Comprehensive FAQ
Frequently asked questions
Extremely detailed answers about Freya, approvals, audit, roles, security, integrations, pricing, ethics, and Antarious AI for small businesses and commercial GTM teams. Use search to jump straight to the answer.
- Searchable
- 16 sections
- Freya
- Trust
- Business & GTM
- Sections
- 16
- Questions
- 86
- Human approval
- 100%
Section 01
Understanding Antarious AI
What is Antarious AI?
Antarious AI is an agentic operating platform for small businesses and commercial go-to-market teams that need to run marketing, sales, CRM, messaging, and money work at speed without surrendering human accountability. At the heart of Antarious is Freya — an autonomous AI Manager that operates as a persistent intelligence and execution layer for your business. Freya monitors connected activity, surfaces problems and opportunities, prepares multi-step commercial workflows, and retains context across decisions — while requiring human approval before any external action is taken. Unlike productivity tools that only help individuals type faster, Antarious AI is designed to change how a lean team operates: connecting channels, drafting the busywork, and producing the intelligence better decisions require.
How is Antarious different from a chatbot or AI assistant?
The difference is fundamental, not cosmetic. A chatbot waits for a human to ask something, answers, and stops. It has little durable memory of how your shop or GTM motion actually runs, and it does not initiate commercial work on its own. Freya is autonomous within governance. She does not wait to be asked before preparing drafts. She monitors connected operations, identifies follow-ups and opportunities, prepares posts, replies, pipeline next steps, and money reminders, and executes complex sequences — halting for human approval before anything leaves your business. Practically: a chatbot is a tool you use. Freya is a layer of operational help that runs alongside your business whether or not someone is actively prompting.
How is Antarious different from a copilot or productivity AI?
Copilot tools assist humans while they work — suggesting text, summarising documents, answering questions inside an existing task. They are human-initiated and human-sustained. Antarious AI is designed to run commercial busywork between human decision points: not merely to assist while you type, but to prepare the work that would otherwise consume your afternoon. Your role is to direct, approve, and decide — not to participate in every drafting step. A copilot makes each task slightly faster. Freya removes large categories of routine preparation from your plate entirely.
What does “agentic AI” mean here?
Agentic AI refers to systems that can take sequences of actions toward a goal without a human directing every micro-step. A non-agentic AI answers a question. An agentic system can receive an objective — draft this week’s content set, prepare outreach for these interested people, assemble invoice reminders from overdue balances — and execute the preparation chain: gather context, draft, sequence, handle exceptions. Antarious adds a critical constraint: every agentic sequence that concludes with an external action requires human approval before execution. Autonomy and control are not in tension; they are the design.
What kinds of organisations use Antarious AI?
Antarious AI on this site is built for small businesses and commercial GTM teams — shops, cafes, salons, studios, trades, local services, agencies, and revenue teams running marketing, sales, SDR-style follow-up, CRM hygiene, messaging, and money collection. The common thread is operational load: multiple channels, multiple follow-ups, multiple stakeholders, and the need for governance over automated preparation.
Is Antarious a SaaS product or a services engagement?
Access is product-led SaaS — many small businesses can start free, connect channels, and begin approving Freya drafts quickly. Larger commercial GTM deployments may still benefit from structured onboarding: connecting sources, mapping who approves what, and training Freya on brand and pipeline context. Think of it as a platform you run daily, with optional structured implementation when complexity warrants it — not as a pure consultancy that disappears after a slide deck.
What is the relationship between Antarious and Freya?
Antarious AI is the company and the platform. Freya is the autonomous agent at the centre — the persistent intelligence layer your team interacts with, and the orchestrator of specialist commercial agents. Antarious is the operating system. Freya is the teammate that runs it, the intelligence that powers it, and the name your team will use every day.
Can Antarious replace our existing team?
No — and that is deliberate. Antarious is built so AI amplifies human capability rather than replacing human judgement. The work that matters most in a small business or GTM team — relationships, creative direction, negotiation, pricing judgement, brand taste, accountability to customers — stays human. Freya handles preparation that does not require that judgement so your team can spend time on what does. The outcome is not a hollowed-out roster; it is the same team producing more with better context and less routine grind.
