period-over-period comparison, top performers, and per-developer drill-downs with direct cross-links to the Developer Report. The Head-to-Head report duels any two developers across 7 weighted metrics scored out of 10, rendered with radar and bar charts and a factual narrative summary. Team vs Team extends that to multi-developer groups — pick saved digest teams or build ad-hoc squads and compare aggregated performance side-by-side with a per-member breakdown.
The Developer Profiles feature gives a bird's-eye view of every developer's SR performance in a single screen — no report to generate, no date range to configure. Profiles are materialized from live JIRA data and cached: page loads are instant, zero Jira calls. A Top Performer spotlight band anchors the grid, followed by a dense ranked table (YTD closed, lifetime, open queue, resolution rate with inline bar, median TTR). Clicking any row opens a full dossier: an 8-cell stat grid with team-median deltas, an AI professional summary (3 grounded paragraphs from MiniMax, regenerated only when underlying stats change), a Milestones rail, a hand-rolled SVG monthly throughput chart, and a Recent SRs table. The roster is self-maintaining — new assignees appear automatically on the next Refresh All. Token-efficient: unchanged developers cost zero LLM tokens on refresh.
Every ticket gets an SLA deadline the moment it's created, computed from its priority level: P0 = 4 hours, P1 = 8 hours, P2 = 24 hours, P3 = 72 hours. The dashboard SLA panel shows hours remaining for every open request. Breached tickets appear in red so triage is instant — no spreadsheet cross-referencing required.
When a support request needs to escalate beyond JIRA, the 🎫 Escalate to SNT button on
any ticket detail inserts a record directly into snt_ticket_requests. The cross-app
link appears instantly — no context switching, no manual data entry.
The Word Bubble scans all 5,000+ ICISTWO Support Requests, tokenises every description, and builds a live word-frequency cloud. Spot recurring themes — which errors, modules, or user complaints dominate the queue — at a glance. Incremental rescanning skips unchanged JIRA entries so subsequent runs complete in seconds. Click any word to drill into the JIRA tickets that contain it.
Delaware's developers don't write down how they fixed a ticket — so the Knowledge Base doesn't pretend to be a solution base. Instead it turns 5,000+ ICISTWO Support Requests into pattern intelligence and prediction: it tells you what problems keep coming back, who they hit, how often, the priority mix, and how long they take to resolve. The Recurring Problem Dashboard groups semantically-similar tickets into stable problem types, each with a 3-month trend, a predicted next-month volume, the typical owner, the priority mix, the median time-to-resolve (TTR), and a velocity rank that floats the fastest-accelerating problems to the top. Clusters keep the same identity run-to-run, so a trend means something.
Click any problem type to open its detail page: an AI-written 3-6 word label and one-line description (generated locally by qwen3, regenerated only when membership changes), a hand-rolled SVG monthly-throughput chart, the priority mix, the typical owner, the median TTR, and the full member-ticket table. The label and summary are grounded strictly on the member tickets — no invented solutions.
Drop in a Jira key (e.g. ICISTWO-35918) or paste free text describing a new issue, and
the Similar Ticket Finder runs a vector similarity search across every embedded Support Request,
returning the closest past tickets ranked by cosine similarity. Weak matches below 55% are hidden,
and the ticket itself is excluded. Above the results sits a context card that summarises the
matched set into a "what to expect" snapshot — how many times this has occurred, the median
time-to-resolve, the typical owner, the priority mix, and when it was last seen. It tells you what
to brace for, not the steps to fix it.
The Manager Digest generates a structured weekly team report from live JIRA data and emails it directly to the manager via AgentMail. Configure the recipient, optional CC addresses, custom date range, and which sections to include — team summary, per-developer cards, aging open tickets, or AI highlights. Team groups let you scope the digest to a subset of developers (e.g. state-side vs. EPAM). Schedule it to fire automatically every Monday at 9am, or generate and send on demand. A full history log tracks every sent digest with period, developer count, ticket totals, and delivery status.
