How Portea went from staff shortages to a record hiring month with Sarvam

Background

Portea Medical is one of India's largest home healthcare companies. It sends trained caregivers, nurses, physiotherapists and doctors into patients' homes, rents and sells medical equipment, and runs patient support programmes for pharmaceutical companies.

In 13 years, Portea has served more than 2 million patients across 120+ cities, completed more than 10 million duties and worked with more than 100,000 healthcare staff.

Behind those numbers are two groups that Portea has to keep bringing together. On one side are families looking for care. On the other are nurses, caregivers and physiotherapists looking for work.

When hiring fell behind demand

Early in 2026, Portea ran short of staff. Between January and May, it could not service a significant number of incoming leads because it did not have enough people available. In March and April, daily new service starts were 30% lower than a year earlier.

The bottleneck was the first call. Recruiters spent too much of their time on first calls, often to people who didn't answer or weren't interested. Each market needed its own language, from Hindi and Hinglish to Tamil, Telugu and Kannada, and non-Hindi markets were the hardest to cover. A new applicant got a call only when a recruiter found the time, and that could take days. In March 2026, before the pilot began, 117 recruitment leads waited more than three days for the first call.

Portea also had a hiring pool it couldn't reach. Over more than a decade it had worked with 100,000+ healthcare staff, and many of them could be open to returning. Calling through years of former caregivers by hand was not feasible.

The same limit applied to patients. Sales teams could try each customer only a limited number of times, and enquiries that arrived after business hours did not always get a timely reply. For a family looking for care, a delay in that first conversation could affect whether they went ahead. Former patients, lapsed physiotherapy customers, equipment rental enquiries and incomplete bookings added up to a large base that human calling alone could not reach consistently.

Why Sarvam

Portea weighed its options. Hiring more recruiters and salespeople cost too much. Building in-house on open speech models and its own telephony would have taken disproportionate effort from a lean team, and dedicated infrastructure made little economic sense for demand that came in bursts. Another voice AI platform Portea evaluated fell short on cost and on speech quality in Indian languages.

Portea chose Sarvam for the quality of its colloquial speech and Indian accents. That mattered most on caregiver calls, which often run in Hinglish and other Indian languages.

It started with a pilot. Over 17 days from mid-March 2026, a voice agent called Neha phoned 10,000 nursing attendants who had worked with Portea in 2024 and 2025 and asked if they wanted to come back. More than 50 people joined Portea after first being connected through the agent.

About two weeks in, two former staff members walked into Portea's office after getting a call. They said the conversation reminded them of their time with Portea and prompted them to get back in touch. It showed Portea that a staff base it had found impractical to call at scale was now within reach.

Where the voice agents work now

After the pilot, Portea expanded Sarvam across its staff and patient workflows.

• Calling former caregivers, from staff lists going back to 2015, to check interest before passing candidates to recruiters
• Making the first call to every new caregiver and nurse lead, so recruiters spend their time on people who have already shown interest
• Reconnecting with former patients of its caregiver services to find those who need care now
• Following up on enquiries that didn't convert, including lapsed physiotherapy enquiries, equipment rentals and incomplete bookings
• Creating staff training and engagement content in local languages with Sarvam's dubbing capabilities

Hinglish is the default on staff calls, with Kannada, Tamil, Telugu, Bengali and Marathi used for local hiring. Patient campaigns currently run in English. The agents place outbound calls and answer people who call back after missing one.

Sarvam works inside the systems Portea already runs. Portea uses Exotel for telephony, and its own CRM builds campaign batches that go to Sarvam through the Campaigns API. When a candidate says yes, the response posts to Portea's recruitment lead API and shows up in the recruiter portal. When a patient says yes, a lead is created in the sales CRM for the sales or operations team to close. Around August 2026, Portea was running about 2,000 calls a day across staff and patient workflows.

The impact

Portea has engaged more than 80,000 prospective staff through Sarvam's voice agents. Even after adjusting for normal seasonal swings, hiring grew more than 33% year on year, and July 2026 was Portea's highest hiring month on record. A few months earlier, staff shortages had contributed to fewer new service starts.

New leads no longer wait. Every new recruitment lead now gets a first call the same day or the next.

On the patient side, during a trial period of about 2.5 months, Portea engaged more than 40,000 patients, and more than 150 customers bought from Portea through these campaigns. Because each added conversation costs relatively little, Portea also reported a meaningful drop in its blended customer acquisition cost.

Recruiters no longer make every first call or chase stale leads. Sales teams spend their time on patients who have shown interest or need to speak to a person.

Looking ahead

Portea plans to extend Sarvam further across the patient side. The next phase includes patient reactivation at a larger scale, beginning with WhatsApp-led campaigns across its historical patient base. It also covers first responses outside business hours on patient enquiry lines, visit confirmation and reminder calls, and wider inbound coverage across patient leads.

Portea also plans voice-to-WhatsApp handoffs, so a patient who says yes on a call can move straight to booking and payment. The longer-term goal is one continuous path from the first conversation to a confirmed visit.