Section 02
Freya the autonomous agent
What does Freya actually do?
Freya performs four core functions for businesses. (1) Continuous monitoring of connected commercial activity — posts performance signals, inbox volume, pipeline movement, overdue money, anomalies worth a look. (2) Proactive surfacing on Home and related views — not raw noise, but packaged context about what needs a human. (3) End-to-end preparation of workflows — drafting content, enriching people records, preparing outreach, assembling reminders, routing approvals. (4) Durable business memory — retaining decisions, brand guidance, and outcomes so context does not vanish when a teammate is busy or leaves a chat thread.
How does Freya know what to monitor and when to alert?
Monitoring is configured as you onboard: which channels matter, which metrics are worth interrupting a human for, who receives which alert types, and what context should accompany each alert. Once set, Freya monitors continuously based on conditions rather than waiting for you to remember to check five apps. Sensitivity can be tuned — too noisy on a metric, raise the threshold; missing a class of problem, add it. Alerts still respect approval rules when an alert implies an external action.
What is the morning brief / Home digest?
Home (and related brief-style digests) is Freya’s proactive summary of what needs you: key commercial signals, anomalies from recent activity, actions pending approval, decisions due soon, and recommended priorities. It is generated from live context, not a static template. You can work the queue item by item — approve, edit, revise, reject — and ask follow-up questions in natural language within your permission scope.
How does Freya handle natural language questions?
Freya interprets questions against connected business context and your permissions, then returns an analysed answer with the figures and sources she used where available. Example: “Which deals stalled more than seven days and what follow-up is drafted?” She will not answer from data outside your role’s access scope. Natural language is a control surface — not a bypass around approvals for sending.
What happens if Freya gets something wrong?
Freya can draft incorrect, incomplete, or off-brand output. That is why human review exists before external action. Catch it in approval: reject, request revision, or edit before approving. Incorrect outputs and the reasoning presented to you are part of the recorded trail. If errors reflect bad source data, fix the source. If they reflect configuration — wrong tone rules, wrong approver routing — adjust configuration. Pattern failures should be raised with support@antarious.com so root causes can be investigated.
Does Freya work 24/7?
Yes. Monitoring, analysis, and draft preparation continue around the clock. Actions that require human approval queue until an authorised person acts — Freya does not self-approve external sends. When you return, the work is waiting with context rather than having silently published overnight.
Can I edit what Freya writes?
Always. Approve as-is, edit first, request a revision with instructions, or reject. Editing before approval is a first-class path, not an awkward workaround.
Does Freya send without me?
No. External actions require human approval. That is structural, not optional. Internal preparation can proceed; customer-facing and money-moving execution waits.
Section 03
Agent architecture (business specialists)
What are agents, and how do they relate to Freya?
Freya is the central intelligence layer — orchestrator, memory, and interface. Agents are specialist functions she directs. One specialist drafts content; another interprets analytics; another prepares outreach; another keeps CRM hygiene; another prepares money artefacts. When a workflow needs multiple skills — for example, turn campaign interest into enriched people records, draft follow-ups, and schedule sends after approval — Freya sequences the specialists and passes context between them so the chain stays coherent.
How many agents are there for business / GTM?
Business and GTM deployments use a commercial specialist set — commonly described as up to around a dozen focused agents spanning strategy, content, leads, outreach, analytics, optimisation, reporting, alerts, CRM, ads monitoring, SEO/content intelligence, pipeline, and agency client reporting where relevant. Not every specialist is active in every workspace. Small businesses typically start with the product surfaces that matter (Posts, Messages, Customers, Money, Home) and deepen specialists as confidence grows.
What do the core business specialists cover?
In commercial terms: strategy and campaign planning support; content generation and messaging variants; lead sourcing/qualification support against your ICP; outreach personalisation across allowed channels; analytics and attribution views; optimisation from performance; reporting drafts; KPI alerts; CRM cleanliness and enrichment; ads performance monitoring; SEO/content opportunity spotting; pipeline risk and stalled-deal flags; and, for agencies, client-facing reporting helpers. Exact availability depends on your plan and configuration.
