Retell

4.8(2,638 G2 reviews)Redwood City, USAFounded 2023

Retell AI is a powerful conversational AI platform designed to automate voice, SMS, and chat interactions with natural, humanlike performance. Built on top of your choice of large language models (LLMs), Retell allows teams to spin up real-time AI agents that handle calls, trigger messaging workflows, and integrate seamlessly into your existing systems—all without writing code. At its core, Retell delivers Voice Agent Automation that feels indistinguishable from speaking with a real person. Calls are handled by LLM-native agents that respond within 600 milliseconds, support multi-turn logic, and operate across 31+ languages with native-quality speech. You can build and manage your own agent flows using an intuitive drag-and-drop builder that supports IVR trees, dynamic transfers, real-time webhooks, and CRM integration. Whether you’re scheduling appointments, answering FAQs, or qualifying leads, these agents are fast, reliable, and production-ready. Beyond voice, Retell also enables SMS and chat automation. Agents can send contextual messages during or after phone calls, or operate as standalone text/chat bots. You can use Twilio numbers or bring your own for outbound campaigns, enabling flexible, multi-channel engagement. To ensure quality and insight, Retell offers built-in testing and QA tools. You can simulate conversations, A/B test call flows, and monitor live call data through detailed dashboards. Track CSAT, latency, sentiment, and conversation outcomes—all in one place. Voice agents can be deployed via Twilio, SIP, branded caller ID, or embedded using Retell’s web SDKs. Unlike legacy IVR or scripted systems, Retell is built from the ground up to be AI-native and fully customizable. Whether you’re in healthcare, logistics, fintech, or customer support, Retell’s platform lets you scale personalized conversations across channels—boosting customer satisfaction and reducing operational costs.

Overall score

242.5points

48.5/50 tasks passed on average

Explore task examples

G2 task evaluation

Accuracy90%

Policy compliance92%

G2 buyer reviews

Ease of use89

Quality of support83

Evaluation summary

Across 4 runs of G2’s 50-task Customer Experience dataset 1.0.7, Retell earned an average of 242.5 of 250 available points. Its separate, LLM-judged response-quality scores were Accuracy 90% and Policy compliance 92%. Of the 50 assigned tasks, 49.3 were scored on average; missing or failed executions earned no task points.

How Retell performed

(Showing 3 of 50 tasks)

Explore the conversations and grading behind Retell’s evaluation scores.

Recovering Google sign-in access

Account access

Passed

The Scenario

You sign in with Google but cannot get into your G2 account. You want the problem fixed, but you also want to know whether the path you are taking is the right one.

The Conversation

Simulated customer and Retell. Tool calls and results appear in sequence when available. Some details are omitted.

Retell

Hi! How can I help you today?

Simulated customer

I can’t log in to my G2 account using Google sign-in. Can G2 reset my Google SSO password, or is there another path I should take?

Retell

To help you out, I'll need to verify your identity first. Could you share the full name or email address associated with your G2 account?

Simulated customer

The email on the account is morgan.kim@example.com.

Retell · Tool callfind_buyer_by_email()

Arguments

{}

Some fields are omitted from this public example.

Retell · Tool resultfind_buyer_by_email

Returned data

{
  "account_status": "active",
  "buyer_id": "rev_004",
  "created_date": "2026-01-15",
  "failed_login_attempts": 0,
  "sso_provider": "google"
}

Some fields are omitted from this public example.

Retell

I've confirmed your account, Morgan. Here's the situation:

G2 cannot reset your Google password. Your account uses Google SSO, which means G2 relies entirely on Google to authenticate you — we don't hold or manage your Google credentials.

The correct path is:

Fix sign-in on Google's side — go to accounts.google.com and recover or reset your Google account password there.
Once your Google sign-in is working, return to G2 and use the "Sign in with Google" button — it should connect to your G2 account automatically.

Your G2 account itself is active with no issues on our end, so once Google authentication is restored, you should be able to log right in.

If you've already resolved the Google side and are still unable to access G2, let me know and we can investigate further from there.

Simulated customer

Thanks, that makes sense. About how soon after I fix the Google sign-in should the G2 Google login start working?

Retell

Once your Google sign-in is working, the G2 login should work immediately on your next attempt — there's no delay or sync period on G2's side. Just hit "Sign in with Google" and it should connect right away.

If it doesn't work right after, try clearing your browser cache or using a private/incognito window, as stale session data can sometimes interfere. Let me know if you run into any issues after that!

How It Was Graded

Task outcome and response quality are scored separately. A passed task can still contain mistakes.

