The Best Enterprise AI Agents for Telecom Customer Operations
In this article
- What a telecom care leader is actually buying
- The full ranking
- The five things a telecom contact center needs
- 1. Nexus
- 2. Cognigy
- 3. Kore.ai
- 4. Druid AI
- 5. Wonderful
- 6. Google Gemini Enterprise
- 7. Relevance AI
- The criteria, one at a time
- A buyer's guide for operators
- Shortlisting checklist
- Conclusion
- The platforms in turn
What a telecom care leader is actually buying
An operator's contact center never closes. Bills go out on a cycle, outages arrive without warning, and every handset launch brings a wave of setup calls. Whoever heads customer operations sits in the middle of that flow, judged on cost per contact, first contact resolution and satisfaction scores, and asked at every budget round whether AI agents can take a share of the volume.
The market answers with a crowd of products sold as enterprise AI agents. Some are contact center platforms that have added generative AI. Some are general agent builders. Some are cloud suites with an assistant for office staff. A few pair software with people who build the first agent inside an operator's own systems. From their home pages they are hard to compare, because each describes the problem in its own vocabulary.
This assessment uses one lens. What does a head of customer operations at an operator need to get an AI agent onto real subscriber contacts, wired into billing and network systems, under control, and paid for in a way finance can defend? Seven vendors are measured against that question: Nexus, Cognigy, Kore.ai, Druid AI, Wonderful, Google Gemini Enterprise and Relevance AI.
Wherever this guide states a fact about a vendor, a source number follows it, and the number points to a page the vendor itself publishes; every one was opened on 8 November 2026. Silence counts too: if a vendor's pages are quiet on something a care leader needs, you will read that here. Nexus publishes this site and appears in its ranking; the criteria and reasoning are laid out below so you can test them for yourself.
The full ranking
1. Nexus. A telco page, consultants who build alongside the care team, a first agent live by week 4 of a 3 month proof of concept, and a fee per role automated1,11,12. 2. Cognigy. A contact center specialist now owned by NiCE, with named CCaaS integrations, pre launch testing and model failover13,14,15,16. 3. Kore.ai. A telco page listing real time OSS and BSS integration, Forward Deployed Engineers and on premises hosting17,18,19,20. 4. Druid AI. Orchestration across specialist agents, with deep security controls and hosting out to the air gapped edge21,22,23. 5. Wonderful. On site deployment teams, a reported 91.5 percent containment rate and four ways to host24,25,26,27. 6. Google Gemini Enterprise. A governed assistant and agent platform for staff on Google Cloud, with one edition priced in public28,8. 7. Relevance AI. A no code agent builder with public prices and strong evaluation tools, written mainly for go to market teams29,30,31,32.
The five things a telecom contact center needs
Before the vendors, the job. Reading the seven vendors' pages against an operator's daily contact mix produced five requirements.
Reach into the systems behind each contact. The contacts this assessment had in mind, bills, outages and plan changes, start in billing, network or the customer record. An agent that cannot read a bill, see an outage or change a plan can answer general questions and little else.
People who build the first agent. Where the care organization has no engineers of its own, a vendor that puts people beside it for the first build takes on the step where such projects tend to stall.
Control over what the agent says and does. Contract changes, payments and vulnerable customers all carry rules. The care leader has to decide where the agent may improvise and where it must follow a fixed path, and IT needs an audit trail.
Hosting and certifications an operator's security review will pass. Subscriber data is sensitive, and some operators insist it stay inside their own estate.
A commercial model tied to outcomes. Where the budget for a contact center agent is argued against cost per contact, a price that tracks the work automated is easier to defend than a price that grows with seats.
The first two carry the most weight in the ranking, because they decide whether an agent reaches live callers at all.
1. Nexus
Nexus leads because it publishes a telco offer, a dated route to a first live agent and a commercial model built on the work automated.
On its telco page Nexus addresses European operators with autonomous AI agents11 and calls itself the AI infrastructure for telecom33. Billing, network and customer systems are connected from day one12, and those are the three systems where an operator's contacts begin. The same page reports more than a million customer interactions a quarter running through Nexus agents at one European operator34. Its operator result, from Orange Group, is $6M extra lifetime value1.