Do agents share information with each other?
Yes. Context sharing is fundamental. Lead analysis informs outreach personalisation; outreach outcomes inform analytics; deal progression informs money drafts. Freya manages the context layer so specialists receive what they need without turning your business into disconnected micro-tools.
Section 04
Human approval and governance
What requires human approval in Antarious AI?
Human approval is required for external actions — anything with consequence outside the platform. That includes customer messages sent, posts published, outreach sequences launched, invoice reminders released, and similar customer-facing or money-moving executions. Internal operations — monitoring, draft preparation, organisation inside the app, anomaly detection — can run without an approval click because they create no external consequence by themselves.
What happens if the approver does not respond?
Antarious holds the action pending. It does not auto-execute on silence. Freya can remind the approver through configured notifications, escalate to a secondary approver if you set that up, and otherwise keep the item held. Time-sensitive windows that expire are logged; Freya does not invent consent.
Can approval requirements be turned off for speed?
No. Approval for external actions is an architectural constraint. What you can do instead: clearer review packages, templates for routine work, delegated approvers for leave, and routing lower-stakes items to the right role so senior owners are not bottlenecks. Friction reduction never means removing the gate.
What does the approver see?
A practical review package: plain-language action summary; supporting context and reasoning; the exact proposed output; risk flags; decision options (approve / approve with edits / request revision / reject) with room for instructions; and deadline indicators when time matters. Routine packages are designed to be reviewable quickly because everything needed is in one place.
Can multiple people be required to approve a single action?
Yes where configured: sequential chains, parallel quorum-style approvals, or hybrids. Each stage is recorded. Rejection halts execution. This matters for agencies and multi-owner small businesses that separate brand voice approval from commercial or money approval.
Who is accountable when something goes wrong after approval?
The human who approved the action is accountable for authorising it. Freya prepares; the human decides. The approval record identifies who approved what, when, and with what package in view. Accountability in customer-facing commerce cannot rest with software — it rests with the person whose judgement released the action.
Section 05
Audit trail
What does the audit trail record?
Meaningful events generate durable records across layers such as: what triggered the work; Freya’s presented reasoning and sources; the review package shown to the approver; the human decision and commentary; the executed output; linked outcome signals where available; and exceptions such as rejects or failed executions. The point is reconstructability: you can answer what happened without archaeology in private chats.
Can audit records be deleted or edited?
No. The audit trail is write-once by architecture. New linked information can be appended; originals are not silently rewritten. That integrity is what makes the trail useful for owners, teammates, and formal reviews.
How long are audit records retained?
Default retention is aligned to common business record-keeping expectations (commonly multi-year, with seven years as a frequent baseline in the wider Antarious trust model). Exact retention for your workspace follows product configuration and legal requirements. Where multiple obligations apply, the longer period typically wins unless instructed otherwise.
How do I retrieve audit records?
Users with audit permission can filter by time, action type, approver, and related dimensions, and export in formats suitable for review (such as PDF/CSV/JSON depending on product capability). Temporary read-only access for an external reviewer can be scoped and expiry-bound where that feature is enabled — and that access is itself auditable.
Section 06
Role-based control
How does role-based access control work?
Each user has a role profile covering dimensions such as data access scope, ability to initiate workflows, approval authority by action type, how they may direct Freya, configuration rights, and audit/reporting access. Enforcement is meant to be real — not merely hiding buttons while still allowing queries.
Can a user grant themselves additional access?
No. Permission self-escalation is structurally prevented. Users cannot modify their own privilege profile into existence, and approval authority is separated from configuration authority in sensible deployments.
What happens when someone leaves or changes role?
Revoke or reshape access immediately in the workspace. Institutional and brand context Freya learned for the business remains with the organisation. The next person receives a role matched to their responsibilities and benefits from retained memory rather than starting from a blank cultural slate.
How do we handle approval authority during leave?
Configure formal delegation: deputy, scope, and end date. Delegation is recorded and should expire automatically. Deputies cannot exceed the delegated scope or silently re-delegate onward.
Section 07
Security and data protection
How is our data encrypted?
