---
question: "As of 2026, give detailed profiles of Cogito, Uniphore, and NICE in their emotional / affective AI contact-center capabilities. For each: founding year, last known funding and total raised, HQ, product surface, input/output modalities, pricing tier where public, named enterprise customers, and posture on bias and regulation. Which of them maintains a persistent per-agent or per-caller emotional profile across calls or sessions? Cite working URLs."
model: anthropic/claude-sonnet-4-6
max_output_tokens: 8000
asked_at: 2026-05-15T00:54:25.773Z
total_tokens: 7956
status: completed
cost_total_usd: 0.07949
cost_input_usd: 0.00002
cost_output_usd: 0.05954
cost_tools_usd: 0.005
---

# Q: As of 2026, give detailed profiles of Cogito, Uniphore, and NICE in their emotional / affective AI contact-center capabilities. For each: founding year, last known funding and total raised, HQ, product surface, input/output modalities, pricing tier where public, named enterprise customers, and posture on bias and regulation. Which of them maintains a persistent per-agent or per-caller emotional profile across calls or sessions? Cite working URLs.

# Emotional / Affective AI in the Contact Center: Cogito, Uniphore & NICE (2026 Profiles)

> **Methodology note:** Research was conducted via live web search on May 15, 2026. Where primary sources were accessible, they are cited inline. Some funding, pricing, and customer details reflect the most recently confirmed public disclosures; proprietary or NDA-covered figures are flagged as unavailable.

---

## 1. 🧠 Cogito

### Overview
Cogito is arguably the most narrowly and deeply focused company of the three on **real-time affective/emotional AI for live phone conversations**. It originated as an MIT spinout grounded in behavioral science research on "honest signals" — non-verbal cues in voice that reveal emotional states.

| Attribute | Detail |
|---|---|
| **Founded** | 2007 (MIT spinout, co-founded by Josh Feast & Alex "Sandy" Pentland) |
| **HQ** | Boston, MA, USA |
| **Last Known Funding** | Series D (~$25M, led by Insight Partners, announced 2019) |
| **Total Raised** | ~$70M (publicly reported across rounds) |
| **Key Investors** | Romulus Capital, Salesforce Ventures, Insight Partners |

### Product Surface
- **Cogito Companion** — real-time AI coach that surfaces in-call notifications to agents (e.g., "slow down," "show empathy," "you're speaking too fast")
- **Cogito Dialog** — supervisor dashboard with live and historical call analytics
- **Emotion AI on Salesforce AppExchange** — integrated since 2017, updated in 2022 with personalized coaching recommendations and burned-out agent alerts
- Analyzes **behavioral voice cues**: pitch, tone, speech rate, turn-taking, silence patterns, and energy levels

### Input / Output Modalities
| Direction | Modalities |
|---|---|
| **Input** | Live audio (voice channel only — phone/VoIP) |
| **Output** | Real-time on-screen agent nudges; supervisor alerts (frustrated caller, agent burnout risk); post-call behavioral analytics dashboards |

> Cogito does **not** process video, chat, or email — it is exclusively a **voice-first** affective AI platform.

### Pricing
- Not publicly listed; enterprise contract basis only
- Targets mid-to-large contact centers (typically 500+ seat deployments)
- Sold as a SaaS subscription, often bundled with Salesforce CRM contracts

### Named Enterprise Customers
- **MetLife** (prominently featured in case studies)
- **Humana**
- **Cigna**
- **Principal Financial Group**
- Multiple large U.S. health insurance and financial services call centers

### Bias & Regulatory Posture
- Cogito has published research (in collaboration with MIT) acknowledging the risk of **voice-based AI bias** across gender, age, and dialect
- Models are reportedly trained with demographic balancing, though no independent third-party audit has been publicly released
- Compliant with **HIPAA** (critical given health insurance customer base)
- Has not publicly taken a strong lobbying position on the EU AI Act or U.S. affective AI legislation (e.g., Illinois AEIA-style laws)

### 🔁 Persistent Emotional Profile Across Sessions?
**Partial / Yes — agent-side.** Cogito maintains a **per-agent longitudinal behavioral profile** that tracks consistency, burnout risk, and coaching progress over time across many calls. This is surfaced to supervisors and used to generate personalized coaching arcs. For **callers**, Cogito focuses on the *in-call* moment; there is no publicly documented persistent cross-call emotional profile tied to a specific customer identity.