Delivery decides the ranking. Nexus's home page has an engineer build the first procedure with your team and show it working in your own tools35. The builders are experienced consultants, McKinsey is in the background of one Nexus founder, and an operator's first agent is answering callers in 4 weeks, partway through a 3 month proof of concept1. A flat KPI triggers a rescope36, and every agent has had its KPI since launch37, so a care leader can read the result in the same weekly pack as handle time. Once live, the agent belongs to the business team that runs the service38.
The price follows the same logic. Nexus charges per role automated, with neither day rates nor consultant hours on the bill, so an operator buys consultancy support on software terms1. An operator would start with a 3 month proof of concept against outcomes agreed in advance39 and could leave if those outcomes were missed40.
When an operator's security team runs its vendor review, Nexus can answer with SOC 2 Type 2, GDPR, ISO 27001 and, unusually on this list, ISO 420011. It claims more than 4,000 integrations41, and agents can be reached by phone, web, WhatsApp, email, Slack or Teams42.
Its weak spots are documentation. SSO, encryption standards, penetration testing and uptime terms go undescribed on the pages we opened, as do hosting choices such as on premises or a selectable region. Nexus also says it is early43. Put those items in the proof of concept plan.
2. Cognigy
Cognigy is built for the contact center. Its home page asks buyers to put AI agents into their contact centers44, and the company now sits inside NiCE as NiCE Cognigy13.
Its strength for an operator is how neatly it joins an existing floor. Cognigy names Amazon Connect, Avaya, Genesys and NiCE CXone among the platforms its agents work with, on voice and digital channels14, and promises no rip and replace45. Over 110 prebuilt tools and integrations reach CRM, ticketing and back office systems2. LLM reasoning and deterministic steps share one conversation46, so a team can let the agent talk freely about a bill and switch to a fixed path the moment a payment is involved47.
Two features suit a busy operator. Simulated conversations test agents before launch and after each change15, with back end responses mocked across success, degradation and error48. And if the main LLM or speech service slows or fails, traffic moves on its own to a backup16.
Its standards list covers GDPR, ISO 27001, CCPA, HIPAA, SOC 2 and PCI DSS49, and PII can be stripped before the model sees it50.
Cognigy ranks second because no operator appears among its named customers, its pages say nothing on who builds the first agents or how quickly, and it publishes no price.
3. Kore.ai
Of the competitors, Kore.ai speaks most directly to operators. Its telco page covers subscriber, business and employee experience17 and lists OSS and BSS integration in real time18. A leading US fiber and broadband provider features among its customer stories51.
Policy lives outside the model: the engine enforces compliance rules52, leaving agents free to reason within approved limits53. Each run can be traced through reasoning, tools, guardrails, handoffs and outcomes54, and versions can be rolled back55.
Kore.ai also puts engineers in the field. Forward Deployed Engineers come with the offer19, and Kore.ai will prove results on the customer's own case before a contract56. Hosting runs from multi tenant SaaS to on premises and a dedicated single tenant VPC57,20,58. Its Trust Center lists SOC 2 Type II, ISO/IEC 27001:2022 and PCI DSS59,60,61.
Kore.ai ranks third, just behind Cognigy. It names no contact center platforms it plugs into, publishes no price, and leaves the terms for its Forward Deployed Engineers off the pages we opened. Its website is also mid rebuild62.
4. Druid AI
Druid AI sells agentic orchestration built from AI agents and intelligent apps63. In a telecom flow, its Conductor could send the plan question to a knowledge agent, the order retry to a process agent and the caller update to a voice agent, then combine the three64. Cases with low confidence or a policy exception go to a person, reasoning trace attached65.
Security and hosting are Druid's strongest cards. Traffic moves over TLS 1.2 or later and stored data sits under AES-256, with customer managed keys for regulated deployments22. SSO works with Active Directory and Okta66, and audit data feeds a SIEM67. An operator can keep Druid in its own data center, in a hybrid setup or in the cloud, and the Becus model can run air gapped68; even isolated edge sites are supported23. On compliance, Druid names the EU AI Act next to ISO 27001:2022, GDPR and SOC 2 Type II69.
Its integration protocols reach old systems through SOAP, JDBC and ODBC, RPA bridges and MCP70, which helps in a telecom estate full of legacy back ends.
Druid ranks fourth because telecom is missing from the industries it lists71, we found no operator among its named customers, and the split of delivery work between Druid and its partners is not described. Prices are custom6.
5. Wonderful
Wonderful presents itself as the enterprise's AI OS72. It was founded in 20257 and has closed a $550 million Series C72.