Data at rest uses strong encryption (AES-256 class controls). Data in transit uses modern TLS. Keys are managed through dedicated key management practices with rotation. Customer-managed key options may be available for organisations with specific sovereignty needs — ask during security review.
Where is our data stored?
Data residency is confirmed during onboarding or commercial planning. Business customers can discuss regional requirements as part of deployment planning.
Does Antarious use our data to train its AI models?
No. Customer data is not used to train shared Antarious models for other customers, and is not shared with third parties for model training. Processing is to provide the Services under your agreement and DPA.
What happens to our data when we end our contract?
You can export within a defined transition window. After that, Antarious securely deletes customer data from systems of record, with written confirmation. Audit records subject to legally mandated retention may remain for the required period, then are deleted.
What security certifications does Antarious hold?
The security programme aligns with widely recognised frameworks such as ISO 27001, SOC 2 Type II, and UK Cyber Essentials principles. Certification programmes may be in progress for specific attestations — request current status under NDA for due diligence via support@antarious.com.
Has Antarious been penetration tested?
Yes. Regular third-party penetration testing covers application and infrastructure layers. Redacted summaries can be shared with business customers under NDA.
What is the incident response process?
On confirmed incident: isolate affected systems quickly; assess scope; notify affected customers within 72 hours of confirmed detection; remediate and restore; share a post-incident review with affected customers.
Is a Data Processing Agreement available?
Yes. A DPA is part of customer onboarding prior to go-live for applicable relationships. It covers categories processed, purposes, sub-processors, security measures, breach notification, data-subject rights support, and deletion. Contact hello@antarious.com or support@antarious.com to request the standard DPA.
Section 08
Deployment and implementation
How long does deployment take?
For many small businesses, meaningful value begins within days: account, brand baseline, channel connect, first approvals. Mid-size commercial GTM deployments with deeper integrations and formal approval graphs often land in a multi-week structured onboarding (commonly on the order of several weeks depending on integration count and stakeholder availability). Larger multi-team rollouts take longer. The most common delay is internal unavailability for decisions — not Freya “needing another month to think.”
What does structured onboarding look like?
A practical sequence: discovery and mapping of channels, roles, and approval owners; configuration and integration; validation of critical workflows; onboarding/training and supervised go-live; then 30- and 90-day optimisation reviews. Small teams can compress this; they should not skip naming approvers or writing brand rules.
Do we need a large technical team?
You need someone who can approve channel connections and basic access decisions. You do not need to hand-build the agent graph yourself. Antarious implementation support or a certified partner can carry configuration for heavier commercial deployments.
Can we use an implementation partner?
Yes. Certified implementation partners can deliver deployments for clients. See the partnership section and /company/partnerships.
What ongoing support exists after go-live?
Help Centre, Documentation, email platform support, training materials, product updates without reinstall theatre, and scheduled review sessions for customers on those plans. Higher commercial tiers may include dedicated success coverage.
Section 09
Integrations
What systems can Antarious integrate with?
Commercial systems across CRM/sales, prospecting, marketing/ads, analytics, communication, productivity suites, and finance/ops connectors — plus REST APIs for custom needs. Examples in the wider ecosystem include common CRM and ads platforms, analytics warehouses/BI tools, and workplace chat. This product story emphasises channels small businesses and GTM teams actually use day to day (social, messaging, customer pipeline, money follow-ups), not public-sector case systems or field M&E toolchains.
How are integration credentials secured?
API keys, OAuth tokens, and service credentials are stored encrypted, not logged in plaintext, and scoped to processes that need them. Rotation schedules are configurable.
What if a connected system is unavailable?
Integration health is monitored. Dependent workflows pause; admins are alerted; work resumes when connectivity returns; pending processing is queued rather than silently dropped.
Can we build a custom integration?
Yes via the Antarious API — authentication, endpoints, schemas, and webhooks are documented in the Documentation library for technical leads and partners.
How does Antarious handle data quality issues?
Freya validates on ingest where configured. Missing fields, format mismatches, duplicates, and unexpected values can be flagged, logged, and either block progress or proceed with explicit gaps called out in outputs. Quality flags can appear on dashboards and alert data owners.