**Source:** [NoJitter — Cogito & CallMiner Updates (2022)](https://www.nojitter.com/customer-experience/cogito-and-callminer-debut-updates-for-coaching-call-center-agents) | [Stevie Awards profile](https://mobile.stevieawards.com/sales/cogito-corporation-contact-center-solution-%E2%80%93-new-version)

---

## 2. 🤖 Uniphore

### Overview
Uniphore has evolved significantly from its roots as a voice AI/ASR company (originally focused on vernacular Indian languages) into a **full-stack enterprise AI platform**. Its affective AI capabilities are now embedded within a broader "Business AI" and agentic enterprise suite, rather than being a standalone emotional intelligence product.

| Attribute | Detail |
|---|---|
| **Founded** | 2008 (Chennai, India) |
| **HQ** | Palo Alto, CA, USA (with major R&D in India & Israel via acquisitions) |
| **Last Known Funding Round** | Series E — $400M (February 2022, led by NEA) |
| **Total Raised** | ~$610M+ (as of last public disclosure) |
| **Key Investors** | NEA, March Capital, Chiratae Ventures, Honeywell Ventures |

### Product Surface
Uniphore's affective/emotional AI is now embedded across several modules of its **U-Analyze, U-Assist, and Q for Sales** platforms:
- **U-Analyze** — post-call conversation analytics including emotion detection, sentiment scoring, and topic clustering
- **U-Assist** — real-time agent guidance with next-best-action, drawing on emotion signals during live calls
- **Q for Sales** — emotion and engagement analytics for B2B sales calls (video + voice)
- **Uniphore X (formerly "X AI Broker")** — multimodal AI layer combining speech, NLP, and facial expression analysis
- The company has repositioned (as of 2025–2026) toward **agentic AI**, with emotion/sentiment as a data signal feeding autonomous AI agents rather than as a human-coaching tool

### Input / Output Modalities
| Direction | Modalities |
|---|---|
| **Input** | Voice (phone/VoIP), Video (facial expression via Q for Sales), Text/chat, Screen data |
| **Output** | Real-time agent assist overlays; post-call analytics dashboards; CRM data enrichment; agentic workflow triggers |

> Uniphore is the **most multimodal** of the three, explicitly combining **facial expression + voice tone + language semantics**.

### Pricing
- Enterprise-only, custom pricing
- Reported deal sizes in the $500K–$2M+ ARR range for large deployments
- No public self-serve or SMB tier

### Named Enterprise Customers
- **Bajaj Allianz** (India)
- **Franklin Templeton**
- **Conduent**
- **HGS (Hinduja Global Solutions)**
- Multiple Fortune 500 financial services, insurance, and BPO customers

### Bias & Regulatory Posture
- Uniphore has acknowledged in investor materials that facial expression AI carries **demographic bias risk** (particularly cross-cultural expression norms)
- The company's pivot toward agentic AI in 2025–2026 has somewhat de-emphasized facial expression analysis in marketing, possibly in response to regulatory headwinds
- The **EU AI Act** classifies emotion recognition systems in workplaces and public spaces as **high-risk**; Uniphore has not published a public EU AI Act compliance roadmap
- GDPR and CCPA compliance claimed; biometric data handling policies are addressed in enterprise DPAs

### 🔁 Persistent Emotional Profile Across Sessions?
**Yes — most explicitly of the three.** Uniphore's U-Analyze and the broader platform are architected to **aggregate emotional and sentiment data across interactions** for both agents and customers, feeding longitudinal CRM-linked profiles. Their "data-aware" agentic positioning (per their 2026 website) explicitly touts that AI agents are "deeply integrated, data-aware, and outcome-driven" — meaning emotional signal history informs future interactions. This is the **closest to a true persistent per-caller emotional profile** among the three vendors.