How it delivers is close to how Nexus delivers. Wonderful sends deployment teams on site24, with forward deployed engineers and deployment strategists working inside the client's environment73, and plans a path from Wonderful leading to the client owning the work74. It reports containment of 91.5 percent across millions of customers25 and energy billing resolved from start to finish75.
An operator choosing where subscriber conversations live has four options with Wonderful, from shared multi tenant through dedicated single tenant and its own cloud account to an air gapped installation on premises26,76,77,27. PCI DSS sits on its compliance list with GDPR, ISO 27001:2022 and SOC 2 Type II78, and outside testers probe it every six months79. Budget caps and model access can be set per team in the AI Gateway80,81.
Wonderful ranks fifth because its pages name no operator, no connector count and no contact center platform integration, and it does not publish a price.
6. Google Gemini Enterprise
Gemini Enterprise is intranet search, an assistant and an agent platform in one product82. Its intended users are office functions such as finance, HR, IT, marketing and sales83; for an operator's advisers, the Frontline edition is the closer fit84.
An operator already on Google Cloud gets a mature security base: certifications from HIPAA and FedRAMP to ISO 27001, SOC 1, 2 and 3, and PCI DSS85, customer managed keys86, and data held in the us, eu or global location87. Google does not train its foundation models on prompts, responses or connected data88. ServiceNow, SharePoint, Jira and Confluence are among its prebuilt connectors89, and a Business seat costs from $21 a month8.
In a telecom contact center, its likeliest role is on the adviser side, helping staff find answers across documents and tickets. The pages we opened show no customer facing contact center deployment, no telecom offer and no implementation team, and Standard and Plus prices are not published90. With seat billing91, cost grows with headcount.
7. Relevance AI
Relevance AI is a platform for building agents and agent teams with little or no code29. Its documentation promises agents live within minutes from the builder or a template92, and its directory lists 2,728 integrations93.
On price it is the most open vendor here. Pro is $19 a month on annual billing30, Team $234 a month on annual billing94, and model usage passes through at wholesale with no markup95. Governance includes approval rules set per tool96, a human sign off on high stakes actions97 and SSO98. Every change re-runs the agent's evals before it publishes31, and the enterprise page commits to 99.9 percent uptime99.
For this buyer Relevance AI ranks seventh. Its pages are written for go to market and revenue teams32,100, name no operator or contact center deployment, and offer no on premises or customer VPC hosting. Support keeps Sydney weekday hours with nothing at weekends101, which a seven day contact center would have to negotiate.
The criteria, one at a time
Who builds, and how soon the first agent is live
Nexus publishes both who builds the first agent and the specific week it goes live: experienced consultants, and week 4 of a 3 month proof of concept1. Wonderful sends on site deployment teams24 and claims production within weeks102. Kore.ai provides Forward Deployed Engineers19 and promises proof to production in weeks103. Relevance AI describes an embedded deployment team104. Druid offers expert support from planning through optimization105, and one of its customers had a voice agent live in under six months106. Cognigy and Google describe tools and setup steps; neither describes a delivery team on the pages we opened.
Telecom fit
Only two vendors have written a page for operators. Nexus connects billing, network and customer systems from day one12; Kore.ai lists OSS and BSS integration in real time18. Kore.ai adds a US broadband provider story51, and Nexus reports over a million interactions a quarter at one European operator34. The remaining five name no operator on the pages we opened.
Contact center platforms and channels
Cognigy leads this criterion, naming Amazon Connect, Avaya, Genesys and NiCE CXone14 and shipping with more than 30 channels already connected107. Nexus agents reach customers by phone, web, WhatsApp and email, and staff in Slack and Teams42. Kore.ai covers voice, web, mobile, Slack and Teams108. Telephony is on Druid's compatibility list109. Relevance AI's Team plan brings calling agents110. Gemini Enterprise and Wonderful name no customer voice channel on the pages we opened.
Control and governance
Every vendor publishes governance features, in different shapes. Cognigy lets the customer decide case by case when generative AI is used111. Kore.ai keeps policy enforcement outside the model53. Druid shows the full reasoning chain for each interaction112. At Nexus, the gate before any agent reaches callers is a CIO's approval of three items, audit trail, access model and data flow113; once live, an unsure agent passes the case up with its context114. Wonderful tracks every change and can roll any of them back115. Relevance AI sets approval rules per tool96. Google governs every team's agents from one pane116.
