background

Back

Chatbot Migration Checklist for Shopify Stores (2026)

AeroChat Team

Chatbot Migration Checklist for Shopify Stores

Switching chatbots on a live Shopify store is not complicated — but it does require a clear order of operations. Done properly, the transition is invisible to your customers. Done carelessly, customers hit gaps, agents miss context, and your team is left firefighting during trading hours.

This checklist covers the full migration process from beginning to end: what to document before you start, how to set up and train the new chatbot, how to test it properly, how to make the switch without downtime, and what to watch for in the first week. Work through it in order and nothing will fall through.

Key Takeaways

  • A chatbot migration has five phases: audit, setup, test, switch, and monitor. Skipping any phase creates risk for your customers and your team.

  • Export your existing chatbot's conversation data, FAQs, and custom responses before you close the account — you will need this content to train the new tool.

  • Never make the switch during a high-traffic period. Choose a quiet Tuesday or Wednesday morning.

  • Run a structured test on the new chatbot before it goes live. An 80% or higher resolution rate on your test questions is the minimum bar.

  • The first 48 hours after switching are the most important. Stay close to the dashboard and be ready to update training content if a gap appears.

  • Your team needs a 30-minute briefing on the new tool before the switchover — not after.

Phase 1 — Before You Start: Audit Your Current Setup

The most common reason chatbot migrations go wrong is that store owners dismantle the old tool before they have properly documented what it was doing. Spend time in your current chatbot dashboard before you touch anything else.

1.1 — Record Your Current Performance Numbers

Pull these metrics from your existing chatbot platform and write them down. They become your baseline — the numbers the new tool needs to beat to justify the switch.

Metric

Where to Find It

Write Down Your Number

Monthly chat volume

Dashboard > Analytics

______

Resolution rate (% resolved without a human)

Dashboard > Reports

______

Average response time

Dashboard > Analytics

______

Most common question categories

Dashboard > Top Questions or Tags

______

Monthly agent escalations

Dashboard > Handoffs or Transfers

______

Customer satisfaction score (if tracked)

Dashboard > CSAT or Ratings

______

If your current platform does not show you these numbers, that is useful information in itself. A chatbot you cannot measure is one you have been running blind — and the new tool should surface this data clearly from day one.

1.2 — Export Everything You Can

Before you close your existing account, export every piece of data the platform allows:

  • Your full FAQ and knowledge base content — every question and answer you have trained the chatbot on.

  • Any custom response flows or conversation scripts you have built.

  • Your top 50 most common customer questions (if the platform tracks these).

  • Any canned responses or quick replies your agents use.

  • Historical conversation transcripts if available — these are useful for training the new AI.

Most chatbot platforms allow you to export this content as a CSV or PDF. Even if the new tool uses a different format, having the content saved locally means you are not starting the setup from scratch.

1.3 — List Every Channel the Chatbot Handles

Write down every channel your current chatbot covers:

  • Website chat widget (which pages does it appear on?)

  • WhatsApp chat (if connected)

  • Instagram DMs (if connected)

  • Facebook Messenger (if connected)

  • Email (if routed through the chatbot)

The new chatbot needs to cover every channel your customers currently use. If your existing tool only handles web chat but your customers also contact you through Instagram, this migration is also an opportunity to close that gap.

1.4 — Note Your Agent Handoff Workflow

Document exactly how your current chatbot hands conversations to human agents:

  • What triggers a handoff? (customer request, keyword, unanswered question?)

  • Which team member or inbox does the handoff go to?

  • Does the agent receive the full conversation history, or just a summary?

  • How are after-hours messages handled?

This workflow needs to be rebuilt in the new tool before you switch. If the handoff does not work on day one, your agents will not have the context they need — and customers will have to repeat themselves.

Phase 2 — Setting Up the New Chatbot

Do not touch your existing shopify chatbot while you complete this phase. The new tool is being configured in the background. Your customers are not affected yet.

