UGS AI Support Agent - Proposal
Prepared for: Unified Global Solutions (UGS).
Prepared by: Automate Your Biz (AYB AI).
Date: Nov 3, 2025.
Executive Summary
Custom-Development of UGS AI Support Agent to handle ~40% inbound support calls automatically, delivering 24/7 coverage, and improve first-contact resolution while significantly reducing operational costs.
What we'll do (Phase 1):
Knowledge Base Integration
Ingest ~180 support knowledge base articles (curated and "AI-ready").
Voice AI Agent
Build a voice AI agent for Support (English) with robust troubleshooting flows for software and hardware related issues.
Zoho Desk Integration
Integrate with Zoho Desk for ticket creation/updates and status checks.
Smart Handoffs
Provide controlled live transfers and handoffs to humans based on SLA and issue severity.
Return On Investment
AI support agent capable of handling inbound tech support calls once knowledge base is integrated and tuned.
Reducing UGS operational costs by ~40%.
Faster time-to-answer and reduced manual load for Tier-1 support.
Consistent, auditable triage with data captured in Zoho Desk.
Table of Contents
  • Goals & Current Context
  • Scope of Work (Support Agent only)
  • Solution Design
  • Integrations & Data
  • Conversation Design & Escalations
  • Implementation Plan & Timeline
  • Deliverables
  • Pricing & Terms
  • AI Usage & Billing
  • Support & SLA
  • Assumptions, Risks & Mitigations
  • Next Steps
1) Goals & Current Context
Your goals (from discovery):
Reduce workload and/or replace front-level support where appropriate.
Provide 24/7 coverage (including after-hours handoffs aligned to SLA).
Improve customer experience and consistency of troubleshooting.
Deflect significant call volume to AI while maintaining quality.
Today:
  • Support inquiries ~50/50 software vs hardware.
  • First-contact resolution (FCR) ≈ 40%.
  • Staff coverage: 8am–7pm CST (Support), on-call after hours for Tier-1 (complete outage).
  • English only for now.
  • Target concurrency: usually 2–3, peaks up to ~15 (we'll configure capacity up to 20 concurrent AI calls).
2) Scope of Work (Support Agent Only - Not Billing)
In-scope (Phase 1):
AI voice agent for Support (no billing/payment handling in this phase).
KB ingestion, normalization, and retrieval strategy (RAG).
Zoho Desk: account lookup (via identifiers collected), ticket creation/update, and status retrieval.
Intelligent call triage, guided troubleshooting, and SLA-aware handoff.
Capacity up to 20 concurrent calls per agent without extra platform fees.
3) Solution Design (High Level)
Channels: Inbound phone (SIP/Netsapiens).
AI Core:
  • LLM-powered reasoning with retrieval-augmented generation (RAG) over curated UGS knowledge base.
  • Guardrails: scenario-based prompts, escalation triggers, compliance boundaries.
Key Capabilities:
Issue Triage
Issue triage (software/hardware) → guided troubleshooting steps.
Caller Identification
Identify caller and collect essentials (acct/company/name/callback).
Ticket Management
Create/update Zoho Desk tickets with structured data.
SLA-Aware Handoffs
SLA-aware handoffs: outage vs non-outage flows.
After-Hours Support
After-hours posture aligned to UGS wording.
Observability:
Call transcripts, dispositions, ticket IDs, escalation markers, and reason codes.
4) Integrations & Data
Zoho Desk/CRM
Lookups: contact/domain/phone where feasible.
Ticket ops: create with category, priority, summary, repro steps, device/app context.
Status checks: "open/in progress/resolved" updates to caller.
Netsapiens / Telephony
SIP routing for Support queue numbers.
Transfer to live agent or voicemail as needed.
Knowledge Base (RAG)
Ingest ~180 UGS knowledge base articles → normalize to stepwise, testable actions.
Metadata: product, device, symptom, prerequisites, escalation criteria.
Data Hygiene
PII minimization; only collect what's necessary for ticketing & follow-up.
5) Conversation Design & Escalations
Troubleshooting flow pattern:
01
Confirm context (account/product/device).
02
Symptom classifier (no dial tone, no inbound audio, handset won't register, etc.).
03
Stepwise checks (power, cabling, network, provisioning, handset firmware, softphone settings).
04
Validate outcome after each step; branch intelligently.
05
Resolve → summarize → ticket note + confirmation.
06
Not resolved → create/update ticket and handoff.