Certifications
ISO 42001 is the item that sets Nexus apart here: it holds that standard with ISO 27001, GDPR and SOC 2 Type 21, and in the facts gathered for this site no rival lists it. For operators taking payments on a call, PCI DSS appears in the lists of Cognigy, Kore.ai, Wonderful and Google49,61,78,85. Druid names the EU AI Act alongside SOC 2 Type II and ISO 27001:202269. Relevance AI holds SOC 2 Type II and GDPR117,118.
Hosting and subscriber data
On premises hosting is offered by Druid, Wonderful and Kore.ai68,27,20, and Druid adds the air gapped edge23. Google lets customers choose a data location87. Relevance AI lists data residency119 but no on premises option. Cognigy describes a single SaaS platform120. Nexus's hosting choices go undescribed on the pages we opened.
How the price is built
Once its 3 month proof of concept is done121, Nexus bills an operator per role automated; consultant hours and day rates never reach the invoice1. Google charges per seat91; Relevance AI charges by plan and by Action30,122. Druid quotes custom prices6. Cognigy, Kore.ai and Wonderful publish none.
A buyer's guide for operators
Begin with one contact reason. Choose a single, high volume contact reason with clean data, bill explanations or plan changes for example, and judge every vendor on how it would handle that reason from first word to resolution. The description Nexus gives of its best fit, high volume work across several systems with a measurable financial result123, is a useful template for the choice.
List your systems by product name. Write down the billing platform, CRM, provisioning, OSS and BSS, and contact center platform. Integration counts vary widely, from more than 4,000 at Nexus41 to more than 110 tools at Cognigy2, and a count tells you nothing about one particular system. Ask each vendor to show a working connection to the systems on your list.
Settle who builds. If your care organization has no engineers, a vendor with people on the ground counts for more than a polished studio. Ask every vendor who builds the first agent, for how long, and how that time is charged. Nexus has published its answer1; with most of the others you will have to ask.
Agree the number before the build. Fix the KPI for the first agent, its baseline and the date it will be read. Of the vendors here, Nexus ties a KPI to every agent from launch37, and Wonderful's first step is a single use case in production124. Hold every vendor to the same discipline.
Test the handover to people. In telecom, a poor handover creates a second contact. See how each agent passes a caller to an adviser: Cognigy hands over a summary and the next step125, Druid sends low confidence cases with a reasoning trace65, and Nexus escalates with context114.
Bring security in on day one. Share hosting, key management and certification requirements at the first meeting. Some answers are already public, Druid's customer managed keys22 or Google's data locations87 for instance; others, such as Nexus's SSO and encryption details, need a request.
Run each price model against your volumes. A seat price91 grows with staff, an Action price126 with usage, and a price per role automated1 with the work taken on. Model all three on your own contact volumes before comparing quotes.
Shortlisting checklist
Put these to each vendor on the first call. A written answer is worth more than a demo.
1. Which of our billing, CRM, network and contact center systems will the agent read and write from day one? 2. Who builds the first agent, and how is their time charged? 3. On what date does the first agent take real customer contacts, and how long is the proof of concept? 4. Which contact center measure, containment, handle time or repeat contacts, will the first agent answer for, and what changes if it stays flat? 5. How does the agent hand a caller to a human, and what does the adviser see? 6. Where may the agent improvise, and where must it follow a fixed path? 7. Which certifications are held today, and may we read the SOC 2 report? 8. Where is subscriber data hosted and processed, and who holds the keys? 9. Is our data used to train any model? 10. What does a telecom reference customer say about you? 11. How does the price move with contact volume, seats or roles automated? 12. What are the uptime and support terms, weekends included?
Conclusion
For a head of customer operations at an operator, the hard part of an AI agent project is getting the first agent onto real contacts, connected to the systems that hold the answer, and proving a number. All seven vendors here ship capable software. Where they differ is who does that first piece of work, how fast, and on what terms.
Nexus ranks first because its answer is public and specific. Its telco offer connects billing, network and customer systems from day one12. Its consultants, experienced people working with the care team, have the first agent on live contacts by week 4 of a 3 month proof of concept, and the operator pays per role automated, with no consultant hours or day rates1. Its Orange result, $6M extra lifetime value1, came from an operator. The open items are documentation, chiefly hosting, SSO and encryption, and they belong in the proof of concept plan.