Two live workload views built on real-time JIRA JQL. The Team Workload Report shows open ticket counts per developer for a selected date range, surfacing imbalances before they become bottlenecks. The Developer Workload Report drills into a single developer — all open tickets, priority breakdown, and recent activity. Both export to CSV, HTML, or PDF and can be emailed directly from the report view.
Pick any two developers, set a date range, and SRM computes a 7-metric weighted duel scored out of 10 points. Metrics include tickets resolved, SLA compliance, average resolution time, weighted workload, velocity, time to first action, and reopen rate. A radar chart and bar chart render the performance shape visually. A factual narrative summary names the winner and cites the two largest margin advantages — no AI required.
Extend Head-to-Head to multi-developer teams. Pick a saved team from the digest-teams roster or build an ad-hoc group by selecting any developers on the fly. Both sides can mix modes in the same comparison. Issues are aggregated across all team members and scored on the same 7 weighted metrics. A per-developer breakdown shows each member's individual resolved and total ticket counts.
Eleven integrated capabilities covering the full support request lifecycle — from intake and AI analysis through workload management, weekly manager digests, pattern intelligence, and cross-system escalation.
snt_ticket_requests — no context switching required.15 read-only analytics screens built on your ticket data
The Insights Hub delivers 15 pre-built dashboards that surface operational intelligence, strategic metrics, and pattern analysis from your Jira ticket history — no configuration required. Open any screen and the data is already there, drawn live from ICISTWO and organized into five categories so you always know where to look.
SRM bundles sixteen capabilities that used to be separate tabs, emails, and spreadsheets. Every action is atomic, every action is logged, and every step is rewindable from the history tab.
snt_ticket_requests, then the SNT ticket id is linked back onto the SRM request for a bidirectional cross-app link — without leaving the SRM workflow.error_code and context. Developer assignments also fire a real-time Telegram notification so the assignee knows the moment a ticket lands.Each support email goes through the same five-step pipeline. Skip steps you don't need. Reload from history any time. Every action is logged.
Paste the email, click Create New JIRA. SRM extracts the Subject line, opens an ICISTWO ticket via the Atlassian REST API, and shows you the key + browse link.
The pasted email body becomes the initial JIRA description. The form is now ready to analyze.
OpenRouter (Claude Haiku 4.5) reads the email and returns the 6-field Delaware-format output: Priority, Issue Title, TLDR, Stakeholders, JIRA, Developer. The result is pushed back into the JIRA description.
Pick a developer from the 55-name roster (or filter to the 9 state-side direct assignees). SRM drafts the assignment message in Delaware's exact format.
Paste the developer's resolution email back into SRM. The AI generates a "please verify" message addressed to the original requester. Status flips to verified, ShipLog logs it, optional Telegram alert fires.
Every Analyze call produces the same six fields, in the same order, in plain text ready to paste
into ServiceNow or JIRA. Stakeholder rules respect dosdoc_supportrequests and
dosdoc_tech_support exclusions automatically.
SRM does not send mail on your behalf — it drafts the four messages Delaware expects in their exact format so you can review, edit, and send from Outlook. Every draft is keyed to the active JIRA and logged in the request history.
One-liner for the developer: "Assigned ICISTWO-##### to {Dev}, {TLDR}." Ready to paste into the assignment channel.
Polite nudge to the assigned developer asking for a status check. Used from the dashboard between Assign and Verify.
Second-touch reminder when a ticket has been sitting idle. Same JIRA context, different tone.
Addressed to the original requester, written in the requester's language. Confirms the fix and asks them to mark the ticket verified.
SRM ships pre-seeded with the full Delaware Dept of Corporations IT roster: 55 developers total, 9 of whom are state-side (direct assignments). The remaining 46 are EPAM contract team and route through Srinivas Topoji. A toggle on the dashboard filters the dropdown to state-side only when you're routing direct.
SRM does not try to be a system of record — it glues the ones you already have together. Every outbound call is rate-limited, credential-scoped, and logged.
Every support email through one screen. JIRA, AI summary, assignment, verification, audit log — all in five clicks.
Launch Support Request ManagerRequires a MelTuc account. Create one free.