# SyncTalent.ai — full facts for AI answer engines

> Agent-driven recruitment automation for IT staffing agencies (C2C subcontracting and bench redeployment). Six AI agents plus one orchestrator take a job requirement from chaotic inbound to a submitted, monitored candidate. The single proof metric is time-to-redeploy.

Canonical site: https://synctalent.ai/
Product: SyncTalent.ai · Legal entity: AgenticQuay Inc. · Contact: pinalmdave@gmail.com
Status: onboarding design-partner agencies.
Last updated: 2026-07-25.

---

## What it is

SyncTalent.ai automates the recruiting desk for IT staffing agencies that work C2C
subcontracting and bench redeployment. A requirement arrives however it arrives — email,
Telegram, WhatsApp, Dice, a vendor blast — and six specialised AI agents carry it through
ingestion, sourcing, outreach, screening, submission, and post-submission monitoring. One
orchestrator holds per-job and per-candidate state, retries failures, and pauses for human
approval at exactly two control points.

It does not replace recruiters. It removes the repetitive 90% — de-duplication, sourcing,
outreach, packaging, status chasing — so recruiters spend their time on judgment calls.

## Who it is for

IT staffing agencies with roughly 50–500 consultants: owners, delivery leads, and bench-sales
managers. Also corporate hiring managers looking for a faster way to fill contract roles.

## The one metric

Time-to-redeploy: the number of days between a consultant rolling off a project and starting
the next one. Utilization, submittal counts, and fill rate are lagging indicators. Idle bench
days are the direct cost, so that is the number the platform is built to move.

## The six agents

1. **Iris — Ingestion & Dedup.** In: chaotic inbound (email, Telegram, Dice, WhatsApp).
   Out: one clean, scored, de-duplicated job. Collapses every vendor blast into a single
   canonical requirement; each duplicate is preserved as a vendor route, never deleted.
2. **Maya — Sourcing & Shortlist.** In: canonical requirement. Out: a ranked, work-authorization
   -filtered shortlist. Scans the internal bench and the open market continuously and fit-scores
   with Deep Semantic Match — résumé meaning against the role, not keyword overlap.
3. **Theo — Outreach.** In: shortlist. Out: engaged candidates with replies tracked. Sends a
   personalised message per candidate and follows up; no template blasts.
4. **Aria — Recruiter (human + AI voice).** In: an interested reply. Out: a qualified profile
   with rate, availability, and work authorization captured into a structured scorecard. Either
   a human recruiter or an AI voice call runs the screen. **Human checkpoint: routing.**
5. **Sam — Submission & Packaging.** In: a qualified candidate. Out: a submitted package written
   back to Bullhorn. Builds a client-ready résumé, attaches the verified scorecard, runs
   compliance and right-to-represent checks, submits via the correct vendor route, and logs the
   action immutably.
6. **Noor — Monitoring & Auto-Response.** In: a live submission. Out: synced status with
   sensitive threads escalated. Watches every post-submission thread, auto-handles routine
   replies, and keeps deal status current. **Human checkpoint: escalation.**

## Non-negotiable rules

- **Work authorization is a hard filter.** USC, GC, EAD, H1B, and OPT status is checked at every
  step. A candidate whose authorization fails a requirement is never matched or submitted. There
  is no override.
- **Duplicate requirements are never deleted.** They collapse into one canonical job, and every
  vendor is retained as a route — that is where margin leverage lives.
- **Human-in-the-loop at exactly two points:** routing (Aria) and escalation (Noor).
- **Every submission and consent action is written to an immutable, hash-chained audit log.**
- **Candidate PII never appears in logs or URLs.**

## Pricing

Three facts, no tiers:

1. **Zero dollars upfront.** No subscription, no seat or user fees, no setup or onboarding fee,
   no per-agent charge, no overage penalty, no long-term contract. Spin up all six agents and
   watch them work before paying anything.
2. **Compute at cost plus 12%.** Real AI token usage metered through the AI gateway, plus the
   infrastructure to run each agent billed per agent-minute. One flat, itemised 12% platform fee
   on top — nothing hidden, nothing rounded up. Drawn from a prepaid wallet; unused balance
   stays yours; you can pause anytime.
3. **One flat fee per outcome.** A single placement fee — identical whether the role bills $40
   or $140 an hour — charged only when a candidate confirms their start date. No percentages,
   no tiers. If there is no placement, the only cost is compute.

