Your AI support employee.
You stay in control.
Inbox investigates customer conversations with Shopify orders, approved knowledge, and Customer 360 context, then drafts the reply or routes the decision for approval.
- Shopify facts, knowledge citations, and conversation history beside every draft
- Cases, investigations, and approval requests for refunds, replacements, and exceptions
- Grounding preflight + audit log before any customer reply is sent
Guided Gmail + Shopify setup for the first commerce brands. Cancel anytime.
V2 Workspace
Customer 360 + grounded drafts
“Where is my Glow Serum order? Tracking has not moved.”
Janelle B. · #LB-2087
Damaged item · approval route
Aisha R.
Sensitive skin · cited answer
Grounded draft
Shopify + knowledge proof
Human-approved actions
Proof should be inspectable before it becomes a customer story.
Until approved customer case studies exist, the homepage uses a transparent demo workspace: visible product flows, public counters, explicit assumptions, and no hidden claim that a demo brand is a paying customer.
Source
Working V2 demo workspace
GlowSkin Beauty is a controlled commerce demo tenant used to exercise the current product, not a fabricated external customer endorsement.
Workflow
Evidence before reply
The demo shows order context, approved knowledge, policy evidence, approval routing, and audit proof in one reviewable path.
Safety
Approval gate stays visible
Unauthorized AI sends are tracked as a public proof point. Customer-facing replies are framed around human-approved decisions.
Economics
Modeled savings, clearly labeled
Savings estimates are modeled from reduced lookup and drafting time. They are not presented as measured customer revenue outcomes.
Evidence path shown in the demo
The same path appears across the product screenshots, video, metrics, and security sections.
Customer issue
Shopify order context
Approved knowledge
AI draft + preflight
Human approval
Audit proof
Product demo
Watch the current V2 workflow end to end
A guided recording from the working GlowSkin demo workspace: dashboard triage, Customer 360, commerce context, approved knowledge, AI health, audit proof, and connector setup.
Watch it work
From customer conversation to evidence-backed action
Every message runs through a visible commerce workflow: customer match, Shopify context, knowledge citations, investigation, draft, approval, and audit.
Receive
A customer conversation lands from Gmail, ready for the V2 workspace.
Customer 360
Inbox matches the customer, order, fulfillment, tags, cases, SLA, and history.
Ground
It retrieves Shopify facts and approved knowledge with citations before drafting.
Investigate
For refunds, damaged items, and exceptions, it creates evidence-backed cases.
Draft & validate
The AI prepares an on-brand reply, recommendation, or follow-up question and checks it before send.
Approve & audit
Humans approve sensitive decisions; every send keeps grounding proof in the audit log.
6 steps · strictly linear · no black box
Sample drafts
Drafts with evidence attached
Each example shows the customer issue, the drafted answer, the proof sources, and the review route. Examples are illustrative, not customer activity.
Maya · #LB-1042
“Where is my Vitamin C Glow Serum? Tracking has not moved for five days.”
Hi Maya - I checked order #LB-1042. Your Glow Serum left fulfillment and BlueDart shows the parcel is delayed in transit. I will monitor it today and open a courier case if there is no scan by 8pm.
Janelle · #LB-2087
“The Rose Barrier Cream jar arrived cracked and leaked in the box. I want a replacement or refund.”
Hi Janelle - I am sorry the Rose Barrier Cream arrived damaged. Please send a quick photo of the jar and packaging, and I will route the replacement or refund for approval right away.
Aisha
“Is the brightening serum okay for sensitive skin? I react to fragrance.”
Hi Aisha - yes, the Vitamin C Glow Serum is fragrance-free and patch-test friendly. If your skin is very reactive, start every other night and avoid layering it with exfoliating acids.
Nora · #LB-3310
“Can I change the shipping address for my cleanser order before it ships?”
Hi Nora - order #LB-3310 is still unfulfilled, so we can update it. Please reply with the full new address and I will confirm before the label is created.
Priya
“Can you pause my cleanser subscription for August? I will be travelling.”
Hi Priya - I can help pause the August cleanser shipment. I am checking the next renewal window now and will confirm the pause before any replenishment order is prepared.
Lina
“I missed the return window by two days. Can you still help? The moisturizer is unopened.”
Hi Lina - thanks for explaining. Since the moisturizer is unopened and you are only two days past the window, I am routing this as a policy exception for approval before I promise the next step.
Every draft waits for your approval before it reaches a customer.
Drafting preview
Paste a customer issue. Watch the AI prepare the first pass.
No signup. This preview shows drafting and safety validation. The real workspace adds Shopify context, knowledge citations, cases, preflight, and audit proof.