2.1 — Install and Connect to Shopify

  1. Install the new chatbot from the Shopify App Store.

  2. Grant it the permissions it requests — most chatbots need read access to products, orders, and customers to function properly.

  3. Confirm the Shopify connection is live by checking whether the app can read your product catalog and recent orders. If it cannot, the integration has not been set up correctly.

  4. Connect any additional channels the new tool supports: WhatsApp, Instagram, Facebook Messenger, Telegram, or email.

2.2 — Import Your Training Content

This is the most important step in the setup phase. The new chatbot will only perform as well as the content it has been given. Use the content you exported in Phase 1 as your starting point.

  • Import your full FAQ content — every question and answer.

  • Add your return and refund policy as a document or URL the AI can read.

  • Add your shipping policy — all regions, timeframes, and carriers.

  • Import your product catalog descriptions if the chatbot does not pull them automatically from Shopify.

  • Add any edge case responses you noted from your old chatbot — for example, how to handle a damaged item claim or a missing parcel.

If your new chatbot connects directly to Shopify and reads product and order data live, you may not need to manually enter product information — the AI will pull it from your store. Check with your provider to confirm what gets imported automatically and what needs to be added manually.

2.3 — Rebuild Your Agent Handoff Workflow

Using the workflow you documented in Phase 1, configure the handoff in the new tool:

  1. Set the trigger conditions — what causes the chatbot to offer a human agent?

  2. Connect the inbox or email address where handoffs should go.

  3. Confirm that the full conversation history is passed to the agent when a handoff occurs.

  4. Set up an out-of-hours message — what does the chatbot say when no agents are available?

  5. Test the handoff manually before moving to Phase 3.

2.4 — Configure the Chat Widget

  • Match the widget colours to your store's brand.

  • Write an opening greeting that sounds like your brand voice — not a generic "Hi, how can I help you?"

  • Set the widget position (bottom right is standard for most Shopify themes).

  • Decide which pages the widget appears on. If you want it on every page, confirm it does not interfere with your checkout flow.

  • Set the widget to visible-to-admin or preview mode only. Customers should not see it yet.

Phase 3 — Testing the New Chatbot Before It Goes Live

This phase is where you confirm the new chatbot is ready for real customers. Do not rush it, and do not skip it — even if the setup felt smooth.

If you want a detailed walkthrough of how to run a structured chatbot test, including the specific question categories to use and the metrics to track, our guide on testing a new chatbot before switching covers the full process.

For the purposes of this migration checklist, here is the condensed version:

Phase 3 Testing Checklist

  • Ask the chatbot your 15 most common customer questions. It should answer all of them correctly.

  • Ask three order-specific questions using a real order number. It should pull live Shopify data.

  • Ask five product detail questions about items from your catalog. Answers should match your Shopify product descriptions.

  • Ask three edge case questions the chatbot was not explicitly trained on. Check how it handles the unknown.

  • Trigger the agent handoff deliberately. Confirm the agent receives the full conversation history.

  • Test on mobile (both iPhone and Android if possible). Check that the widget loads, the text is readable, and the input field works with the keyboard open.

  • Run the test on desktop Chrome, Safari, and Firefox.

  • Ask a question in a second language if your customers are international. Note how the chatbot responds.

  • Send a complaint message. ("I am not happy with my order.") Check that the response is empathetic and offers a clear next step.

  • Review the chatbot dashboard after testing. Confirm it is logging conversations and showing accuracy data.

Only move to Phase 4 when you are satisfied with the test results. If the chatbot is getting questions wrong, go back to Phase 2 and update the training content. Then re-run the relevant tests. Most gaps take less than an hour to fix.

A good benchmark: the new chatbot should resolve 80% or more of your test questions correctly and handle the agent handoff cleanly every single time before you go live.

Phase 4 — Making the Switch

You have tested the new chatbot and it is performing well. Now it is time to switch. Done in the right order, this takes less than 30 minutes and your customers will not notice any gap.