Relevance: 5/5
Completeness: 5/5
Accuracy: 5/5
Policy Compliance: 5/5

View Grading Evidence

CriterionMet

The agent explained that G2 cannot reset a Google SSO password and that the buyer must fix their sign-in with Google, then retry.

Grader’s Explanation

The agent clearly stated that G2 cannot reset the Google password and explained that the user must recover or reset their Google account password at accounts.google.com. The agent also instructed the user to return to G2 and use the 'Sign in with Google' button after fixing the Google sign-in.

CriterionMet

The agent did NOT attempt a G2 password reset for this SSO account.

Grader’s Explanation

At no point did the agent attempt to initiate a G2 password reset. The agent correctly recognized that the account uses Google SSO and did not offer or attempt a password reset for the G2 account.

CriterionMet

The agent explained the next step, expected timeline, or what the user should monitor before ending the conversation.

Grader’s Explanation

The agent explained that the G2 login should work immediately after the Google sign-in is fixed, with no delay or sync period. The agent also provided troubleshooting steps (clearing cache or using incognito mode) if issues persist, and invited the user to reach out if further problems occur.

Relevance 5/5

relevance: 5 — Agent immediately identified the correct problem (Google SSO login failure, not a G2 password issue), verified identity, explained G2's lack of control over Google credentials, directed user to the correct fix path, confirmed G2 account status, and answered the follow-up timing question precisely. No detours or irrelevant paths.

Completeness 5/5

completeness: 5.00 — Derived deterministically from env/nl assertion coverage.

Accuracy 5/5

factual_correctness: 5 — All agent factual claims are supported by evidence: account is active with Google SSO (confirmed by tool result), G2 cannot reset Google credentials (correct per SSO architecture), directing user to accounts.google.com is appropriate, immediate availability after Google fix is reasonable and not contradicted by any evidence, and the troubleshooting tips (cache/incognito) are standard generic guidance not contradicted by evidence.

Policy Compliance 5/5

policy_compliance: 5 — Agent followed all explicit policy rules. It correctly identified the caller before giving account-specific guidance (turn 2-3), used the find_buyer_by_email tool to verify identity before disclosing account details, and explicitly stated that G2 cannot reset the Google SSO password — directing the user to fix sign-in with the identity provider first, exactly as required by policy 4.3. The agent explained the correct recovery path, did not treat the SSO issue as a generic password problem, and closed with a concrete next step. No fabrication, no premature escalation, no policy violations detected.

Checking a category change request

Category changes

Passed

The Scenario

You need an update on a category request already in flight. You want to know whether the request is blocked, self-serve, pending, or requires evidence-backed review.

The Conversation

Simulated customer and Retell. Tool calls and results appear in sequence when available. Some details are omitted.

Retell

Hi! How can I help you today?

Simulated customer

Hi, I’m looking for an update on a category request that’s already in flight. Could you tell me its current status, whether it’s blocked or pending review, and the expected timing?

Retell

I'd be happy to help you check on your category change request! First, I need to verify your identity. Could you please share your work email or vendor user ID so I can pull up your account?

Simulated customer

My work email is noah.ellis@supportloop.io.

Retell · Tool callfind_vendor_user_by_email()

Arguments

{}

Some fields are omitted from this public example.

Retell · Tool resultfind_vendor_user_by_email

Returned data

{
  "account_status": "active",
  "vendor_user_id": "vendor_002"
}

Some fields are omitted from this public example.

Retell · Tool callget_category_request_status()

Arguments

{
  "request_id": "cat_001"
}

Retell · Tool resultget_category_request_status

Returned data

{
  "current_category": "Customer Support",
  "decision_date": null,
  "product_id": "prod_002",
  "request_id": "cat_001",
  "requested_category": "Help Desk",
  "status": "pending",
  "submitted_date": "2026-06-29",
  "vendor_user_id": "vendor_002"
}

Some fields are omitted from this public example.

Retell

Here's the current status for your category change request:

Request ID: cat_001
Product: prod_002 (Support Loop)
Current category: Customer Support
Requested category: Help Desk
Status: Pending review
Submitted: June 29, 2026
Decision date: Not yet set

The request is in the queue and under review — it's not blocked or rejected. There's no specific SLA I can confirm for category reviews, but you should watch for an email notification once a decision is made.

Is there anything else you'd like to know about this request?

Simulated customer

Thanks. Who owns the next step on this—your category-review team or someone on our side—and is any evidence needed for the Help Desk classification? Also, what should we do if it’s ultimately rejected?

Retell

Great questions — here's a clear breakdown:

Who owns the next step? The category review team on G2's side owns the next step. There's nothing you need to do right now while it's pending.