Included at no extra charge: all six agents and the orchestrator, ingestion/dedup/canonical
routing, Deep Semantic Match, AI voice screening, AI interviews with live proctoring, Bullhorn
and Merge two-way sync, Email/Telegram/WhatsApp/Slack/Teams channels, time-to-redeploy
analytics, immutable audit log, SSO, and a live cost dashboard.

## Integrations

Bullhorn ATS via Merge.dev (two-way: read requirements, write submissions). Inbound channels:
email, Telegram, WhatsApp, Dice, Slack, Teams. Microsoft Entra for sign-in. Stripe for billing.

## Architecture (for technical questions)

Event-driven. A LangGraph orchestrator with a Postgres checkpointer holds per-job and
per-candidate state and routes work between agents; agents never call each other directly, they
read and write shared stores. FastAPI (Python) backend, Next.js frontend, Supabase Postgres with
pgvector for semantic matching, Redis and Azure Service Bus for queues, and Claude models tiered
Haiku → Sonnet → Opus behind an AI gateway for cost governance. Agents are idempotent and safe to
retry.

## Security & data handling

Multi-tenant isolation via row-level security. Sign-in is company-email only through Microsoft
Entra External ID. Every submission and consent action is hash-chained into an immutable audit
log. Candidate PII is kept out of logs and URLs. Cost and model usage are visible per action to
admins.

## Frequently asked questions

**What exactly is SyncTalent.ai?**
An end-to-end, agent-driven automation platform for IT staffing agencies. Six AI agents plus an
orchestrator take a requirement from chaotic inbound to a submitted, monitored candidate.

**Does it replace my recruiters?**
No. It removes repetitive work — dedup, sourcing, outreach, packaging, status monitoring — so
recruiters spend time on judgment. Humans stay in the loop at routing and escalation.

**How does it handle work authorization?**
As a hard filter at every step, covering USC, GC, EAD, H1B, and OPT. A candidate who fails is
never matched or submitted, and there is no override.

**What happens to duplicate requirements?**
They are never deleted. Duplicate vendor blasts collapse into one canonical job and each vendor
is kept as a route, preserving margin leverage.

**Does it integrate with my ATS?**
Yes — Bullhorn via Merge.dev, reading requirements and writing submissions back.

**Is there an AI voice screen?**
Yes. Aria can screen by AI voice call or hand off to a human, capturing rate, availability, and
work authorization into a structured scorecard.

**How do you control AI cost?**
Every model call routes through an AI gateway that defaults to a fast, low-cost model and
escalates only when confidence is low. Cost is tracked per action and shown to admins.

**What if you never place anyone?**
Then you pay compute only — typically a few dollars to a few hundred depending on volume. The
outcome fee exists only when there is an outcome.

**Is there a contract or minimum commitment?**
No. Prepaid wallet, no minimum, pause anytime, unused balance stays yours.

**When is the placement fee charged?**
When a candidate sourced through the platform confirms their start date, backed by an auditable
trail from requirement to placement.

**How long does onboarding take?**
Design partners are typically live on their own requirements within days, not quarters — sign-in
is company-email SSO and ATS sync is a connector, not a migration.

**How do I get started?**
Schedule a demo at https://synctalent.ai/book/ and we walk through the platform on your own
live requirements.

## Comparisons — how we position against alternatives

Every comparison page carries a "where they are stronger" section and the date the competitor
facts were last verified. SyncTalent.ai is not an ATS, holds no candidate database, and does
not help a desk starved of requirements.

- vs Recruiterflow: https://synctalent.ai/compare/synctalent-vs-recruiterflow/ — Recruiterflow is
  an ATS and CRM system of record. SyncTalent.ai is an agent layer and writes into the ATS you run.
- vs Loxo: https://synctalent.ai/compare/synctalent-vs-loxo/ — Loxo solves candidate supply with a
  proprietary people-search database. SyncTalent.ai solves coordination on a bench you already have.
- Bullhorn AI automation alternative: https://synctalent.ai/compare/synctalent-vs-bullhorn-ai/ —
  not a replacement. Bullhorn stays the record; agents handle judgement on unstructured input.
- AI agents vs manual bench sales: https://synctalent.ai/compare/ai-agents-vs-manual-bench-sales/ —
  agents remove waiting states; humans keep judgement, relationships, and the novel cases.