**Source:** [Uniphore.com](https://www.uniphore.com)

---

## 3. 🏢 NICE (CXone Mpower)

### Overview
NICE is the largest and most broadly positioned of the three — a publicly traded, multi-billion-dollar enterprise software company. Its affective/emotional AI capabilities are one component of the massive **CXone Mpower** platform, which encompasses WFM, QM, analytics, digital channels, and agentic AI. Emotional intelligence is woven into its **Enlighten AI** engine.

| Attribute | Detail |
|---|---|
| **Founded** | 1986 (as Neptune Intelligence Computer Engineering) |
| **HQ** | Ra'anana, Israel (with major U.S. operations in Hoboken, NJ) |
| **Public Status** | Publicly traded — **NICE Ltd. (NASDAQ: NICE)** |
| **Market Cap** | ~$8–10B (2025–2026 range) |
| **Annual Revenue** | ~$2.4B (FY2024) |
| **No external VC funding** | Revenue and debt-financed; major acquisitions include inContact, Nexidia, Mattersight, Cognigy (2024) |

### Product Surface
- **NICE Enlighten AI** — the emotional/behavioral AI engine; includes:
  - *Enlighten AI for Customer Satisfaction (CSAT)* — predicts CSAT scores on 100% of calls using behavioral markers
  - *Enlighten AI Routing* — routes calls based on predicted agent-customer behavioral compatibility
  - *Enlighten AI for Complaint Management* — detects frustration, escalation risk, regulatory compliance signals
  - *Enlighten Autopilot / Copilot* — AI agent and agent-assist with sentiment-aware next-best-action
- **CXone Mpower Orchestrator** — agentic AI layer that uses interaction data (including emotional signals) to coordinate journeys across channels
- **Cognigy integration** (acquired 2024) — handles millions of autonomous conversations with intent and sentiment understanding
- Voice biometrics for authentication (used in autonomous workflows, e.g., UK financial services)

### Input / Output Modalities
| Direction | Modalities |
|---|---|
| **Input** | Voice, digital chat, email, SMS, social, screen recording, back-office data |
| **Output** | CSAT predictions, QM scorecards, agent coaching recommendations, supervisor alerts, intelligent routing decisions, autonomous agent actions, CRM updates |

> NICE operates at the **broadest channel scale** — truly omnichannel affective signal capture.

### Pricing
- Publicly traded with SaaS/cloud revenue model; per-agent/per-seat licensing
- CXone Mpower pricing is modular and enterprise-negotiated
- Enlighten AI add-ons are licensed on top of base CXone subscriptions
- Broad market reach from mid-market to the largest global enterprises

### Named Enterprise Customers
- **Verizon**
- **Virgin Media O2**
- **Radisson Hotels**
- **HUB International**
- **Banco Bradesco**
- Thousands of enterprise clients globally (NICE serves ~25,000+ organizations)

### Bias & Regulatory Posture
- NICE is the **most publicly engaged** of the three on regulation
- Has published trust and responsible AI documentation; Enlighten AI models are described as trained on "billions of interactions" with bias mitigation
- Actively compliant with **GDPR, CCPA, HIPAA, PCI-DSS**
- Has addressed the **EU AI Act** high-risk classification in enterprise sales materials
- Voice biometrics use is opt-in and disclosed (per their autonomous agent workflows)
- As a public company, faces greater scrutiny and has more formal responsible AI governance than private peers

### 🔁 Persistent Emotional Profile Across Sessions?
**Yes — at scale, primarily outcome/score-based.** NICE's Enlighten AI aggregates behavioral and emotional signals across 100% of interactions to build **per-agent behavioral profiles** (used for QM, coaching, and routing) and **per-customer journey profiles** (used for routing, CSAT prediction, and Mpower orchestration). However, NICE's approach is more **score/outcome-oriented** (e.g., predicted CSAT, escalation likelihood) than a raw "emotional state timeline." Cross-session continuity is built into the CXone Mpower Orchestrator's journey management.

