Fixed quote after a free call

AI customer service for ecommerce that stays on your policies

We build an AI customer service agent for your store's chat, email and help desk that answers product, shipping and policy questions from your approved knowledge base and live order data, drafts replies for agents to approve, and hands refunds, complaints and anything uncertain to a person. It tells shoppers they are talking to an AI and never invents a policy. Projects typically cost $2,500 to $7,500 and take five to eight weeks.

The problem

Why most ecommerce chatbots make support worse

Generic chatbots frustrate shoppers because they guess. They promise free returns you do not offer, quote stock levels they cannot see or loop customers through menus. The result is more tickets, not fewer.

A useful AI support agent for a store is narrower and better connected: it answers only from your own policies and product information, looks up the shopper's actual order, knows when to stop, and gives a human agent a clean summary when it hands over.

What we build

What our AI support agent for Shopify includes

Approved knowledge base

Shipping, returns, sizing, care and product facts written and approved by your team; the only source the agent answers from.

Order lookup

Verified customers can ask about their own order; the agent reads status and tracking through the store API.

Product and sizing help

Answers from your catalog data and size guides, with links to the right product page.

Agent assist mode

Draft replies inside Gorgias or Zendesk for agents to edit and send, before any automatic sending.

Clear hand-off

Refunds, complaints, damaged goods and low-confidence answers go to a person with a summary.

AI disclosure and review

The assistant says it is automated, and a weekly report lists questions it could not answer.

Who does what

Automation does the chasing, people make the calls.

Every step is marked: what runs on its own, what a member of staff reviews and what stays with a person who decides.
  1. 01 Automated

    Question asked

    A shopper asks on chat, email or the contact form.

  2. 02 Automated

    Context gathered

    The agent retrieves matching policy and product facts and, if verified, the order.

  3. 03 Automated

    Answer or draft

    Routine questions are answered; others become a draft for a human agent.

  4. 04 Team review

    Agent approves

    Support agents edit drafts and take over handed-off conversations.

  5. 05 Founder or ops lead decides

    Policy and refunds

    Refunds, exceptions and policy changes are decided by the ops lead or founder.

How it behaves

How the ecommerce AI agent behaves

What each person sees and what runs in the background, including what happens when something goes wrong.

What the shopper sees

  • A clear note that they are chatting with an AI assistant
  • Answers that match your actual policies, with a link
  • Their own order status after a simple verification
  • A person whenever they ask or the issue is a refund or complaint

What agents see

  • Drafts ready to approve in the help desk
  • Handed-off conversations with a short summary
  • A weekly list of unanswered questions

What the automation does

  • Retrieves approved facts and order data
  • Refuses to guess and hands over below a confidence threshold
  • Logs every conversation

Edge cases and fallbacks

Shopper cannot verify the order
General policy answers only; order details stay private.
Question about a product claim not in the catalog
No claim is made; the question goes to the team.
Prompt injection or abuse
The agent ignores instructions outside its rules, closes politely and logs it.
AI provider outage
Chat falls back to a contact form and tickets go to the normal queue.
Peak sale traffic
Rate limits and queueing keep the help desk stable; agents see a backlog alert.

Architecture

An AI customer service agent for a store at a glance.

The setup we usually start from. We adapt it to the phone system and practice software you already use.
Shopper Chat, email, form Agent service n8n or Laravel Language model Claude, Gemini, GPT Knowledge base Approved policies Order data Store API Answer or draft With source link Support agent Refunds, issues Transcripts Weekly review
  • Input
  • Logic and services
  • AI step
  • Data
  • Output
  • People

How it flows

  1. 1 A shopper asks a question and is told they are chatting with an AI assistant.
  2. 2 The agent service retrieves approved policy and product facts and, after verification, the shopper's order.
  3. 3 The language model writes an answer from those facts only, or a draft for a human agent.
  4. 4 Refunds, complaints and uncertain cases go to a person, and every conversation is logged for review.

Project planner

How the project runs, phase by phase.

Typical durations for a first version. Your proposal lists the exact phases, dates and what you sign off at each step.

AI customer service agent for your store project plan

Typical time to launch: about 6 weeks, then support

  1. 01

    Discovery

    About 1 week

    Export and group real tickets, agree what the agent may and may not answer.

  2. 02

    Knowledge base

    About 1 week

    Write and approve policy, shipping, sizing and product answers.