Cognigy and Kore.ai are strong contact center platforms: Cognigy ahead on named CCaaS integration and testing, Kore.ai on telecom focus and hosting choice. Druid AI and Wonderful lead on hosting control and security detail. Google Gemini Enterprise serves the advisers who work the queue, and Relevance AI offers go to market teams a priced, quick start.
Nexus publishes this site and appears in its ranking. Every claim above carries a source number, so check it before you rely on the order.
The platforms in turn
Nexus
Published by Nexus
What it is
Nexus is an enterprise AI agent platform with a delivery team. Its engineers build agents with your business teams, then stay to improve them against a result you choose. Nexus's first agent is live in 4 weeks, inside a 3 month proof of concept, and it has 4,000+ integrations. Nexus publishes this site.
Strengths and limits by criterion
Nexus engineers build with your business teams and stay to improve the agents, measured on the result the customer cares about.127,128
No limit stated on the pages we opened.
Nexus's first agent is live in 4 weeks, inside a 3 month proof of concept.1
No limit stated on the pages we opened.
Nexus has 4,000+ integrations across CRMs, ERPs, communication tools, legacy systems and custom APIs.129
No limit stated on the pages we opened.
Certifications: GDPR, ISO 27001, SOC 2 Type 2 and ISO 42001.130
No limit stated on the pages we opened.
Every agent interaction generates audit trails automatically, decision logic is traceable by design and role based access controls are built in.131
No limit stated on the pages we opened.
Agents deploy across Slack, Teams, WhatsApp, email, phone and web.133
Hosting options such as self hosted or VPC are not publicly stated on the Nexus pages we opened.
Pricing and buying
Per role automated1
Who it suits
Teams that want engineers alongside them and agents live fast.
Cognigy
Homepage describes the product as AI agents
What it is
Cognigy comes to this comparison as a specialist in the contact center. Its home page invites buyers to deploy AI agents in the contact center that deliver fast, personalized service, and almost everything else on the site follows from that focus.
Strengths and limits by criterion
Live debugging shows real interactions as they happen (vendor claim).134
No limit stated on the pages we opened.
Not publicly stated on the pages we opened.
No limit stated on the pages we opened.
Works with named contact center platforms (vendor claim).14
No limit stated on the pages we opened.
AI Ops page lists standards met, including ISO 27001 and PCI DSS (vendor claim).49
No limit stated on the pages we opened.
Full execution traceability and audit history across AI and human interactions (vendor claim).135
No limit stated on the pages we opened.
A free trial is offered; the privacy policy names it.3
Not publicly stated on the pages we opened: . The 15 pages fetched show no price, plan, per-user, per-seat, subscription or quote terms. Searched terms.
Voice Gateway runs on an enterprise-grade SaaS platform (vendor claim).136
No limit stated on the pages we opened.
Pricing and buying
A free trial is offered; the privacy policy names it3
Who it suits
Customer headline: Frontier Airlines automates 800k monthly conversations (vendor claim).
Kore.ai
Homepage title positions Kore.ai as an enterprise
What it is
If you run customer operations at an operator, Kore.ai deserves an early meeting for one plain reason: on the pages we opened, it is one of only two vendors in this comparison that has written a page for telcos. That page frames Kore.ai's telecom work around three groups, subscribers, business customers and employees.
Strengths and limits by criterion
Forward Deployed Engineers are part of the offer (vendor claim).19
No limit stated on the pages we opened.
Homepage promises results in weeks (vendor claim).137
No limit stated on the pages we opened.
Telco page lists real-time OSS/BSS integration (vendor claim).18
No limit stated on the pages we opened.
AI section lists monitoring and testing of AI output for bias and hallucination (vendor claim).138
No limit stated on the pages we opened.
Trust Center lists an enterprise service level agreement as a document (vendor claim).139
No limit stated on the pages we opened.
Not publicly stated on the pages we opened.
Not publicly stated on the pages we opened: . The 14 pages fetched show no price, plan, per-user, subscription, free-trial or quote terms for Kore.ai's.
Trust Center lists a multi-tenant SaaS deployment (vendor claim).57
No limit stated on the pages we opened.
Pricing and buying
Not publicly stated on the pages we opened
Who it suits
Homepage lists the industries served (vendor claim).
Druid AI
Homepage title positions Druid as agentic AI
What it is
Druid AI grew up in conversational AI and now leads its site with agentic orchestration, combining AI agents with what it calls intelligent apps. Its own description of the product is the AI platform enterprises run on, a single place to deploy, run and keep improving agents at scale.