Runs the real drafting engine live. The full workspace adds Shopify context, knowledge citations, and audit proof.
Your numbers
See what support operations are costing you
Drag the sliders. Inbox gives back the research, order lookup, evidence gathering, and first-draft time while keeping decisions under human control.
Operator time recovered
42 hours
Support capacity unlocked
$1,664/mo
You still approve sensitive work. Inbox removes the Shopify lookup, knowledge search, evidence gathering, and first-draft work. Flat price, no per-resolution fees.
Start saving — free for 30 daysSee it in action
Every page has a job in the support workflow
Step through the current V2 workspace: operations, conversations, knowledge, AI observability, audit proof, reports, reusable replies, settings, team controls, privacy, integrations, and billing.
Why founders pick Inbox instead of...
| Feature | Manual Support | Gorgias | Siena | Inbox |
|---|---|---|---|---|
| Shopify Order Context | Manual lookups only | Limited integration | Basic order data | Real-time, 360 view |
| Knowledge Citations | Memory-based | Not emphasized | Not emphasized | Verified sources |
| Human Approval (default) | Manual always | Automation available | Automation available | Approval-first model |
| Audit Trail (tamper-proof) | No record | Limited | Limited | Immutable log |
| Pricing Model | $3-5K/mo per person | Tiered and usage-based | Tiered and usage-based | $129-159 flat/mo |
| Setup Time | 1-2 weeks | 1-2 days | 1-2 days | 30 minutes |
| Customer Data Risk | Low (human review) | Depends on automation setup | Depends on automation setup | Low (approval gate) |
Shopify-First
Order context, not guesses
Approval-First
You stay in control
Predictable Pricing
$129, not per-resolution
Why founders switch
Inbox vs the alternatives
Manual support doesn't scale. Fully autonomous AI creates trust risk. Inbox gives you speed while keeping approval in the loop.
| Feature | Inbox ✓ | Shared inbox | Generic AI bot |
|---|---|---|---|
| Commerce-native Customer 360 beside every conversation | |||
| AI drafts with Shopify facts and knowledge citations | |||
| Cases and investigations for refunds, returns, and damaged items | |||
| Human approval for replies and sensitive commerce actions | |||
| Grounding preflight before a customer-facing send | |||
| Searchable audit log with actors, filters, and grounding proof | |||
| AI health, eval gates, and model-call observability | |||
| Knowledge source publishing, freshness, and policy control | |||
| Assignment, SLA filters, notes, and team handoff | |||
| Flat monthly price without per-resolution fees | |||
| DPA, export/delete, MFA, tenant isolation, and hardened auth | |||
| Designed for Shopify-first D2C teams instead of generic tickets |
Who it's for
Built for how you actually run support
However you handle customer conversations today, here's exactly what Inbox takes off your team - and what still stays under human control.
The solo founder
You run the whole store and still answer every “where’s my order?” and refund message yourself.
- Inbox pulls Shopify order, fulfillment, and customer history into the thread
- AI drafts the reply with evidence and shows what facts it used
- You approve the send before anything reaches the customer
Fewer manual lookups, faster replies, and control over every customer-facing send.
The small CX team
Two or three of you split the inbox and worry about staying fast and on-brand.
- Use assignment, SLA filters, internal notes, and case counts in the V2 workspace
- Publish team-approved policies and saved replies into the Knowledge base
- Review AI health, grounded sends, and approval decisions in one audit trail
A tighter operating rhythm without turning your support desk into a generic ticket maze.
The scaling brand
Volume spikes hard around launches and BFCM — and per-ticket tools punish you for it.
- Let AI investigate repeatable order-status, damaged-item, return, and policy questions
- Route refund, replacement, discount, and exception approvals to the right person
- Use support data to spot courier, warehouse, product, and knowledge gaps
Support becomes a commerce operations signal, not just a cost center.
The control-first operator
You want AI leverage but cannot risk a wrong answer or a compliance gap reaching a customer.
- Grounding preflight checks recipients, citations, Shopify facts, and blockers
- Approval requests keep sensitive commerce actions human-reviewed
- DPA, export/delete, MFA, tenant isolation, and audit logs are built into the product
Real AI speed with evidence, accountability, and a reviewable trail.
Start with your use case
Different teams, different paths. Pick yours to get started the right way.
Solo Founder
One person handling all support
Small CX Team
Growing support operation (2-5 people)
Scaling Brand
Multi-brand or high-volume (5-50 people)
Enterprise
Security-first, compliance-heavy
Why founders pick Inbox
Commerce support work, not generic tickets
V2 conversation workspace
Open, assign, filter, and resolve commerce conversations with SLA status, case counts, notes, and reply history in one dense operator view.