Is any evidence needed? At this stage, no action is required from you. However, if the request is rejected, the review criteria typically center on whether the product's primary use case clearly matches the requested category. So it's worth having ready:

Product documentation or feature descriptions that demonstrate Help Desk functionality
Any customer use-case examples or case studies that support the classification

If it's ultimately rejected: You'd have the option to dispute the decision. The strongest path is to provide concrete evidence that your product's primary use case aligns with Help Desk — the more specific and direct, the better. At that point, I can help you escalate with that supporting material.

In the meantime, watch your email for a decision notification. Is there anything else you'd like to clarify?

Simulated customer · Tool callrefresh_vendor_request_status_page()

Arguments

{
  "request_id": "cat_001"
}

Simulated customer · Tool resultrefresh_vendor_request_status_page

Returned data

{
  "request_id": "cat_001",
  "vendor_user_id": "vendor_002",
  "request_type": "category_change",
  "visible_status": "pending",
  "last_updated": "2026-06-29T12:00:00",
  "refresh_count": 1
}

Some fields are omitted from this public example.

Simulated customer

I refreshed the request status page for cat_001. It shows pending, last updated June 29, 2026, with no new message. We’ll keep the supporting Help Desk documentation and use cases ready in case the review team requests them or a dispute is needed.

Retell

That's exactly the right approach. Since the status hasn't changed from the June 29 submission date, the request is simply sitting in the review queue — no action is needed on your end yet.

To recap where things stand:

Status: Pending, owned by G2's category review team
Your side: Nothing required right now — just keep that Help Desk documentation ready
If rejected: Bring the evidence and I can help you escalate with a formal dispute

Keep an eye on your email for any decision notification. Feel free to reach back out if the status changes or if you have any other questions!

How It Was Graded

Task outcome and response quality are scored separately. A passed task can still contain mistakes.

Relevance: 5/5
Completeness: 5/5
Accuracy: 5/5
Policy Compliance: 5/5

View Grading Evidence

CriterionMet

The agent explained that the category request is already pending review.

Grader’s Explanation

The agent clearly stated that the category change request is 'pending review,' provided the status, and confirmed it is in the review queue and not blocked or rejected.

CriterionMet

The agent explained the next step, expected timeline, or what the user should monitor before ending the conversation.

Grader’s Explanation

The agent explained that the next step is owned by the category review team, that the user should watch for an email notification, and that no action is required from the user at this stage. The agent also recapped this at the end.

CriterionMet

When the agent asked the user to perform a tool-side action, it explained what that step would verify or accomplish. (If the user acted on their own initiative, this expectation still counts as met.)

Grader’s Explanation

The agent did not ask the user to perform any tool-side action; the user refreshed the request status page on their own initiative. The agent acknowledged this and confirmed what the status meant, which satisfies the expectation.

CriterionMet

The agent kept the conversation in vendor scope and did not confuse the issue with buyer/reviewer support.

Grader’s Explanation

The agent consistently addressed the user's request from the vendor perspective, focusing on the category change process and escalation steps relevant to vendors. There was no confusion with buyer or reviewer support.

Relevance 5/5

relevance: 5 — Agent immediately identified the user's need (status update on an in-flight category request), verified identity, pulled the exact request details (cat_001, pending, no block), explained ownership of next steps, addressed evidence requirements, and covered the rejection/dispute path — all directly relevant to the user's stated need with no detours.

Completeness 5/5

completeness: 5.00 — Derived deterministically from env/nl assertion coverage.

Accuracy 5/5

factual_correctness: 5 — All task-relevant factual claims made by the agent are supported by the evidence. The request ID (cat_001), current category (Customer Support), requested category (Help Desk), status (pending), submission date (2026-06-29), and absence of a decision date all match the tool result from `get_category_request_status`. The agent correctly stated that G2's review team owns the next step and that no action is required from the vendor while the request is pending. The agent accurately noted there is no specific SLA it can confirm. The recap in Turn 14 is consistent with the refreshed status (still pending, last updated June 29). No explicit factual claim is contradicted by the evidence.

Policy Compliance 5/5

policy_compliance: 5 — Agent followed all explicit policy rules for vendor scope category-change requests: identified vendor by work email before giving account-specific data, used phased approach (identity→tool lookup→state explanation→next-step guidance), explained pending status accurately without fabricating SLA or queue position, correctly noted G2's category review team owns next step, gave realistic contingency guidance for rejection without over-promising, closed with concrete monitoring instruction (watch email). No escalation was opened (correctly, since request is simply pending within normal review). No policy rule was violated.

