Customer service AI

Brand-trained AI chat. Keystone escalation. CAD-priced.

Concierge is brand-voice-trained AI customer service - the first-touch chat handler that knows your products, your brand voice, your return policy, your shipping, and your FAQ. Anything it can't resolve escalates to Keystone with full transcript attached. Multi-storefront SMBs get per-brand voice routing at Pro+.

Concierge Starter at $29/mo CAD covers single-storefront SMBs with brand-voice chat, 24/7 first-touch handling, transcript export. Pro at $59/mo unlocks multi-storefront, sentiment analysis, live agent override, knowledge base integration, and Keystone-routed escalation. Enterprise at $159/mo adds unlimited storefronts and custom brand corpus.

Built for the Canadian SMB customer service operator who's tired of either paying $135-$2,500/mo CAD for the popular enterprise chat AI or running an unstaffed shared inbox. Concierge sits in the middle: brand-voice AI handles 60-80% of inbound questions; Keystone catches the rest.

15-day free trial - No credit card - CAD-priced
The problem

Why SMB customer service AI either fails or breaks the budget

The popular generic SMB chat widget is $30-$60/mo CAD with canned-response templates - your customer gets generic answers that don't sound like your brand, and complex questions fail entirely. The popular AI-powered chat platform is $135-$2,500/mo CAD per storefront, billed in USD, with most useful features (brand voice training, custom corpus, escalation routing) at the upper end of that range.

Multi-storefront SMBs are charged per-storefront. Run 3 brands and your chat costs $400-$7,500/mo CAD. Brand voice per storefront is rare - you usually end up with one shared corpus across all your stores, which sounds wrong on every single one.

Escalation routing is the other gap. The popular AI chat tools either stop at the AI or escalate into a generic email queue. Routing into a proper ticketing system (Keystone) with full transcript attached, escalation policy, SLA clock, and asset linking simply doesn't exist as an integrated SMB-tier offering outside Vastura.

Concierge is positioned at the gap: $29-$159/mo CAD, brand-voice trained per storefront, escalates to Keystone with full context. The unit economics work because Concierge runs on Anthropic Haiku 4.5 inference at fractional cost per conversation, and the tier caps prevent unbounded inference cost from eating the operator's margin.

The Vastura solution

Brand-voice AI chat + Keystone escalation + per-storefront voice routing.

Concierge Starter at $29/mo CAD is single-storefront, brand-voice chat (corpus pulled from Studio Brand Kit or Brand Lab if present), 24/7 first-touch handling, transcript export (PDF + JSON), email support. Entry tier for the small operator wanting a smart first-touch handler without enterprise pricing.

Concierge Pro at $59/mo CAD unlocks multi-storefront with per-brand voice routing (each storefront gets its own brand-correct chat), sentiment analysis (real-time customer-sentiment scoring during the conversation), live agent override (the agent can take over a chat mid-conversation), knowledge base integration (Concierge searches your KB articles for answers), and Keystone-routed escalation (chats Concierge can't resolve route into your Keystone queue with full transcript and customer record).

Concierge Enterprise at $159/mo CAD adds unlimited storefronts (no per-brand cap), custom brand corpus (upload your past customer service transcripts; we train a custom corpus per brand for higher resolution rates), custom integrations, and dedicated CSM. Built for the multi-storefront agency or holdco running customer service at scale.

Architecturally, Concierge is the customer-facing AI in the Vastura stack. It runs on Haiku 4.5 inference with per-brand context loaded per conversation. Cost basis is fractional per conversation; tier pricing caps your monthly spend. Pairs naturally with Launch Pad (storefront context auto-attached), Keystone (escalation), Marketing Suite (lead handoff).

Feature deep-dive

Everything you need, built into one platform

Eight customer-service AI modules - brand voice, sentiment, escalation, KB integration - priced for the Canadian SMB.

B

Brand-Voice-Trained AI Chat

Concierge pulls brand voice corpus from Studio Brand Kit (or Brand Lab if present) and responds in your brand's voice. Multi-storefront SMBs at Pro+ get per-brand voice routing - same Concierge backend, different voice per storefront.

2

24/7 First-Touch Handling

Concierge handles inbound chat 24/7 - no staffing required. Most SMBs report 60-80% first-touch resolution rate. Customer gets an answer in seconds, not hours, and only the genuinely complex queries escalate to a human.

K

Keystone-Routed Escalation (Pro+)

Pro+ tier. Chats that Concierge can't resolve route into Keystone with full conversation transcript, customer record, storefront context, and any attached data. SLA clock starts; the right agent picks up where Concierge left off without re-asking the customer.

T

Transcript Export (PDF + JSON)

Every Concierge conversation is exportable as PDF (for customer-facing review) or JSON (for analytics and training). Per-storefront archive at Pro+. Useful for sentiment trending, common-question identification, and KB gap analysis.

S

Sentiment Analysis (Pro+)

Pro+. Real-time customer-sentiment scoring during the conversation. Drops in sentiment trigger live-agent-takeover alerts. Sentiment trending dashboards show which storefronts have happy customers and which need product-or-policy attention.