4.1 — Choose the Right Time

Timing the switchover well reduces risk significantly. Pick a moment that gives you:

  • Low chat volume — a Tuesday or Wednesday morning between 9am and 11am local time is ideal for most stores.

  • No active promotions or sales — avoid switching during a discount period, product launch, or seasonal peak.

  • You or a team member available to monitor for the first two hours after the switch.

  • At least one week away from any major trading event like Black Friday or a brand campaign.

If you have team members who handle agent escalations, make sure they know the switch is happening and have had a briefing on the new platform before this day. More on that below.

4.2 — Brief Your Team Before You Switch (Not After)

One of the most avoidable migration problems is a support agent who has never seen the new platform receiving a live escalation on the day of the switch. Give your team at least a 30-minute walkthrough of the new tool before switchover day. Cover:

  • How to find incoming escalations in the new inbox.

  • How to read the conversation history that comes with a handoff.

  • How to close a resolved conversation.

  • Where to find customer order details inside the platform.

  • Who to contact if something is not working as expected on the first day.

If your team will be using the new chatbot's inbox to respond to direct messages across WhatsApp, Instagram, and web chat from a single place, walk them through that workflow too. A unified inbox works differently from handling each channel separately, and the first day is not the time to learn it on the job.

4.3 — The Switchover Steps

Complete these steps in this exact order:

  1. Log into your Shopify admin and verify the new chatbot is connected and showing as active.

  2. Make the new chatbot widget visible to all visitors in its settings.

  3. Log into your old chatbot platform and pause or disable the widget. Do not close the account yet.

  4. Visit your own store as a customer and open the chat. Confirm the new widget loads correctly.

  5. Send a test message. Confirm it receives a response.

  6. Trigger a handoff and confirm it routes correctly to your team.

  7. Check the new chatbot dashboard to confirm it is logging the test conversation.

  8. Notify your team that the switch is live and monitoring has begun.

Switchover Day Quick Checklist

  • New chatbot set to visible for all visitors.

  • Old chatbot paused or disabled (account kept open for one week).

  • Widget tested on your own store as a customer.

  • Test message sent and confirmed.

  • Agent handoff tested and confirmed.

  • Dashboard confirming conversations are being logged.

  • Team notified that monitoring is live.

Phase 5 — The First Week After Switching

The migration is not complete on switchover day. The first week is when you validate that the new chatbot performs as well on real customer traffic as it did during testing — and when you catch any gaps that did not show up in the test phase.

5.1 — Monitor Daily for the First 48 Hours

Check the following every morning for the first two days:

  • Resolution rate — is the chatbot resolving conversations at the rate you expected from testing?

  • Escalation volume — are more conversations escalating to human agents than you anticipated? If yes, find out which question types are causing it.

  • Unanswered questions — does the dashboard show any question categories where the chatbot said it could not help? These are training gaps to fill immediately.

  • Agent feedback — ask your team at the end of each day whether they noticed any handoff issues or received escalations where the chatbot gave incorrect information.

  • Customer complaints — check your email and social channels for any customers reporting issues with the chat.

5.2 — Compare Against Your Phase 1 Baseline

At the end of week one, pull the same metrics you recorded in Phase 1 and compare them side by side:

Metric

Old Chatbot (Phase 1)

New Chatbot (Week 1)

Change

Monthly chat volume

______

______

______

Resolution rate

______

______

______

Average response time

______

______

______

Agent escalations

______

______

______

Customer satisfaction score

______

______

______

The new chatbot should show improvement across resolution rate, response time, and escalation volume. If resolution rate has dropped compared to the old tool, something was not migrated correctly — go back to the training content and find the gap.

5.3 — Fill Training Gaps as You Find Them

Real customer traffic will surface questions that your test phase did not. This is normal — no test can replicate every possible question a customer might ask. When you find a gap:

  1. Open the conversation in the chatbot dashboard.

  2. Read exactly what the customer asked and what the chatbot answered.

  3. Decide whether the answer was wrong, incomplete, or simply missing from the knowledge base.

  4. Add the correct answer to the training content.

  5. Verify that the chatbot now answers that question correctly.

Keep a simple log of every gap you find and fix during the first week. This log becomes your quality record and is useful if you ever need to retrain the chatbot after updating your store policies.

