Indition is AI customer support, trained on your Ecommerce store and running on premium models. Let it resolve customer issues on its own, or assist your live agents with accurate, on-brand replies — consistent, governed, and fully under your control.
Five AI actions power every reply — whether Indition sends it autonomously or your agent reviews it first. No tab-switching, no copy-pasting, no lost context.
Generates a context-aware reply grounded in the full conversation and your store data. Send it autonomously, or let your agent review and edit first — you decide per site.
Searches your knowledge base and business data to surface precise answers, with citations so answers can be verified against the source.
Cleans up and sharpens any draft — clearer structure, professional language — while preserving every fact and commitment in the original.
Rephrases any reply as Professional, Friendly, Concise, or Empathetic, matching the moment without changing the facts.
Creates a concise internal recap for handoffs, escalations, and case documentation. Inserts into an internal note, never the customer reply.
Choose Suggest, Assist, or Auto per site with full audit logs and instant human takeover.
You decide how much Indition runs on its own — keep every reply under agent review, hand routine conversations to fully autonomous mode, or anywhere in between. Whatever you choose, every AI action is grounded in your data, logged for audit, and reversible — with instant human takeover on any conversation.
You handle a few simple setup choices. Indition connects the data, trains the AI, and keeps improving once your team is live.
Administrators control every aspect of AI Assist per site — which tools appear, which tones are allowed, who gets access, and how data is handled. No surprises.
Enable only the AI actions your team is prepared to support. Choose available tones per site — Professional, Friendly, Concise, Empathetic. Set sensible defaults for new agents.
Business data access is off by default and requires explicit approval at the site level, role level, and AI Assistant policy level — the most restrictive policy always wins.
Supports Global Administrator, Administrator, Manager, Agent, and External Agent roles. Performance reports and logs are restricted to admin and manager roles only.
Show citations in AI results so agents can verify answers against source documents. Source detail view is available to authorized admin roles only — policy-gated at every level.
Control how many recent messages (5–100, default 30) are included in AI context. Tune for your conversation lengths without exposing irrelevant history or overloading prompts.
Site-scoped AI Assist performance dashboards and detailed event logs track usage, adoption rates, latency, error rates, and generation-to-send linkage across your entire team.
Start with a full-power 14-day trial — the entire product, your data, our best models. Then pick the plan that fits.
14-day free trial · full product · no credit card to install · cancel anytime
"I’m running the store most of the day, so I don’t have time to sit down and answer every message that comes in. I’m also not very technical, but the AI is easy to use and helps me respond to customers quickly without having to figure everything out myself. It feels like having an extra set of hands."
"My husband and I run the store together, and before this we were constantly asking each other who replied to a customer or what was already discussed. Now the Indition AI helps summarize conversations and makes it much easier for either of us to pick up where the other left off. We communicate a lot better than before!"
"Customer messages used to take up a big part of my day. Writing the same answers over and over was exhausting, especially during busy hours. Now Indition Ai Chat does all the work for me. I still have control over what gets sent, but I spend a lot less time typing and a lot more time actually running my store.”
Use industry-leading AI to power faster replies, smarter recommendations, and better customer conversations.
Claude helps power thoughtful, context-aware conversations that feel natural and easy to understand. Use it to summarize conversations, generate helpful replies, and work through complex customer questions.
OpenAI models help generate fast, accurate responses using the context of each customer conversation and your business information, helping your team spend less time typing and more time running the business.
Yes. Run it in Auto mode to resolve conversations on its own, bounded by the guardrails and data policy you set, with every action logged and one-click human takeover. Or keep it in Suggest/Assist mode so agents approve replies first — your choice, per site.
The most common causes are: the conversation has ended (AI actions are intentionally disabled for ended conversations), Agent Assist has not been enabled for the Chat site, the agent's role doesn't have access, or the site has no active AI Assistant and collection configured. Ask your administrator to verify site settings, role access, and active AI Assistant.
Yes. Find Answer can search your configured business data collection — which can include your knowledge base, product documentation, policy pages, and other approved internal sources. Business data access is policy-gated: the site administrator, agent role, and AI Assistant policy must all permit it before any internal data is used in a response.
There are five actions: Suggest Reply, Find Answer, Rewrite, Change Tone, and Summarize. Administrators can enable or disable each one per Chat site. You can roll out with just Suggest Reply at first, then expand as your team gets comfortable. The default action and available tones are also fully configurable per site.
The four tones are Professional (clear and formal, good as a default), Friendly (warm and conversational, great for repeat customers), Concise (brief and direct, ideal for simple answers), and Empathetic (compassionate and supportive, best for complaints or difficult situations). Administrators choose which tones are available per site and set a default. All tone changes preserve the original facts — only the wording changes.
Every generation and subsequent agent action (Insert, Copy, Discard, Send) is recorded in three schema-managed tables. Site-scoped performance reports and detailed logs are accessible at /app/chat/report/ai-assist and /app/chat/report/ai-assist/logs for administrators and managers. Standard agents do not have access to these reports. The AI assisted badge appears on any sent message that was linked to an inserted AI generation.
Clarification is a valid AI response — not an error. It means the current context wasn't specific enough to generate a reliable final answer. When this happens, Insert is intentionally disabled (you can't send a clarification question as your reply). Use Copy to adapt the clarification as a question to the customer, collect their answer, then run the action again with better context.
AI Assist will warn you that the displayed result may be stale because a newer visitor message arrived after generation. Do not send the stale result without reviewing the new message first. For Suggest Reply or Find Answer, regenerate against the updated context. For Rewrite or Change Tone, update your draft first, then run the action again. Discard any result that is no longer relevant.
Yes — all paid plans include a 14-day free trial and no credit card is required to install. The Starter plan is free to install with 200 generations per month. You can start with Suggest Reply and the Professional tone, run a smoke test in a real conversation, and expand from there. Downgrade, upgrade, or cancel anytime from your Shopify admin.
Yes. Turn off Enable Agent Assist in Chat Site → AI Settings → Agent Assist and save. AI actions immediately stop appearing in the live-agent workspace. No destructive migration is required — all generation and event records are retained as audit evidence. You can re-enable at any time without data loss.
Trained on your store, running on premium models, and governed by you. Setup in 2 mins and resolve more issues automatically or assist your agents on the rest — and stand behind every answer.