Customer 360
Every thread carries the customer, Shopify orders, fulfillment, tags, previous cases, notes, and risk context beside the composer.
Knowledge with citations
Publish policies, FAQs, manuals, saved replies, and product facts so AI drafts can cite exactly what they used.
Cases & investigations
For refunds, damaged items, replacements, and exceptions, Inbox gathers evidence and turns messy messages into structured cases.
Approval engine
Reply sends, refunds, returns, replacements, discounts, and policy exceptions can all route to human decision before action.
Grounding preflight
Before a reply sends, Inbox checks recipients, thread IDs, Shopify facts, citations, validation, blockers, and warnings.
AI health & evals
Track model calls, success rate, grounded sends, repairable proof gaps, latency, and scenario evaluation gates.
Tamper-evident audit log
Every approve, reject, edit, grounded send, actor, and repair is searchable and filterable for trust reviews.
Flat, predictable pricing
The product is built to avoid per-resolution fees that punish growing brands during launches, weekends, and BFCM spikes.
Get started in 6 minutes
Each step takes less than 2 minutes. You'll be handling customer support with AI drafts and your approval before you finish your coffee.
Connect Gmail
Inbox reads your support inbox via secure OAuth.
- Click 'Connect Gmail'
- Approve OAuth scope
- Status turns green
Connect Shopify
Link your store for order context in every reply.
- Enter your store URL
- Approve Shopify app
- Orders now visible
Set Your Brand Voice
Configure how the AI should sound and behave.
- Choose tone (friendly/professional)
- Set refund policy
- Add signature
Paste an Email → See Draft
Test with a real customer support email.
- Copy a customer email
- Paste it in the live demo
- See AI draft appear
Click Approve → Email Sent
Your first human-approved AI reply.
- Review the draft
- Click Approve
- Reply sent via Gmail
Check Audit Trail → See Proof
Every send is recorded in the audit log.
- Visit Audit Log
- See grounding proof
- Know it's traceable
Total setup time
6 minutes
Then you're live and handling customer support with AI drafts + your approval.
Security & trust
Built to be safe with your store and your customers
Grounding, approvals, AI health, tenant isolation, and audit logs make the system reviewable before it becomes autonomous.
A trust posture your buyer can actually inspect
Inbox sells an AI employee, so trust cannot be a footer badge. The homepage now shows the exact controls behind approval-first automation, data boundaries, OAuth safety, auditability, and the enterprise roadmap.
Approval plus preflight
Sensitive sends, refunds, replacements, discounts, and policy exceptions wait for a human decision after grounding and recipient checks run.
Protected access
TOTP two-factor login is available, OAuth tokens are encrypted at rest, and secrets are never exposed to the browser.
Tenant isolation
Every brand's data is isolated at the database level with row-level security so queries cannot cross stores.
Untrusted customer input
Customer email, attachments, and order notes are treated as data, never instructions to the AI, to reduce prompt-injection risk.
AI health and evals
Model calls, success rates, grounded send proof, scenario gates, and repairable evidence gaps are visible in the analytics workspace.
Searchable audit proof
Approvals, edits, sends, actor IDs, grounding proof, and repair actions are reviewable from the audit log.
AI support should make founders faster without making them careless.
Most support tools make the inbox neater. They do not answer the hard part: what happened with this order, what policy applies, what tone should we use, and who is allowed to make the final call?
Inbox is built around a stricter idea: the AI does the investigation, but the brand keeps the decision. That is why the product centers Shopify context, approved knowledge, preflight checks, and an audit trail before a reply reaches a customer.
No customer-facing reply without grounding proof and an approval path.
No fake customer logos, fake testimonials, or hidden automation claims.
No connector expansion unless it preserves tenant isolation, evidence, and audit logs.
No per-resolution pricing that punishes a brand for growing.
Prince Roy
Founder & CEO, The Agentic Group
Prince Roy
Founder-led onboarding for early commerce brands. The goal is not to replace your judgement. It is to remove the research, lookup, and drafting work around it.
Problem
Support grows faster than the founder can hire
Commerce teams lose hours to order lookups, policy checks, refund judgement calls, and tone matching. Generic ticketing tools organize the pain, but they do not remove the work.
Insight
Commerce support has evidence
Every customer has orders, fulfillment events, policies, product facts, and conversation history. That context is what lets AI investigate instead of guessing.
Principle
Approval is the trust layer
The AI can research, draft, and validate faster than a human. The human still owns the decision, especially when money, exceptions, or brand trust are involved.
Bring every support decision into one workspace.
Inbox drafts with Shopify context, knowledge proof, preflight checks, and approval routing before anything reaches a customer.