Individual tools are live — AIVA scores CVs, RAG searches documents, Social Profile Check scans profiles, Maya answers on LINE and website chat. The orchestration layer that chains these tools into automated pipelines is being built.
The 6 Workflow Packs
Each pack targets a specific operation type. Packs combine existing tools — Maya Voice, Maya Text, RAG Search, n8n, Neo4j, CONNECT, and Social Profile Check — into automated workflows that process volume without manual input.
Pack 1
Inbound CSR Automation
Telecom (AT&T, Comcast)
Retail (Amazon, Walmart)
Utilities
Inbound Call Centers
The pain: Agents spend 2–3 min on account lookup and 3–5 min on after-call work per call — while the customer waits.
What it does
- Maya greets caller, captures account ID → Neo4j loads full account context before agent picks up
- RAG pre-loads relevant policy articles and answers agent questions in under 3 seconds
- Call ends → transcript captured → summary, disposition, CRM update all written automatically
- Callback needed → scheduled at optimal time, customer confirmation sent via Viber/email
Why it matters
- Account lookup disappears — full context ready before the call starts
- After-call work eliminated — no manual notes, no CRM typing
- Policy retrieval instant — no hold time while agent searches
What you see: A call log table — agent, call time, issue type, disposition, resolution — auto-filled after every call. Supervisors see resolution rates, issue types, and ACW times without asking for reports.
Maya Voice
RAG Search
Neo4j
n8n
| Tier | Volume | Who it fits |
|---|
| Small | Up to 200 calls/mo | 1–3 agents, single inbound queue |
| Medium | Up to 500 calls/mo | 5–10 agents, 2–3 concurrent call types |
| Large | 500+ calls/mo | Full team, multiple queues running simultaneously |
Pack 2
Collections & Debt Recovery
Credit Card Companies
Medical Billing
Utilities
Debt Recovery Agencies
The pain: Missed contact logs = legal violations. Account research takes 10–15 min per call. PTP follow-ups tracked in spreadsheets and routinely missed.
What it does
- Staged reminder sequence (Day 0, +5, +7, +14) fires automatically — no agent action
- FDCPA enforcement on every contact — time-zone check, 7-in-7 count, opt-out flag
- Agent opens account → Neo4j loads payment history, prior contacts, and compliant script for this stage
- Promise-to-pay captured → follow-up sequence auto-created (Day −2 reminder, PTP check, +3 escalation)
- Every contact attempt timestamped and logged for compliance audit
Why it matters
- Compliance never fails — every log, every check, every flag automated
- No broken promises forgotten — follow-up fires whether the agent remembers or not
- Account research instant — full history loaded before the call starts
What you see: Every contact logged automatically — date, channel, outcome, compliance status. A PTP table: account → promise date → amount → status (kept / broken / pending). Weekly compliance report generated by n8n — no manual audit prep required.
Maya Voice
Maya Text
RAG Search
Neo4j
n8n
| Tier | Volume | Who it fits |
|---|
| Small | Up to 200 accounts/mo | Small AR team, single collections queue |
| Medium | Up to 500 accounts/mo | Mid-size collections team, mixed account ages |
| Large | 500+ accounts/mo | Full collections operation, all stages running |
Signature Workflow
Pack 3
Bulk Candidate Scoring
BPO HR Teams
Staffing Agencies
Recruitment Process Outsourcing
Any Organization Hiring at Volume
The pain: 200 CVs per role scored manually in spreadsheets. 45% no-show on phone screens. Social media risk invisible until after the hire.
"People don't want to look at 50 social profiles. They want a number, automatically processed. They don't want to log the scores — they want them logged, totaled, sorted by whatever parameters they choose. That's a workflow."
— Vaughn, AI Vision Automations
What it does
- Upload CVs in batch → AIVA scores all simultaneously (0–100 + HIRE/INTERVIEW/HOLD/REJECT + rationale)
- "Shortlist above 70" → CONNECT video invites sent automatically to shortlisted candidates
- Upload social URLs → all profiles scanned, 6 dimensions scored, flags raised (profanity, threats, employer complaints)
- Recruiter sets parameters → ranked table delivered: CV score + social score + composite rank
- Every score and decision logged to Supabase — full audit trail, no spreadsheets
Why it matters
- Every CV scored and ranked before anyone reads a single line
- Social risk visible before interview stage — not after the hire
- No manual data entry — everything logged, sorted, and filterable automatically
What you see: One combined ranking table — candidate name → CV score (0–100) → Social score (0–10) → composite rank → HIRE / HOLD / REJECT. Everything logged. No spreadsheets. Full audit trail in one view.