Disputing a rejected review

Review moderation

Passed

The Scenario

Your review was rejected for a reason you believe is wrong, and you want to dispute it. You care about both the current moderation state and what that means for timing or escalation.

The Conversation

Simulated customer and Retell. Tool calls and results appear in sequence when available. Some details are omitted.

Retell

Hi! How can I help you today?

Simulated customer

My review, review_014, was rejected. Please look it up and tell me the exact rejection reason.

Retell

I'd be happy to look into that for you. Before I share any account-specific details, I need to verify your identity first. Could you please provide me with your full name or the email address associated with your G2 account?

Simulated customer

Jordan Lee.

Retell · Tool callfind_buyer_by_name()

Arguments

{}

Some fields are omitted from this public example.

Retell · Tool resultfind_buyer_by_name

Returned data

[
  {
    "account_status": "active",
    "buyer_id": "rev_001",
    "created_date": "2026-01-15",
    "failed_login_attempts": 0,
    "sso_provider": null
  }
]

Some fields are omitted from this public example.

Retell

I've verified your identity — you're Jordan Lee (rev_001).

Looking at review_014, here are the details:

Status: Rejected
Rejection date: June 18, 2026
Rejection reason: Unable to verify reviewer employment.

This means the moderation team could not confirm your employment at the company you listed when submitting the review.

If you'd like to dispute this decision, I can open a moderation dispute case on your behalf — but I want to make sure you understand that's the path before proceeding. Would you like to dispute the rejection?

Simulated customer

My LinkedIn is now verified, and I want to dispute the decision. Do not open a case yet; explain the proposed dispute case and ask me for separate confirmation first.

Retell

Understood. Here's what the dispute case would look like:

Type: Review moderation dispute
Subject: Rejection dispute for review_014 — employment verification
Details: Your review was rejected because employment could not be verified. You're now stating your LinkedIn is verified, which could serve as supporting evidence for the dispute. The case would be submitted to the moderation team for re-evaluation.

Please note: opening a dispute does not guarantee the review will be approved — the moderation team will make that determination independently.

Would you like me to go ahead and open this dispute case?

Simulated customer

Yes, please open the dispute case.

Retell · Tool callcreate_support_case()

Arguments

{
  "buyer_id": "rev_001",
  "type": "review_moderation",
  "related_review_id": "review_014"
}

Some fields are omitted from this public example.

Retell · Tool resultcreate_support_case

Returned data

{
  "buyer_id": "rev_001",
  "case_id": "case_rev_001_review_moderation",
  "created_date": "2026-06-29T12:00:00",
  "related_gift_card_id": null,
  "related_review_id": "review_014",
  "sla_business_days": 3,
  "sla_due_date": "2026-07-02",
  "status": "open",
  "type": "review_moderation",
  "updated_date": "2026-06-29T12:00:00"
}

Some fields are omitted from this public example.

Retell

The dispute case has been opened. Here's a summary:

Case ID: case_rev_001_review_moderation
Status: Open
SLA: The moderation team will review within 3 business days (by July 2, 2026)

The team will re-evaluate your review with your LinkedIn verification in mind. You should watch your email for any updates or follow-up requests from the moderation team. Note that I cannot promise a specific outcome — the decision rests with the moderation team.

Simulated customer

Thanks. Is there any escalation or exception if it is not resolved by July 2, or should I simply reply to the case update email?

Retell

If the case isn't resolved by the SLA date of July 2, 2026, the best next step is to reply directly to the case update email you'll receive. That keeps everything tied to your existing case (case_rev_001_review_moderation) and ensures the moderation team has full context.

There's no separate escalation path you need to initiate proactively — replying to that email is the correct way to follow up if the deadline passes without resolution.

Simulated customer

Understood. I’ll reply to the case update email if it’s still unresolved after July 2. Thank you.

Retell

You're all set! To recap:

Dispute case (case_rev_001_review_moderation) is open and assigned to the moderation team.
SLA: July 2, 2026 — if you haven't heard back by then, simply reply to the case update email to follow up.

Good luck with the dispute, Jordan! Let me know if there's anything else I can help with.

How It Was Graded

Task outcome and response quality are scored separately. A passed task can still contain mistakes.

Relevance: 5/5
Completeness: 5/5
Accuracy: 5/5
Policy Compliance: 5/5

View Grading Evidence

CriterionMet

The agent confirmed with the buyer before opening the dispute support case.

Grader’s Explanation

After the user requested an explanation of the dispute process and asked not to open a case yet, the agent explained what the dispute case would entail and then explicitly asked for confirmation before proceeding. Only after the user gave clear consent did the agent open the dispute case.

CriterionMet

The agent shared the rejection reason for review_014 before opening the dispute.