  3. 03

    Build

    About 2 weeks

    Agent service, retrieval, order lookup, help desk integration, hand-off and logging.

  4. 04

    Test

    1 to 2 weeks

    An evaluation set of real, tricky and adversarial questions; the team signs off.

  5. 05

    Launch

    2 to 3 days

    Draft-only mode first, then automatic answers for approved topics.

  6. 06

    Support

    Ongoing

    Weekly transcript review at first, then monthly knowledge base updates.

What it runs on

Real products only, connected through the access each vendor provides. See all integrations

Claude, Gemini or OpenAI models n8n AI Agent node or Laravel Shopify Admin API or WooCommerce REST API Gorgias or Zendesk API Website chat widget Slack for hand-off alerts

Cost estimate

What it costs, line by line.

Estimates to plan with. Your fixed quote follows a free call, once we have seen your systems and volumes.
  • Discovery and knowledge base

    Ticket analysis, scope and approved answers.

    $500 – $1,200

  • Build

    Agent service, retrieval, order lookup, chat and hand-off.

    $1,200 – $4,000

  • Integrations

    Store, help desk, chat widget and alerts.

    $300 – $1,000

  • Testing and evaluation set

    Real, tricky and adversarial questions tested before launch.

    $300 – $900

  • Launch and training

    Draft-only launch, transcript review and team walkthrough.

    $200 – $400

  • Total project estimate

    $2,500 – $7,500

Optional monthly care after launch is quoted with your proposal. Fixed quote after a free call. Model usage, help desk and hosting are billed to your accounts directly. Voice support costs more to build and run than chat.

Monthly running costs you pay directly

Estimates from the vendors' published pricing, checked November 2026. Billed to you by each vendor, never marked up.

  • AI model usage: Billed by the provider per use; scales with conversation volume.
  • Help desk: Your Gorgias, Zendesk or other help desk plan.
  • Hosting: n8n Cloud from about €20 a month, or your own server.

What pushes cost up

  • Systems without an API, or with integration access that must be requested from the vendor
  • Writing data back into your core system rather than only reading from it
  • Several locations, brands or very different processes to support
  • Custom apps, dashboards or a voice line on top of the workflows

What keeps cost down

  • One location and one clear process to start with
  • Cloud software with a documented API
  • Using tools you already pay for instead of adding new ones
  • Approved wording and sample data ready at the start
Get a fixed quote

Privacy and consent

Rules we design for, with the sources.

General information to plan with, not legal advice. Your counsel or data protection adviser confirms what applies to your practice.

EU: tell people they are talking to AI

Article 50 of the EU AI Act requires AI systems that interact with people to inform them they are interacting with AI, from 2 August 2026. We add the disclosure in every market.

Source: artificialintelligenceact.eu

EU and UK: processors and personal data

Order lookups process personal data, so model and hosting providers act as processors under a written contract and the agent sees only the fields it needs.

Source: gdpr-info.eu

US: honest claims

The agent answers only from approved wording so it cannot invent shipping promises or policies; under the FTC Mail Order Rule a stated shipping time needs a reasonable basis.

Source: www.ftc.gov

How we deliver

Agile delivery, in 1 to 2 week sprints.

We deliver in short Agile sprints using Scrum, so you see working software every one or two weeks instead of waiting for a big reveal at the end.

Every sprint

  1. 1 Sprint planning. At the start of each sprint we agree which backlog items we will finish, each with written acceptance criteria you can check.
  2. 2 Daily stand-up. A short internal check-in on progress and blockers. You get a written update in your Slack or Teams channel, not another meeting.
  3. 3 Sprint review and demo. At the end of each sprint we demo what was built on test data or in a shadow run, and you accept it or ask for changes.
  4. 4 Retrospective. The team looks at what went well and what to change in the next sprint, and we share anything that affects you.

How scope stays under control

Product backlog
Every feature, fix and question in one prioritised list you can see at any time.
Acceptance criteria
Plain-language conditions for "done" on each item, agreed before work starts.
Change requests
New ideas go into the backlog with an estimate. Nothing changes scope or price without your written approval.
User acceptance testing (UAT)
Before launch your team tests the real workflow against the acceptance criteria and signs off.

Tools we use

Jira, ClickUp or Trello for the backlogSlack or Microsoft Teams for updatesGitHub for code and version historyLoom or a live call for demos

Project manager / Scrum master

Runs the sprints, keeps the backlog and timeline honest, and is your single point of contact.