L

Live Agent Override (Pro+)

Pro+. A human agent can take over a chat mid-conversation - Concierge steps aside and the human handles. Useful for high-stakes conversations (large refund, complex customization, escalating complaint). Agent sees full transcript and customer context.

K

Knowledge Base Integration (Pro+)

Pro+. Concierge searches your knowledge base (in Keystone or Launch Pad) for answers to customer questions. Article references included in the chat response. Helps the customer find self-service answers without leaving chat.

C

Custom Brand Corpus (Enterprise)

Enterprise. Upload your past customer service transcripts; we train a custom corpus per brand for higher first-touch resolution rates. Multi-storefront agencies get distinct custom corpora per client brand.

Compared head-to-head

Vastura vs the industry SMB average

Compared at $30-$2,500/mo in the SMB AI customer chat category - Anonymized industry-average benchmarks - CAD-priced throughout - No FX surcharge on Vastura tiers
FeatureVasturaIndustry SMB average
Brand-voice-trained AI chat from Starterpartial
24/7 first-touch handling from Starter
Keystone-routed escalation Pro+×
Transcript export PDF + JSON from Starter
Sentiment analysis Pro+ Pro+
Live agent override (mid-chat) Pro+ Pro+
Knowledge base integration Pro+
Multi-storefront support Enterprise (unlimited)partial
Per-storefront brand voice routing Pro+×
CAD-priced (no FX surcharge)×
Custom brand corpus training Enterprisepartial Enterprise-only typically
Cost-controlled inference Haiku 4.5 cost basispartial flat-fee unbounded common
Tier-priced (not per-conversation)partial
Launch Pad context auto-attached×
Marketing Suite lead handoff×
Audit log on every conversationpartial
Multi-language support (EN/FR)

Concierge sits in the under-served middle of the SMB AI customer service market. Cheaper than the popular enterprise AI chat platforms ($135-$2,500/mo CAD), smarter than the popular template-based chat widgets ($30-$60/mo CAD), with the unique Keystone escalation routing that closes the loop from AI chat to ticketing.

If you run a single storefront and want a smart first-touch handler, Starter at $29/mo CAD is the right tier. If you run multiple storefronts and need per-brand voice routing, Pro at $59 is the right tier. If you're an agency or holdco with 5+ brands, Enterprise at $159 is the only SMB-tier subscription with unlimited storefront support + custom corpus.

Pricing

Choose the tier that fits your operation

Every tier includes a 15-day free trial. Cancel anytime. Annual billing saves 25%. All prices in Canadian dollars.

Starter

Single storefront wanting a first-touch chat handler

$29
/month CAD
or $261/yr billed annually (save 25%)
  • Single storefront
  • Brand-voice chat
  • 24/7 first-touch handling
  • Transcript export (PDF + JSON)
  • Email support
Start 15-day free trial

Enterprise

Multi-storefront agency with Keystone-routed escalation

$159
/month CAD
or $1,431/yr billed annually (save 25%)
  • Pro + unlimited storefronts
  • Custom brand corpus per storefront
  • Custom integrations
  • Dedicated CSM
Start 15-day free trial

15-day free trial - Cancel anytime - CAD-priced - No FX surcharge

Use cases

Built for real Canadian SMB operators

Four personas we built this product for. Find yours below.

Single Storefront

Solo Shopify operator

Pat runs one Shopify store. Starter at $29/mo gives her brand-voice chat 24/7, transcript export, basic first-touch handling. Resolves 60-70% of inbound; the rest go to her support email.

Best fit: Concierge Starter

Multi-Storefront SMB

Three-brand operator (Vastura-style)

Marcus runs Ag, Ltd, Digital. Pro at $59/mo gives per-brand voice routing - each storefront has its own Concierge voice. Sentiment analysis flags angry customers; live agent override lets Marcus jump in. Keystone catches anything escalating.

Best fit: Concierge Pro

Agency

Multi-brand commerce agency

Lisa's agency runs 8 client brands. Enterprise at $159/mo handles unlimited storefronts, custom corpus per client brand, dedicated CSM. Throughput per support agent up 3x; first-response time under 30 seconds across all 8 brands.

Best fit: Concierge Enterprise

Holdco

Multi-entity ecommerce holdco

Devon's holdco markets 5 brands. Enterprise gives unlimited storefronts, custom corpus per brand, per-brand voice routing. Conductor cron pulls overnight sentiment digest; LP Assistant routes follow-ups.

Best fit: Concierge Enterprise + Conductor

Integrates with the Vastura stack

How this fits with the rest of your operation

Vastura is a stack, not a single tool. Each product passes data into the next so you stop running glue scripts and CSV exports.

Launch Pad - storefront context auto-attached

Concierge running on an LP storefront automatically sees the product, order, customer, and brand context. Per-brand voice routing handled.

LP storefront chat -> Concierge -> Brand-correct response
Keystone - escalation with full transcript

Chats Concierge can't resolve route into Keystone with full transcript, customer record, storefront context. SLA clock starts. Agent picks up without re-asking the customer.