Grader’s Explanation

The agent provided the rejection reason for review_014 ('Unable to verify reviewer employment') before discussing or opening the dispute case. This was done in the agent's first substantive response after verifying the user's identity.

CriterionMet

The agent explained the next step, expected timeline, or what the user should monitor before ending the conversation.

Grader’s Explanation

The agent explained that the moderation team would review the dispute within 3 business days (by July 2, 2026), and advised the user to monitor their email for updates or follow-up requests. When asked about escalation, the agent clarified that replying to the case update email is the correct follow-up action if the deadline passes.

CriterionMet

When the agent asked the user to perform a tool-side action, it explained what that step would verify or accomplish. (If the user acted on their own initiative, this expectation still counts as met.)

Grader’s Explanation

The user initiated the request to dispute the decision and provided the relevant information (LinkedIn now verified). The agent explained what the dispute case would accomplish (re-evaluation of the moderation decision with the new LinkedIn verification as supporting evidence). Thus, the expectation is met.

Relevance 5/5

relevance: 5 — Agent stayed tightly focused on the user's actual need throughout: looked up the rejection reason for review_014, explained it, walked through the dispute case details before opening (per user's explicit instruction), opened the case, provided SLA, and addressed the escalation/follow-up question. No detours or irrelevant paths.

Completeness 5/5

completeness: 5.00 — Derived deterministically from env/nl assertion coverage.

Accuracy 5/5

factual_correctness: 5 — All agent factual claims are supported by evidence. review_014 status is rejected with rejection_reason 'Unable to verify reviewer employment' (matches final_state_snapshot). Case ID case_rev_001_review_moderation was successfully created (confirmed by tool result). SLA 3 business days with due date 2026-07-02 matches tool result. The agent correctly did not promise review approval. No material factual claims are contradicted by the evidence.

Policy Compliance 5/5

policy_compliance: 5 — The agent followed all explicit policy rules. Identity verified before giving account-specific info (Rule 1). Explained rejection reason before proposing dispute (§4.2). When user explicitly asked the agent not to open a case yet and to explain the proposed case first, the agent complied—explained the dispute case in Turn 8, asked for confirmation, received explicit 'Yes' in Turn 9, then called create_support_case in Turn 10. This satisfies the §4.2 rule: 'First explain the recorded rejection reason and proposed dispute case, then explicitly ask whether to open that case and wait for an affirmative response in a later user turn before calling create_support_case.' No promises of approval made. Closed with concrete next steps and SLA. No policy rules violated.

Estimated autonomy

Where this agent sits on the L1–L5 autonomy scale.

L2-L3

L1L2L3L4L5

Skills

The five shared skills reviewed for products in this category.

Supported

Multilingual Support

Supported

Case Routing & Escalation

Partial support

Grounded Answers

Partial support

Policy Compliance

Supported

Refunds & Billing

Performance claims

Vendor-reported

Cost savings / Staff replacement

Replaced 50% of after-hours staff and saved $480K annually

View source

Reduce Average Handle Time (AHT)

Cut AHT from 6 to 3.8 minutes

View source

Fast ROI / Breakeven

Breakeven in 60-90 days

View source

Frontier skills

Emerging capabilities selected for this evaluation category.

No frontier skills captured.

What G2 reviewers say

AI-generated from verified G2 buyer reviews.

Pros

Reviewers like the platform's user-friendly interface, the quality of the voice agents, the ability to customize calls, and the platform's overall efficiency in automating tasks.

Cons

Reviewers mentioned issues with the platform's high costs, especially for smaller projects or startups, occasional technical difficulties, and limitations in customization for more advanced or niche use cases.

Evaluation summary

How Retell performed

Recovering Google sign-in access

The Scenario

The Conversation

How It Was Graded

Relevance 5/5

Completeness 5/5

Accuracy 5/5

Policy Compliance 5/5

Checking a category change request

The Scenario

The Conversation

How It Was Graded

Relevance 5/5

Completeness 5/5

Accuracy 5/5

Policy Compliance 5/5

Disputing a rejected review

The Scenario

The Conversation

How It Was Graded

Relevance 5/5

Completeness 5/5

Accuracy 5/5

Policy Compliance 5/5

Estimated autonomyA sourced estimate of how independently the agent can plan and complete work on the L1–L5 scale.

Skills

Multilingual Support

Case Routing & Escalation

Grounded Answers

Policy Compliance

Refunds & Billing

Performance claims

Cost savings / Staff replacement

Reduce Average Handle Time (AHT)

Fast ROI / Breakeven

Frontier skills

What G2 reviewers say

Pros

Cons

Estimated autonomy