Not a contact-centre suite. Not a voice bot. The cross-enterprise intelligence and action layer for communications embedded in business workflows — collections, service, sales, employee and student support — delivered with partners, governed on Gemini Enterprise Agent Platform.
We are not selling a bot or a CC suite — we are creating the layer through which enterprises design, govern, deploy, and improve agentic interactions across every channel. Broader than conversational AI vendors, orthogonal to CCaaS, native to the Gemini Enterprise narrative.
Models are table stakes. Telephony, channel integration, compliance, state, language handling, observability, and workflow integration are the hard 80% — Google provides the platform, partners industrialize the gap into repeatable vertical IP.
In-tenant deployment (data, models, state, telemetry inside the customer's cloud) plus Indic/regional language quality is a combination point-solution vendors can't match and hyperscaler rivals haven't packaged.
Every contained interaction is model + speech + search + data consumption. One lighthouse use case ramps into an enterprise-wide agent platform commitment — track expansion, not just the first deal.
Every figure below is sourced and confidence-tagged. Use them verbatim; do not improvise market sizes in front of customers.
Eight weighted criteria. 70+ is a priority pursuit; under 45 goes back to nurture. Be honest — the pilot factory only has so many slots.
Customer Engagement Suite (GECX) owns contact-centre modernization. We own the cross-enterprise agent layer. The tree and the capability map below are the rules of engagement — deviating from them creates channel conflict that kills both deals.
No stage skipping: the pilot only performs when discovery baselined real metrics and the customer signed off one use case.
| Package | Buyer | Trigger | Handoff model | Timeline | Commercial |
|---|---|---|---|---|---|
| Collections Agent | Collections / Risk leader (BFSI, Telecom, Education fees) | Account enters DPD bucket (e.g., 1–30 days past due) or a promise-to-pay is broken. | Disputes, hardship or vulnerability cues, and legal-stage accounts route to human collectors with full context. | 8–12 weeks to production pilot on one bucket and one language pair | Platform consumption + per-completed-interaction implementation fee; expand by bucket and language |
| Inbound Service Agent | Customer Service / Contact Centre leader | Customer calls, messages, or emails on any serviced line or number. | Sentiment, complexity, or policy triggers route to human agents inside the existing contact-centre stack. | 6–10 weeks for a top-5-intents production pilot | Consumption-led; price per resolved interaction as maturity grows |
| Outbound Lead Agent | Sales / Growth / Admissions leader | New lead from web form, campaign, marketplace, or walk-in registration. | Hot leads transfer live or book meetings; complex product questions route to specialists. | 6–8 weeks on a single campaign source | Per-qualified-lead framing on top of platform consumption |
| Employee & Student Support Agent | CHRO, CIO, GCC site leader, University registrar / dean of students | Any employee or student question via chat, voice, email, or portal. | Sensitive HR matters, grievances, and judgment calls route to named human owners. | 6–10 weeks for two functions (e.g., HR + IT) in one geography | Per-seat platform economics; strong Gemini Enterprise seat-activation motion |
| Agent Assist & Conversation Intelligence | Contact Centre / Quality / Operations leader | Live human-assisted interaction, or batch analysis of recorded interactions. | Human agent stays in control throughout; the AI advises and documents. | 4–8 weeks; lowest-risk entry point because no customer-facing automation | Per-monitored-seat or per-analyzed-interaction consumption |
Tilicho Labs is the communications and implementation layer: we leverage its existing platform and IP (one-time integration fee + usage-based consumption) instead of funding redevelopment of capabilities it already has. Gemini Enterprise Agent Platform provides intelligence, agents, and governance. Validate exact Tilicho capability coverage during implementation — never present unconfirmed capabilities as production-ready.
One narrowly-defined use case, baseline metrics before build, production-like traffic before verdict. Central pilot funding (e.g., ~$250K for a strategic partner build-out) is decided per-vertical on evidence from the first two customer pilots — it is not assumed.
Those deflect within one channel. This is the layer that reasons and acts across all channels and systems — keep the channels, upgrade the intelligence behind them. Offer a conversation-intelligence audit of real traffic as a no-risk first step.
Customer Engagement Suite modernizes the contact centre. This is for communications embedded in wider workflows — collections, sales, HR, education. Where the customer needs both, we run a joint motion with explicit swimlanes. Use the decision tree; never improvise the boundary.
Demos aren't deployments. The gap is telephony, compliance, state, observability, and integration at enterprise scale — inside the customer's tenant. Startups are potential partners on our platform, not necessarily competitors.
Design decides this: disclosure, language, latency, and instant human escape hatches. Pilot metrics include CSAT and complaint rate against the human baseline — we measure it rather than argue it.
We never present outcome claims — the calculator uses the customer's own volumes and costs, shows every formula, and pilots establish the real containment rate before any scale decision.
Agents are grounded in governed sources, act only through permissioned tools, follow policy matrices they cannot override, and log everything. Compare against today's baseline: sampled QA on 2% of human calls.
Many try; the platform is the hard part — orchestration, state, evals, governance, channel infrastructure. Internal teams that start from raw APIs rebuild all of it. Offer the reference architecture as the accelerant either way.
No unsupported savings claims · no fabricated references · no unverified competitive or partner claims · no GECX conflict · no internal metrics (seat activation, consumption targets) in customer material · every number carries its source and confidence.