Business analyst

Maps how the work runs today, writes the workflow specification and acceptance criteria with your team.

Automation engineer

Builds the n8n workflows, integrations, AI steps and error handling.

Backend developer

Writes custom code where a no-code step is not enough: APIs, data sync, secure storage.

Frontend / mobile developer

Builds forms, booking pages, dashboards, web apps or Android apps when the project needs them.

QA engineer

Tests every workflow against the acceptance criteria, including edge cases and failure paths.

DevOps engineer

Sets up hosting, access, backups, monitoring and alerts, and the deployment pipeline.

Who does what at each phase

Free call
Project manager, business analyst
Understand the problem, check it is worth automating and outline a rough range.
Discovery
Business analyst, automation engineer
Map the current process, systems and data, and agree success measures.
Design
Business analyst, automation engineer, backend developer
Write the workflow specification, data flows, compliance notes and backlog with acceptance criteria.
Sprints
Automation engineer, developers, QA
Build, test and demo in 1 to 2 week sprints; you accept or adjust after each demo.
UAT and launch
QA, DevOps, project manager
Your team tests and signs off, we switch on, monitor closely and train the team.
Support
Automation engineer, DevOps
Stabilisation window, then optional monthly care with monitoring and changes.

A point of contact

One person who can answer questions and approve decisions, with about an hour a week during the project.

Access

Admin or API access to the systems involved, created as separate accounts with the least access needed.

Sample data

Real but non-sensitive records (or de-identified ones) so we can design and test against reality.

Approved wording

Message templates, policies and consent wording (for SMS, your opt-in and opt-out text) reviewed by you.

Vendor agreements

Signed agreements with the software vendors that will handle personal data, such as a BAA or processor contract.

Time for testing

A few hours from the people who will use the system, for the demos and user acceptance testing.

Working system

Workflows, integrations and any apps running in your accounts, under your keys.

Documentation

Workflow specification, data-flow map, run book and exported workflow files.

Training

A walkthrough for the team and a short recorded video for new starters.

Handover

All access, credentials and code handed over; you can keep us or take it in-house.

Stabilisation window

Fixes for anything that does not meet the agreed acceptance criteria after launch.

Support options after launch

  • No plan

    You run it yourselves after the stabilisation window and contact us when you need a change.

  • Monthly care

    Monitoring, error alerts, small changes and updates each month, with an agreed response time.

  • Priority care

    Faster response times for business-critical workflows and a monthly review call.

How a project moves through our team

Free call PM and analyst Discovery Process and data Design Spec and backlog Sprints Build, test, demo Your review Accept or adjust UAT and launch Sign-off, go live Support Monitor and improve
  • Input
  • Logic and services
  • People
  • Output
  • Data

How it flows

  1. 1 A free call with the project manager and business analyst to understand the problem and whether it is worth building.
  2. 2 Discovery and design produce a written specification, data-flow map and a backlog with acceptance criteria.
  3. 3 The team builds in 1 to 2 week sprints; you review a demo at the end of each and accept or adjust.
  4. 4 Your team runs user acceptance testing, we launch, then monitor and improve under a support plan.

FAQ

AI customer service agent for your store questions, answered.

It is an AI agent that answers shopper questions about orders, shipping, returns and products, using your own policies and order data, and hands anything it cannot handle to a person. Done well, it removes routine tickets without making customers fight a bot.

Both words get used. Ours is an agent in the sense that it can look things up (orders, tracking, product data) and take limited actions, but it is fenced in: it only answers from approved sources and hands over anything outside its scope.

We do not let it. It can explain your refund policy and start a return request, but releasing money stays with your team.

It is built to refuse rather than guess. If the knowledge base has no answer, it says so, offers a person and adds the question to the weekly review list.

Yes. We often start in agent-assist mode, where it writes drafts inside the help desk for agents to approve, then turn on automatic replies for topics that prove reliable.

Watch the share of conversations resolved without hand-off, customer satisfaction on AI conversations, reopened tickets and the questions it could not answer. We set up the logging from day one.

Free 20-minute call

Want this for your e-commerce brand?

Book a free 20-minute call. We will look at how the work runs today, tell you honestly whether it is worth building and give you a fixed quote if it is.

Book a free 20-minute call