Section 10
Sector: Business and GTM / small business
What does Antarious replace in a GTM tech stack?
Many GTM teams run a fragmented pile of content tools, prospecting tools, sequencers, attribution add-ons, reporting assemblers, and alert utilities. Freya’s commercial specialists consolidate preparation across those jobs so context does not die at every handoff. Small businesses feel the same fragmentation as “five apps and a spreadsheet.” Antarious AI collapses the operating rhythm into Posts, Messages, Customers, Money, and Home — with approvals as the control plane.
How does an SDR / follow-up workflow work with Freya?
Typical flow: define ICP or offer signals → specialists help source/score interested people → enrichment context informs personalised outreach drafts → human approval before sending → sequenced follow-ups under rules → reply classification surfaces hot intent → briefing cards prepare the human for the call or close. Humans spend time on conversations and judgement; Freya spends cycles on list prep and draft cadences.
How does multi-touch attribution thinking show up?
Analytics specialists can evaluate journeys with multiple lenses (first touch, last touch, linear, time decay, and more data-driven views where configured) so leadership is not trapped in a single vanity model. For small businesses, the practical win is clearer “what actually created the customer” narratives tied to campaigns, posts, and inbox threads you already run.
How do agencies run multi-client work?
Agencies can separate client workspaces with isolation, distinct brand memory, separate approval graphs, and client reporting helpers. The operational goal is higher-quality output per operator without mixing Client A’s voice into Client B’s send queue.
What kind of small business is this for?
Shops, cafes, salons, studios, trades, local services, and similar teams that post, reply, follow up, and chase money themselves today. If your growth bottleneck is fragmented busywork rather than lack of a thirteenth analytics warehouse, you are in the audience.
Does it help with funding?
Operating on Antarious AI builds a live funding profile reflecting real activity for Antarious Finance when you choose to explore funding. It is an outcome of tidy commercial operations — not a separate donor-reporting product.
Is it free to start?
Yes. You can start free — no credit card required for that entry path — and see Freya draft from day one inside the product.
Section 11
Pricing and commercial
How is Antarious priced?
Pricing reflects scope: seats, active commercial capability, integrations depth, memory/scale needs, and support tier. Public pages emphasise starting free for small businesses; broader commercial proposals are tailored because deployments vary. Contact the Antarious team to discuss fit.
Is there a minimum contract term?
Self-serve entry can begin without enterprise paperwork theatre. Standard commercial agreements for larger deployments are commonly annual, with multi-year options and occasional shorter terms by arrangement.
Is implementation included?
Platform access and structured implementation are often scoped separately for heavier rollouts. Certified partners may deliver implementation commercially. Many SMBs onboard primarily through in-product setup plus Help Centre / Documentation.
Who do I talk to about commercial terms?
Email hello@antarious.com or use /company/contact. For partner-led deals, register through the partnership programme channels.
Section 12
Compliance and regulatory (business-relevant)
Is Antarious designed to support UK GDPR and EU GDPR?
Yes, with features that support controller obligations: human oversight for significant automated effects (Article 22 alignment by approve-before-execute structure), data minimisation in configured processing, support for data-subject rights workflows as documented in the DPA, accountability evidence via the audit trail, and a DPA executed for applicable customers. GDPR compliance remains the customer’s determination as controller; Antarious provides infrastructure and contractual processing terms.
How does Antarious align with the EU AI Act’s human oversight expectations?
Design aligns with logging, transparency, and human oversight themes: audit records for post-hoc review, visible reasoning/context on recommendations, approve-before-execute for external actions, and robustness aids such as validation and anomaly flags. Organisations should confirm their specific AI Act obligations with legal advisors; Antarious can provide technical documentation to support assessments.
What about ISO alignment notes?
Security and governance programmes align toward ISO 27001-style control thinking (access control, encryption, incident response, supplier management, auditability). Request current certification status and control mapping under NDA during procurement.
Where should commercial teams start on compliance?
Start with GDPR/AI Act/ISO-oriented questions above, plus the Security and Human Approval trust pages, for regulatory due diligence relevant to commercial operations. Confirm obligations with your own counsel.