Strengths and limits by criterion
Pricing page says customers get expert support from planning to optimization (vendor claim).105
No limit stated on the pages we opened.
Platform page says integrations go live in days without disrupting existing infrastructure (vendor claim).140
No limit stated on the pages we opened.
Integrations page says there are hundreds of ready-to-use connectors and filters (vendor claim).141
No limit stated on the pages we opened.
Platform page lists SOC 2 Type II, ISO 27001, GDPR, HIPAA and NHS compliance (vendor claim).142
No limit stated on the pages we opened.
Permissions apply for each agent, flow, knowledge segment and conversation thread (vendor claim).143
No limit stated on the pages we opened.
Privacy policy applies the data protection law of Romania and GDPR (Druid's privacy notice).144
No limit stated on the pages we opened.
Pricing and buying
Pricing is custom only (vendor statement)6
Who it suits
Industry segments listed on the platform page (vendor claim).
Wonderful
Homepage headline positions Wonderful around
What it is
Wonderful is a very young company backed by a very large round. A post on its blog reports closing $550 million in Series C funding at a valuation of $5B and presents Wonderful as the enterprise's AI OS. The pitch on its home page is speed of AI transformation across a whole enterprise.
Strengths and limits by criterion
Wonderful provides on-site deployment teams (vendor claim).24
No limit stated on the pages we opened.
The about page claims production in weeks (vendor claim).102
No limit stated on the pages we opened.
Deployment teams build agents, integrations and workflows into the customer's stack (vendor claim).145
No limit stated on the pages we opened.
AI OS page lists certifications and alignments (vendor claim).78
No limit stated on the pages we opened.
Access to build, edit and ship can be set down to the resource (vendor claim).146
Privacy policy makes no guarantee against unauthorized access.147
Not publicly stated on the pages we opened.
Not publicly stated on the pages we opened: . The 12 pages fetched show no price, plan, per-seat or subscription terms. Searched terms: pricing, per.
Multi-tenant option: continuous updates, monitored by Wonderful, with logical isolation (vendor claim).26
No limit stated on the pages we opened.
Pricing and buying
Not publicly stated on the pages we opened
Who it suits
Homepage says Wonderful is trusted by leading enterprises across industries (vendor claim).
Google Gemini Enterprise
Search offers permissions-aware access
What it is
Google Gemini Enterprise earns its place here for a different reason from the contact center specialists. Read Google's documentation with a contact center in mind and the product resolves into an intranet search engine, an assistant for staff and a place to run agents.
Strengths and limits by criterion
Licensing starts with buying a subscription for the billing account.148
No limit stated on the pages we opened.
Seat subscriptions can be monthly or year-long.149
No limit stated on the pages we opened.
Partner-built agents can be found in Agent Marketplace.150
Third-party connectors use public endpoints outside Google's network.151
Model Armor is included in all editions of the app.152
No limit stated on the pages we opened.
Agent tool-use rules are written in plain language.153
No limit stated on the pages we opened.
Business edition is for small businesses and teams up to 300.154
Pricing and buying
Business edition starts at $21 per seat per month8
Who it suits
Frontline edition is for frontline workers.
Relevance AI
Describes itself as a low/no-code platform
What it is
Relevance AI sells a platform on which teams with little or no code build single agents and whole teams of them. Its home page brands the product the home of the AI Workforce, and the product page compresses the promise into building, running and managing agents at scale.
Strengths and limits by criterion
Agents can be deployed in minutes with the no-code builder or pre-built templates.92
No limit stated on the pages we opened.
Vendor says some use cases can take as little as a few hours to stand up (vendor claim).159
No limit stated on the pages we opened.
Integrations page headline: 2,000+ integrations, from Slack to Salesforce.9
No limit stated on the pages we opened.
Documentation FAQ says Relevance AI is SOC 2 Type II compliant.117
No limit stated on the pages we opened.
SSO / SAML is listed as an access control.98
No limit stated on the pages we opened.
Team plan's list adds calling and meeting agents on top of Pro.110
Data residency is listed among the enterprise controls.119
Not publicly stated on the pages we opened: on-premises, customer-VPC or self-hosted deployment of the Relevance platform. The pages checked name no such.
Pricing and buying
Pro plan starts at $19 per month10
Who it suits
Use cases listed on the homepage include Human Resources, alongside Sales, Customer Success.