Sarvam

Sarvam prices speech-to-text at ₹30/hour and text-to-speech at ₹3 per 1,000 characters, so an agency can quote a client retainer against metered usage instead of staffing a multilingual call desk. The Starter tier carries ₹0 with universal credits that never expire and no monthly commitment, which means an agency can prototype a Hindi or Tamil voice agent before signing a delivery contract. The strategic stake is geography: agencies serving India-focused clients gain a defensible capability, while English-only Western shops get little from the same subscription.

Sarvam's economics reward agencies that reuse a single voice agent build across many Indic languages. Text-to-speech covers 11 Indic languages and speech-to-text with diarization covers 12, so a Hindi appointment-booking agent can be redeployed for Marathi, Tamil, or Bengali clients by swapping language configuration rather than rebuilding dialogue logic. The Sarvam Voice Agent Starter is priced at $2,250 with 20h setup for one language; each additional language adds configuration hours, not a new build, which lifts effective margin per retainer. Usage sits on the client side: ₹30/hour for speech-to-text and ₹3 per 1,000 characters for TTS, so quote volume estimates before signing. A Pune agency serving clinics across three states can sell one agent architecture three times, then layer document digitization or dubbing as upsells. English-only Western clients see no multiplier, so screen for Indic-language call volume first.

Residual Labor Ratio is the share of call handling that still needs a human after an AI voice agent goes live: exceptions, escalations, consent callbacks, and transcript review. It matters because agencies price retainers on the assumption that deployment removes labor, when in practice it relocates labor into supervision. The ratio is measurable before you quote: pull 100 real call recordings, count how many end without human touch, then multiply the remainder by your loaded hourly rate. A clinic booking line may resolve 80 of 100 calls end to end, leaving 20 escalations at six minutes each, roughly two hours of oversight per 100 calls. Platforms differ in how much they absorb: Trillet verifies caller identity and executes actions in live systems, while Goodcall leans on knowledge sources and CRM connections, and Ruby keeps trained humans in the loop by design. Quote the residual, not the demo.

Escalation debt is the accumulated cost of every call an AI voice agent mishandles, misroutes, or drops without a clean human handoff. It does not appear on a usage invoice; it surfaces later as client churn, refunds, and remediation hours. Agencies should track three numbers per deployment: escalation rate, time-to-human, and repeat-contact rate. A missed-call follow-up agent that recovers bookings but routes billing disputes to voicemail builds debt faster than it books revenue. Trillet's audit-trail design and identity verification show why regulated clients treat escalation logs as compliance evidence, not just support metrics. The framework matters because voice retainers are priced on call volume while risk is priced on escalation quality. Before quoting a monthly fee, run 50 real calls through the agent, count every unresolved path, and price the human labor that remains. Escalation debt compounds quietly until renewal, when the client has a call log you cannot defend.

Adopt Sarvam when the client's customer conversations happen in Indic languages and the use case is one of Sarvam's pre-built agent patterns; skip it when the client's callers are English-only.

Quote the retainer only after you have priced the residual labor: the minutes of human review, exception handling, and transcript QA that remain per 1,000 calls after go-live.

IF your client base takes calls, WhatsApp messages, or documents in Hindi, Tamil, Telugu, or any of the 11 Indic languages Sarvam's text-to-speech covers, THEN the Starter plan at ₹0 with universal credits that never expire lets an agency prototype a cart-recovery or appointment-booking agent before committing to a retainer. IF the client is English-only or the deal depends on white-label rebranding, THEN skip Sarvam: it is an API and pre-built agent platform, not a white-label reseller product, and its value collapses outside Indic-language demand.

A productized Sarvam deployment that puts a Hindi or regional-language voice agent on a client's existing phone or WhatsApp line, billed as a monthly retainer with usage passed through at Sarvam's published rates.