Section 13
AI ethics
What is Antarious’s position on AI replacing human jobs?
AI should amplify human capability, not replace human judgement. Freya takes routine preparation — drafting, monitoring, reconciliation busywork — so humans keep strategy, relationships, negotiation, creativity, and accountability. The observable goal is capacity expansion for the same team, not hollow automation theatre.
How does Antarious address biased recommendations?
Outputs reflect your data and configuration, so historical bias in sources can appear in drafts or rankings. Mitigations include transparent reasoning, mandatory human review before external action, data-quality flags, and multi-source context where available. Teams should actively examine biased patterns in their own pipelines and correct configuration accordingly.
Is Antarious suitable for decisions that affect individuals?
Freya can support decisions that affect customers — pricing exceptions, outreach prioritisation, credit-like commercial judgements inside your own policies — by preparing recommendations. She does not replace the human decision. For high-impact individual outcomes, configure stricter review and clearer role ownership.
Does Freya take political positions for my brand?
Freya produces commercial drafts from your data and instructions. She should not be treated as a political actor. Sensitive public statements remain human-owned: approve carefully or reject. Brand exclusions belong in your baseline rules.
Section 14
Technical architecture (high level)
What AI models does Freya use?
Antarious does not treat model brand names as the customer contract. The platform orchestrates large language capabilities and specialised analytical components behind Freya’s behaviour: transparent review packages, consistent commercial output quality, and human gates on external actions.
Is Antarious cloud-based?
Primarily yes. Dedicated or private cloud options may be available for enterprise customers with stricter isolation needs. On-premises is not the standard SMB path; discuss dedicated cloud if you have hard network constraints.
What uptime commitments exist?
SLAs are specified in customer agreements by tier. The platform is engineered for high availability with redundant infrastructure patterns. Ask commercially if you need contractual numbers.
Is there an API?
Yes — for ingestion, output delivery, approval triggers, and audit retrieval patterns. See Documentation for API reference material.
Does Antarious support SSO?
Yes for organisations that need it — common enterprise IdPs and SAML 2.0-compatible providers. SSO is configured during deployment for teams that require it; many small businesses begin with standard account access.
Section 15
Support
What support is available after go-live?
Help Centre, Documentation, email support, training materials, continuous product updates, and scheduled reviews on eligible plans. Response targets typically prioritise critical platform issues first, then impaired features, then standard queries — exact SLAs depend on your commercial tier.
Is there a dedicated customer success manager?
Available for higher commercial / strategic customers. Standard workspaces work through email support and the resource library.
How are product updates delivered?
Continuously to the cloud product without reinstall rituals. Significant changes are communicated in-product and by email; release notes live with Documentation.
What if a feature requires configuration changes?
Updates aim to be backwards-compatible. Where a major change needs action, expect advance notice, migration notes, and support help — not a surprise broken approval graph.
How do I get help right now?
Browse /resources/help-center, read /resources/documentation, email support@antarious.com, or keep working in the app at https://app.antarious.com while you wait for a reply.
Section 16
Partnership programme
What types of partnerships does Antarious offer?
Referral partners introduce Antarious and earn commission without delivering implementation. Implementation partners deploy for clients after certification. Strategic partners build a deeper practice with volume commitments, enhanced terms, and roadmap input. Details live at /company/partnerships.
How do I become a certified implementation specialist?
Certification covers platform architecture, deployment configuration, business/GTM sector configuration, trust architecture, and client onboarding/change management, plus assessment. Curriculum hours are on the order of a multi-day professional programme with annual renewal.
How does deal registration work?
Partners register opportunities through partner channels. Attribution attaches to your organisation; commission follows successful deployments per agreement. You can typically register referrals once partner paperwork exists even before full implementation certification.
What territories are covered?
The partner programme is global without artificial geographic fences. Regional go-to-market expertise is welcome.
How do I apply?
Email partnerships@antarious.com or hello@antarious.com with your organisation, commercial focus, and the partnership type you want. Expect a response within about two business days. You can also start at /company/partnerships.
Next
Still need a human?
If your question is not answered above, use the Help Centre for operational blockers, Documentation for foundations, or contact the team for commercial and security reviews.