Company details

HEADCOUNT

100-499

ORG TYPE

Startup

SECTOR

Technology & Digital

WEBSITE

https://www.poly.ai

About the company

PolyAI builds voice AI agents for enterprise customer service, aiming to handle real phone conversations in a natural, human-like way. PolyAI was founded in 2017 by Nikola Mrkšić, Tsung-Hsien Wen, and Pei-Hao Su, with roots in Cambridge’s dialogue systems research community. The company markets its platform around large-scale deployments, multilingual coverage, and serving major brands with high call volumes. Recent public announcements focus on scaling product capability and customer adoption.

Locations and presence

PolyAI is set up for flexible work, with roles spanning the UK, US, and other countries depending on the job. Public listings show office presence across London, New York, San Francisco, and Belgrade, alongside location-specific roles in places like Turkey and Serbia.

Palpable Score

62.6

/ 100

PolyAI has a real on-ramp for new grads in at least one clearly-labelled role, and the company pairs that with solid benefits and equity language. The limiting factor is consistency: most openings skew senior, salary transparency varies a lot by location, and early-career outcomes data is patchy and mixed depending on team.

Pillar 1: Early-career access

Score

11.0

/ 20

The company runs at least one explicitly entry-level opening, with the Agent Designer role stating “new graduates welcome.”
PolyAI’s current openings list is dominated by senior requirements such as the Platform Software Engineer role asking for 5 years and multiple management-level roles.
The company includes some potentially accessible non-senior routes (for example a part-time Language Specialist contract), but there are not many 0–3 year roles live at once.

Pillar 2: Hiring fairness and transparency

Score

13.7

/ 20

The company writes unusually concrete responsibilities and minimum requirements in the Agent Designer posting, which helps candidates self-assess quickly.
PolyAI includes practical transparency signals in postings, like stating when visa sponsorship is not offered and when office attendance is required for hybrid roles.
The company shares salary ranges for some US roles (for example Community Manager and Revenue Operations Manager), but that transparency is not consistent across regions and functions.

Pillar 3: Learning and support

Score

12.3

/ 20

The company offers an annual learning and development allowance and positions “learning and leadership development” as a standard benefit.
PolyAI’s Agent Designer role bakes in hands-on feedback loops like listening to real customer calls and iterating using measurable metrics, which can accelerate early skill growth.
The company does not publish much role-level detail on mentoring, pairing, structured onboarding, or formal early-career coaching, which caps confidence in day-to-day support.

Pillar 4: Pay fairness and stability

Score

13.3

/ 20

The company publishes base salary ranges for some US roles (for example $90,000–$107,000 for Community Manager and $115,000–$125,000 plus bonus for Revenue Operations Manager).
PolyAI repeatedly includes equity in role descriptions and pairs that with a benefits package covering healthcare, PTO policies, and family support.
The company does not consistently publish salary ranges across roles and locations, and at least one opening is explicitly part-time contract, which reduces the “stability and clarity” signal for early-career candidates.

Pillar 5: Early-career outcomes

Score

12.3

/ 20

The company has strong overall employee sentiment signals in public reviews, including high “recommend to a friend” rates and strong scores for career opportunities, but these are not early-career specific.
PolyAI has at least one publicly visible early-career datapoint from a Graduate Software Engineer review describing smart colleagues and strong learning-by-proximity.
The company also has function-specific negative feedback in public reviews (for example sales-focused comments about limited career growth and lack of training), suggesting outcomes can vary significantly by team and role.