After-hours language (per UGS preference): "A human member of our team will call you back as soon as possible, based on our SLA."
Handoff rules (examples):
  • Tier-1 outage keywords or repeated failure on critical steps → live transfer if within staffed hours; else ticket + callback per SLA.
  • Authentication or security red-flags → no troubleshooting; route to human.
6) Implementation Plan & Timeline
Phase 0 — Kickoff & Access (1 week)
  • Confirm success metrics, SLAs, and support queues.
  • Provision API keys (Zoho Desk/CRM), SIP routing for test numbers.
Phase 1 — knowledge base Prep & RAG Setup (1 week)
  • knowledge base audit, normalization templates, ingestion pipeline.
Phase 2 — Build & Integrate (1 week)
  • Voice agent + Zoho Desk integration + disposition schema.
  • Netsapiens call flows; test handoffs.
Phase 3 — User Acceptance Testing & Tuning (1–2 weeks)
  • Internal + UGS pilot: measure containment, FCR lift, false-positive guardrails.
  • Iterate on prompts, KB gaps, and flows.
Phase 4 — Go-Live & Hypercare (2 weeks)
  • Controlled rollout → full Support queue.
  • Weekly review + optimization.
Total Time (typical): ~5–7 weeks from kickoff.
7) Deliverables
AI Support Agent (voice) integrated with Zoho Desk & Netsapiens routing.
KB ingestion & retrieval pipeline with normalization guidelines.
Conversation maps (top issues), escalation matrix, and after-hours posture.
Admin guide (how to update KB, review calls, tweak prompts).
Reporting: containment rate, FCR proxy, escalation reasons, ticket linkage.
Go-live hypercare and monthly optimization cadence.
8) Pricing & Terms
$3,495
One-time
(setup, integrations, ~180 Knowledge Base ingestion pipeline, pilot)
$1,195
Monthly maintenance
/month
Monthly maintenance fee includes support, reporting, monitoring, error handling, minor flow updates, keeping systems up to date, and monthly reviews. Capacity configured up to 20 concurrent calls for the Support agent.
9) AI Usage & Billing
Usage billed at cost (no markup)
All AI voice usage charges will be billed at cost to UGS. Typical range: $0.12–$0.14 per minute for the voice stack. Final per-minute rate: Confirmed after the agent is developed and vendors/configurations are finalized.
10) Support & SLA
AYB-AI support hours: Monday–Friday, 9am–5pm EST via Zoom, Slack, or email.
Response targets (AYB-AI):
P1 (prod-down)
response within 2 business hours.
P2 (degraded)
same business day.
P3 (cosmetic/low)
within 2 business days.
11) Assumptions, Risks & Mitigations
Assumptions
  • UGS provides ~180 KB articles and API credentials (Zoho Desk/CRM, SIP routes).
  • UGS confirms escalation paths and SLA definitions.
  • English-only scope for Phase 1.
Key Risks & Mitigations

Edge cases (devices/firmware variance)
log gaps; prioritize top-frequency issues first.

False confidence from AI
strict guardrails and confidence thresholds; early escalation when uncertain.
12) Next Steps
Approve scope and pricing.
Share access: Zoho Desk/CRM API keys, SIP test route, and initial KB dump.
Schedule kickoff (60 minutes).
UGS AI Support Agent will be custom-developed and delivered in ~5–7 weeks from kickoff.