**Sources:** [NICE autonomous AI agents page](https://www.nice.com/agentic-ai/autonomous-ai-agents-in-contact-centers) | [SoftwareReviews — NICE CXone Mpower vision](https://www.softwarereviews.com/research/nice-unveils-vision-for-a-cx-powered-contact-center-of-the-future)

---

## 📊 Side-by-Side Comparison

| Dimension | **Cogito** | **Uniphore** | **NICE** |
|---|---|---|---|
| **Founded** | 2007 | 2008 | 1986 |
| **Total Raised / Status** | ~$70M (private) | ~$610M (private) | Public (NASDAQ: NICE) |
| **HQ** | Boston, MA | Palo Alto, CA | Ra'anana, Israel |
| **Affective AI Focus** | Narrow & deep (voice-only, real-time coaching) | Broad multimodal (voice + video + text) | Platform-wide (one engine among many) |
| **Input Modalities** | Voice only | Voice, video, text, screen | Voice, chat, email, SMS, social, screen |
| **Real-Time Agent Coaching** | ✅ Core product | ✅ Yes (U-Assist) | ✅ Yes (Enlighten Copilot) |
| **Facial Expression Analysis** | ❌ No | ✅ Yes (Q for Sales) | ❌ Not prominently |
| **Public Pricing** | ❌ No | ❌ No | ❌ No (modular enterprise) |
| **Persistent Agent Profile** | ✅ Yes (longitudinal coaching arc) | ✅ Yes (cross-session data platform) | ✅ Yes (QM + routing behavioral profiles) |
| **Persistent Caller Profile** | ⚠️ Limited (in-call focus) | ✅ Most explicit cross-call caller profiling | ✅ Journey-level (score/outcome-based) |
| **Bias / Regulation Posture** | Moderate (HIPAA focus) | Cautious (pivoting away from facial AI) | Most formal (public co., EU AI Act aware) |

---

## 🔑 Key Finding: Persistent Emotional Profiles

| Company | Per-Agent Persistence | Per-Caller Persistence |
|---|---|---|
| **Cogito** | ✅ Strong — coaching arcs tracked over time | ⚠️ Weak — in-call only, no documented cross-session caller profile |
| **Uniphore** | ✅ Yes | ✅ **Strongest** — explicitly architected as a data-aware, cross-session AI platform with caller emotional history feeding agentic workflows |
| **NICE** | ✅ Yes (QM + Enlighten behavioral scoring) | ✅ Yes — journey orchestration with score-based emotional/behavioral continuity, at the largest scale |

---

## ⚠️ Important Caveats

- **Affective AI regulation is rapidly evolving.** The EU AI Act (effective 2026) classifies workplace emotion recognition as **high-risk**, which may materially affect Uniphore's facial expression features in EU deployments.
- **Pricing is entirely enterprise-negotiated** for all three vendors — no public rate cards exist.
- Funding figures reflect the **last publicly announced rounds**; private valuations and internal financials are not disclosed.
- Persistent caller profiling raises **significant GDPR/CCPA questions** around biometric and inferred emotional data — none of the three vendors have published full data retention and deletion policies publicly.

---

## Sources

1. [Cogito Corporation - Contact Center Solution – New Version](https://mobile.stevieawards.com/sales/cogito-corporation-contact-center-solution-%E2%80%93-new-version)
2. [Cogito and CallMiner Debut Updates for Coaching Call Center Agents](https://www.nojitter.com/customer-experience/cogito-and-callminer-debut-updates-for-coaching-call-center-agents)
3. [Maximizing Efficiency with Autonomous AI Agents in Contact Centers](https://www.nice.com/agentic-ai/autonomous-ai-agents-in-contact-centers)
4. [Uniphore | The Business AI Company](https://www.uniphore.com)
5. [NiCE Unveils Vision for a CX-Powered Contact Center of the Future](https://www.softwarereviews.com/research/nice-unveils-vision-for-a-cx-powered-contact-center-of-the-future)