Bland AI
by Bland
Enterprise voice AI phone agents on in-house speech and language models, with self-hosted deployment for regulated calls
Bland AI is an enterprise platform for building, running and monitoring AI phone agents that handle long, regulated calls in healthcare, insurance and financial services. It runs its own speech-to-text, language and text-to-speech models and offers on-prem or VPC deployment, which suits contact-center leaders replacing IVRs and outbound call teams.
Bland AI, the San Francisco company also known as Bland, builds AI phone agents for inbound and outbound calls, with SMS and web chat on its Enterprise tier. Bland runs its own speech-to-text, language model and text-to-speech stack, and does not let customers swap in OpenAI or Anthropic models. The company says this lets it tune for latency, interruptions and long conversations; its typical calls run 30 to 45 minutes, according to The Next Web. Agents are built, tested against simulated callers and unit tests, monitored for why each decision was made, and changed with version locking. A July 2026 update added five interruptibility levels per conversation node. The platform supports more than 40 languages, with real-time translation in 23. Sensitive workloads can run self-hosted, on-prem or in a VPC. Bland lists SOC 2 Type I and Type II, HIPAA with a BAA on Enterprise, GDPR and self-attested PCI DSS. In June 2026 it raised a $50 million Series C led by Dell Technologies Capital, taking total funding past $100 million, with HubSpot Ventures, Emergence Capital, Scale Venture Partners and Y Combinator among backers. Co-founded by CEO Isaiah Granet and Sobhan Nejad, it reports more than 250 enterprise customers, including Samsara, Kin Insurance and CNO Financial Group, and over 3.5 million calls a week. It competes with PolyAI, Retell AI, Vapi, Cognigy and Replicant.
A contact-center or operations leader in healthcare, insurance or financial services who needs to automate long, regulated phone calls and may need the stack self-hosted.
Phone agents that handle complex calls end to end on Bland's own models, with on-prem or VPC deployment and a BAA available on Enterprise.
At a Glance
- Category
- Audio & Voice
- Pricing
- Usage-based, Subscription, Contact for pricing
- Target Market
- Contact Center Leaders, CIOs, Enterprise Developers, Healthcare Operations, Financial Services
- Deployment
- Cloud-first, Self-hosted, API-based
- Headquarters
- San Francisco, United States
- Customers
- 250+ enterprise customers, 3.5M+ calls per week (company-reported, June 2026)
Key Features
- ✓In-house speech and language models
Bland runs its own speech-to-text, LLM and text-to-speech, so call data does not pass to third-party model vendors and latency can be tuned end to end.
- ✓Self-hosted and VPC deployment
Enterprise customers can run the stack on-prem or in a VPC with a dedicated orchestration server, which matters for HIPAA and financial-services workloads.
- ✓Testing and evaluation tools
Environments, simulated-caller evaluations, unit tests and a knowledge coverage score let teams catch agent failures before a change reaches live callers.
- ✓Call observability
Monitoring shows why an agent made each decision on a call, so operators can debug failures and measure the effect of changes.
- ✓Interruption and resumption controls
Since July 2026, five interruptibility levels and a resumption speed setting can be set per conversation node to handle callers who talk over the agent.
- ✓Multilingual calls
Agents speak more than 40 languages natively, translate in real time in 23 and can switch language mid-call for mixed-language customer bases.
Capabilities
Use Cases
- •IVR replacement
A housing finance agency retires its phone tree for a conversational agent; Bland cites Idaho Housing and Finance Association saving $750K a year.
- •Patient care coordination
A healthcare provider uses agents to guide patients through check-ins such as blood-pressure readings and escalate to staff when answers cross a threshold.
- •Insurance and benefits qualification
An insurance or benefits seller automates outbound qualification calls, which Bland says added $430K in annual revenue for American Way Health.
- •After-hours and overflow coverage
A service team routes after-hours and peak overflow calls to an agent that schedules appointments, collects data and transfers to humans when needed.
Ideal For
Best For
- ✓Replacing legacy IVR phone trees with conversational agents for inbound calls
- ✓Appointment scheduling, patient care coordination and after-hours coverage in healthcare
- ✓Outbound qualification, verification and data-collection calls at high volume
- ✓Payments and collections calls that need PCI-aware handling
- ✓Regulated enterprises that require voice AI running self-hosted or in their own VPC
Not Ideal For
- ✗Teams that want to plug in their own frontier LLM, since Bland runs only its in-house models
- ✗Small teams without someone to own call flows, testing and conversation design; production agents typically take two to six weeks per one published review
- ✗Buyers who need warm transfers, SSO, HIPAA BAAs or data residency without signing an Enterprise contract
- ✗Support operations where most volume comes through email, chat or ticketing, given no native Zendesk or Freshdesk integration was found
Deployment
Market & Ratings
250+ enterprise customers, 3.5M+ calls per week (company-reported, June 2026)
Market Analysis
Pros
- ✓Own model stack keeps call data away from third-party model vendors, a selling point for HIPAA and PCI workloads
- ✓Self-hosted and VPC options that most voice-agent platforms do not offer
- ✓Public self-serve tier with no platform fee for testing before an Enterprise contract
- ✓Testing, simulation and observability tooling for production call flows
Cons
- ✗Customers cannot swap in OpenAI or Anthropic models, which The Next Web identifies as a constraint as well as a differentiator
- ✗Many production features (warm transfers, SSO, BAA, data residency, SMS) are Enterprise-only
- ✗Pricing has extra line items (transfer minutes, outbound attempt minimums) and rose from $0.09 to $0.14/min on Start in December 2025
- ✗Competitor reviews report real-world latency of roughly 700-900ms against Bland's sub-400ms claim
- ✗Agents on the old builder must migrate to the new Agents product by November 15, 2026
Pricing
Start
$0
- ✓$0.14/min talk time
- ✓$0.05/min transfer time
- ✓10 concurrent calls, 100 calls per day
- ✓Free credits and an inbound number, no card required
Build
From $299/mo
- ✓$0.12/min talk time
- ✓$0.04/min transfer time
- ✓50 concurrent calls, 2,000 calls per day
- ✓50 knowledge bases, 5 voices
Enterprise
Contact for pricing
- ✓Unlimited calls
- ✓On-prem or VPC deployment
- ✓BAA, SSO and data residency
- ✓Warm transfers, SMS, web chat, guardrails
- ✓Forward-deployed engineers
Bland bills per minute of talk time plus separate transfer minutes; Start is $0.14/min with no platform fee and Build is $299/month plus $0.12/min. A review by eesel (a competitor) notes a $0.015 minimum per outbound attempt including failed calls, and that Start rose from $0.09 to $0.14/min in December 2025. Warm transfers, SSO, BAAs, data residency and self-hosting are Enterprise-only with custom pricing.
Security & Compliance
Sources
This page was written from 6 sources, 4 on domains other than bland.ai.
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Pipecat
Open-source Python framework for real-time voice and multimodal AI agents, with a managed Pipecat Cloud
Retell AI
Developer platform for building, deploying and monitoring AI phone agents, billed per minute
AssemblyAI
Speech-to-text, speech understanding and voice agent APIs priced per hour of audio
Cresta
Contact center AI platform for autonomous AI agents, real-time agent assist and conversation intelligence