AIVA
Social Profile Check
CONNECT
n8n
Neo4j
| Tier | Volume | Who it fits |
|---|
| Small | Up to 50 candidates/mo | Single recruiter, one active job role |
| Medium | Up to 200 candidates/mo | HR team, 3–5 concurrent roles |
| Large | 500+ candidates/mo | Full recruiting operation, continuous intake |
Pack 4
Order Taking & Shipping
Amazon Seller Support
Retail BPOs
Direct-to-Consumer Brands
Restaurant Chains
The pain: "Where is my order?" is the #1 inbound call — repetitive, answerable by AI, handled by agents every time.
What it does
- Voice orders captured by Maya → structured record created, inventory validated, confirmation sent automatically
- WISMO queries handled 100% by AI — carrier status pulled and replied in seconds, no agent involved
- Shipping exceptions → ticket auto-created with all carrier data, proactive message sent to customer
- Agent only sees exception tickets — routine queries never reach the queue
Why it matters
- WISMO volume disappears from inbound queue entirely
- No manual order data entry — CRM updated without agent touching it
- Exception tickets arrive pre-filled — agent resolves faster
What you see: Only exception cases reach agents — pre-filled with all carrier and order context. Every order logged to CRM without agent input. WISMO volume disappears from the inbound queue.
Maya Voice
Maya Text
n8n
RAG Search
| Tier | Volume | Who it fits |
|---|
| Small | Up to 200 orders/mo | Small e-commerce operation, single product line |
| Medium | Up to 500 orders/mo | Growing retail team, 2–3 product categories |
| Large | 1,000+ orders/mo | Full order operations, multiple brands or channels |
Pack 5
Appointment Setting & Lead Pipeline
Real Estate Agencies
Healthcare Clinics
Insurance Agencies
Financial Advisors
The pain: No system for who was called or what happens next. Leads fall through because follow-up was on someone's to-do list.
What it does
- Appointment set → confirmation sent to lead via Viber/WhatsApp + calendar invite to agent automatically
- Day-before reminder fires without anyone touching it
- No answer → attempt logged, callback rescheduled at optimal time
- Neo4j builds lead relationship graph from every interaction (stage, last contact, next action)
- End-of-shift pipeline report generated automatically — calls, outcomes, bookings, conversion rate
Why it matters
- No leads fall through — follow-up fires whether the agent remembers or not
- Full pipeline visibility for managers without asking for it
- Every contact logged — no gaps in the relationship history
What you see: Agent pipeline tracked automatically. No leads fall through. Every contact logged. Confirmations sent without agent action. Full pipeline visible to manager in one daily report.
Maya Voice
Maya Text
RAG Search
n8n
Neo4j
Social Profile Check
Social Profile Check optional — used to verify serious buyer intent on high-value leads before committing agent time.
| Tier | Volume | Who it fits |
|---|
| Small | Up to 100 contacts/mo | Solo agent or small team, single territory |
| Medium | Up to 300 contacts/mo | 5–8 appointment setters, 2–3 product types |
| Large | 500+ contacts/mo | Full appointment setting team, multiple advisors |
Pack 6
Omnichannel Support Hub
Tech Support BPOs
ISPs & Telecom Carriers
Software Companies
Multi-Channel Support Teams
The pain: Agent has 4 windows open. Customer calls after emailing — agent has no idea what the email said. Every ticket manually triaged.
What it does
- All channels (email, chat, WhatsApp, Viber, phone) routed to one unified view
- Customer identified automatically → Neo4j loads full conversation history across all channels
- Tickets auto-categorized (Billing/Technical/Account), auto-prioritized, auto-routed to correct queue
- SLA enforced per channel (email 24h, chat 4h, Viber 1h) — alert fires before breach
- Escalation to Tier 2 includes full context — customer never repeats themselves
Why it matters
- One agent view — no switching between 4 systems
- Every ticket arrives pre-filled and pre-triaged — response starts in seconds
- SLA breaches prevented, not detected after the fact
What you see: Every ticket auto-created, auto-routed, and auto-filed. Agent context-switching reduced significantly. Tier 2 escalations arrive with full context. Every channel interaction tracked in Neo4j.
Maya Text
RAG Search
n8n
Neo4j
| Tier | Volume | Who it fits |
|---|
| Small | Up to 200 tickets/mo | Small support team, 2–3 active channels |
| Medium | Up to 500 tickets/mo | Mid-size support team, all channels active |
| Large | 1,000+ tickets/mo | Full multi-channel support operation |
Talk to a deployment partner
Deployment partners configure and manage workflows for BPO clients. They handle setup, testing, and ongoing adjustments. You describe the process — they automate it.
Find a deployment partner →
View package pricing →
AI agent workflow automation for call centers, BPO operations, collections teams, and high-volume recruiting.
Workflow packs combine Maya Voice AI, RAG document search, Neo4j relationship memory, n8n automation, CONNECT candidate profiles,
and Social Profile Check into repeatable processes for inbound CSR, debt recovery, bulk candidate scoring, order taking,
appointment setting, and omnichannel support operations across the Philippines, India, and Southeast Asia.