Concierge chat -> Escalation -> Keystone ticket
Studio - brand voice corpus

Studio Brand Kit voice corpus trains Concierge per brand. Updated brand voice in Studio flows immediately to Concierge conversations.

Studio voice -> Concierge per-brand -> Branded chat
Marketing Suite - lead handoff

Marketing Suite leads (email captures, social DMs) hand off to Concierge for first-touch qualifying chat. Brand voice preserved.

Marketing lead -> Concierge chat -> Qualified -> CRM
Ledger - refund / billing chat -> JE

Concierge chats that result in refunds or billing adjustments feed into Ledger as the proper journal entry (customer credit memo, refund payable). Per-entity routing.

Concierge refund decision -> Ledger Refund JE -> Customer credit
Conductor - overnight sentiment digest

Conductor schedules an overnight Concierge sentiment digest - which storefronts trended down today, which got escalated to Keystone, which need product attention.

Conductor cron -> Concierge analytics -> Morning digest
Inventory - stock queries

Concierge can query Inventory for real-time stock levels per SKU. 'When will this be back in stock?' is answered with current Inventory data.

Customer query -> Concierge -> Inventory data -> Reply
Pilot - service-product fulfillment status

If your storefront sells service products, Concierge queries Pilot for the project status. Customer asks 'where's my deliverable?'; Concierge returns current Pilot status.

Customer query -> Concierge -> Pilot status -> Reply
Frequently asked

Questions before you start the trial

How long is the free trial and is a credit card required?

Every Vastura tier includes a 15-day free trial. No credit card is required up front - you only enter payment when you decide to keep the subscription past day 15. If you cancel within the trial window, your data stays accessible in read-only mode for 30 days so you can export everything cleanly.

What is the refund policy if I subscribe and change my mind?

If you subscribe and decide within 15 days that Concierge is not the right fit, contact support and we refund the prorated month in full - no questions, no retention call. Annual subscribers get a prorated refund on unused months minus a 5% admin fee, processed back to the original payment method within 7 business days.

Can I export my data if I leave?

Yes - your data is yours. Vastura supports full CSV / JSON / PDF export across every product. Concierge exports include all transactional records, configuration, audit log entries, and attachments. We also provide a Cutover Toolkit (Enterprise tier) for migrating to or from competing platforms with mapping templates included.

How does Vastura handle multi-entity setups?

Multi-entity is native, not an afterthought. From the Pro tier upward, you can run multiple businesses (Inc / sole prop / partnerships) under one login with per-entity books, branding, and identity. The Enterprise tier unlocks unlimited entities plus consolidated reporting across the holdco. One operator, multiple businesses, one subscription instead of one-per-entity.

Why is Vastura priced in Canadian dollars when most SaaS is USD?

Vastura Ltd is Alberta-incorporated and built specifically for Canadian operators. USD-priced SaaS hits Canadian SMBs with a 35-40% FX surcharge after exchange and credit card cross-border fees. We price in CAD, charge in CAD, and write all tax surfaces (GST/HST, CCA, T2 prep) to Canadian rules first. International expansion comes later, but Canada is home base.

How secure is my data?

All Vastura data sits on Cloudflare's enterprise tier with end-to-end TLS 1.3, at-rest encryption, role-based access control, and a full audit log on every write. We're on the SOC 2 Type 1 - Type 2 roadmap (Type 1 target Q2 2027) and ISO 27001 (Q1 2028). PIPEDA compliance is built in by default - Canadian data residency is on the roadmap for late 2026.

How does Vastura compare to the popular SMB-tier incumbents in this category?

The honest answer: most SMB-tier alternatives are single-tool, USD-priced, and don't talk to each other. Vastura's moat is the integrated stack - Ledger feeds Pilot feeds Studio feeds Launch Pad through the MCP bridge, and CAD-pricing means no FX surcharge. We don't claim to beat every dedicated single-tool incumbent on every micro-feature, but we beat the total cost of running 6 separate USD SaaS subscriptions glued together with Zapier.

Do I need technical knowledge or a developer to set this up?

No. Onboarding is guided - connect your bank or Shopify, walk through the entity wizard, and you're posting transactions or running storefronts within the first 30 minutes. The optional MCP layer (for AI agent operation) does benefit from a technical user, but the dashboard, reports, and day-to-day workflow are designed for the operator-without-an-IT-department.

What happens if I outgrow my tier mid-month?

Upgrades are prorated and instant - switch from Starter to Pro and you only pay the difference for the remaining days in the cycle. Downgrades take effect at the next renewal so you keep what you paid for. The Conductor agent can also recommend tier changes based on actual usage patterns at the Pro tier and above.

Is there a discount for annual billing or non-profits?

Yes - annual billing is 25% off the monthly rate, applied across every tier and every product. Registered Canadian non-profits and charities can apply for an additional 20% discount (contact support with your CRA registration number). Bundle suites (Ops / Ecomm / Pro / Enterprise) include built-in discounts on top of that.