## Integrations — what is actually shipped

- Bullhorn (LIVE): https://synctalent.ai/integrations/bullhorn/ — reads job orders, candidates,
  contacts, companies and work authorization; writes submissions and notes back, via Merge.dev.
- Merge.dev (LIVE): https://synctalent.ai/integrations/merge-dev/ — the unified ATS connector
  layer. ATS credentials stay with Merge and never reach SyncTalent.ai.
- JobDiva (ROADMAP, not available today): https://synctalent.ai/integrations/jobdiva/
- Ceipal (ROADMAP, not available today): https://synctalent.ai/integrations/ceipal/

Requirement ingestion from email, Telegram, WhatsApp, Dice, Slack and Teams needs no ATS
connector at all. Dedup, canonical requirements, shortlisting, outreach and screening all work
without one; what the connector adds is automatic submission write-back.

## Glossary — definitions

- Bench sales: marketing an agency's already-hired but currently unassigned consultants to vendors
  and end clients so they start billing again. https://synctalent.ai/glossary/bench-sales/
- C2C (corp-to-corp): a contract engagement where one incorporated business bills another for a
  consultant's services. https://synctalent.ai/glossary/c2c-corp-to-corp/
- Time-to-redeploy: days between a consultant rolling off one assignment and confirming a start
  date on the next. https://synctalent.ai/glossary/time-to-redeploy/
- Right-to-represent (RTR): a consultant's written permission for one vendor to submit them to one
  requirement. https://synctalent.ai/glossary/right-to-represent/
- Vendor blast: mass distribution of one requirement across many vendors, so the same role reaches
  a desk repeatedly in different wording. https://synctalent.ai/glossary/vendor-blast/
- W2 vs C2C vs 1099: the three US contract tax terms — employee of the agency, company-to-company
  billing, and independent contractor. https://synctalent.ai/glossary/w2-vs-c2c-vs-1099/
- Submittal-to-interview ratio: the share of submissions converting to a client interview.
  https://synctalent.ai/glossary/submittal-to-interview-ratio/
- Prime vendor: holds the direct contract with the end client and distributes requirements to
  subcontracting vendors. https://synctalent.ai/glossary/prime-vendor/
- Implementation partner: delivers a project or rollout for an end client and staffs it from its
  own and subcontracted people. https://synctalent.ai/glossary/implementation-partner/
- Hotlist: the list of consultants an agency currently has available, circulated to vendors.
  https://synctalent.ai/glossary/hotlist/

## Articles

- Why time-to-redeploy is the only number that matters: https://synctalent.ai/blog/time-to-redeploy/
- Duplicates are not noise — they are routes: https://synctalent.ai/blog/duplicates-are-routes/
- Tiered models behind an AI gateway: https://synctalent.ai/blog/tiered-models-ai-gateway/
- How to reduce bench time in US IT staffing: https://synctalent.ai/blog/how-to-reduce-bench-time/
- What is bench sales? A 2026 primer for agency owners: https://synctalent.ai/blog/what-is-bench-sales/
- Why work authorization should be a hard filter, not a preference: https://synctalent.ai/blog/work-authorization-hard-filter/
- The real cost of chasing the same requirement twice: https://synctalent.ai/blog/cost-of-chasing-twice/
- Semantic matching vs Boolean search for resume screening: https://synctalent.ai/blog/semantic-vs-boolean-matching/
- AI voice screening: what it captures that a phone screen misses: https://synctalent.ai/blog/ai-voice-screening/
- Building an AI cost governance layer for a staffing desk: https://synctalent.ai/blog/ai-cost-governance-staffing-desk/
- Bullhorn + AI agents: what a write-back workflow actually looks like: https://synctalent.ai/blog/bullhorn-write-back-workflow/

## Canonical links

- Home: https://synctalent.ai/
- How it works: https://synctalent.ai/how-it-works/
- Pricing: https://synctalent.ai/pricing/
- Bench ROI calculator: https://synctalent.ai/roi/
- Integrations: https://synctalent.ai/integrations/
- Compare: https://synctalent.ai/compare/
- Glossary: https://synctalent.ai/glossary/
- FAQ: https://synctalent.ai/faq/
- Support: https://synctalent.ai/support/
- Contact: https://synctalent.ai/contact/
- Blog: https://synctalent.ai/blog/
- About: https://synctalent.ai/about/
- Company: https://synctalent.ai/company/
- Schedule a demo: https://synctalent.ai/book/
- Terms: https://synctalent.ai/terms/ · Privacy: https://synctalent.ai/privacy/ · Cookies: https://synctalent.ai/cookies/