5.4 — Close the Old Account After One Full Week

Keep your old chatbot account open but inactive for seven days after the switch. This gives you a rollback option if something goes seriously wrong during the first week. After seven days without any major issues, you can safely close the old account. Before you do, confirm you have saved all exported data locally — once the account is closed, that data is typically gone.

The Complete Chatbot Migration Checklist at a Glance

Use this as your working checklist. Tick each item as you complete it.

Phase

Task

Done?

Phase 1

Record current chatbot performance metrics

[ ]

Phase 1

Export FAQ, knowledge base, and custom responses

[ ]

Phase 1

List all channels the chatbot currently covers

[ ]

Phase 1

Document the agent handoff workflow

[ ]

Phase 2

Install new chatbot and connect to Shopify

[ ]

Phase 2

Verify Shopify order and product data is accessible

[ ]

Phase 2

Import FAQ and policy content

[ ]

Phase 2

Connect all required channels (WhatsApp, Instagram, etc.)

[ ]

Phase 2

Rebuild agent handoff workflow

[ ]

Phase 2

Configure and brand the chat widget

[ ]

Phase 2

Set widget to admin-only preview mode

[ ]

Phase 3

Test 15 common customer questions

[ ]

Phase 3

Test order-specific questions with real order data

[ ]

Phase 3

Test product detail questions

[ ]

Phase 3

Test edge cases and complaint handling

[ ]

Phase 3

Test agent handoff and confirm history passes through

[ ]

Phase 3

Test on mobile (iPhone and Android)

[ ]

Phase 3

Test on desktop (Chrome, Safari, Firefox)

[ ]

Phase 3

Confirm resolution rate is 80% or higher on test questions

[ ]

Phase 4

Choose a low-traffic switchover window

[ ]

Phase 4

Brief the support team on the new platform

[ ]

Phase 4

Set new chatbot to visible for all visitors

[ ]

Phase 4

Pause or disable old chatbot

[ ]

Phase 4

Test widget live on your own store

[ ]

Phase 4

Confirm handoff routes correctly

[ ]

Phase 4

Notify team that monitoring is live

[ ]

Phase 5

Monitor resolution rate and escalation volume daily for 48 hours

[ ]

Phase 5

Collect and act on team feedback after day one

[ ]

Phase 5

Compare week-one metrics against Phase 1 baseline

[ ]

Phase 5

Fill any training gaps found during live traffic

[ ]

Phase 5

Close old chatbot account after seven days

[ ]

Choosing the Right Chatbot to Migrate To

If you have not yet decided which platform to move to, the checklist above applies to any tool — but the choice of platform itself matters as much as the migration process.

The most common question we hear from Shopify store owners considering a switch is whether to choose a human-first live chat tool or an AI-first platform. These are different products built for different store types, and the right answer depends on your team size, your chat volume, and how your customers contact you.

For a detailed head-to-head between two of the most-discussed options for Shopify stores in 2026, our LiveChat vs AeroChat comparison covers pricing, Shopify integration depth, AI capability, channel coverage, and which type of store each tool is best suited to.

If you are considering AeroChat specifically, it offers a free plan with 100 AI responses per month — which gives you a real working environment to complete your Phase 3 testing before committing to a paid tier.

Frequently Asked Questions

How long does a full chatbot migration take for a Shopify store?

Most Shopify stores complete a chatbot migration in two to three weeks from start to finish. Phase 1 (audit and export) takes one to two hours. Phase 2 (setup and configuration) takes two to four hours depending on how much content you need to import. Phase 3 (testing) should run for a minimum of two weeks to cover a realistic range of customer questions. Phase 4 (the switch itself) takes under 30 minutes. Phase 5 (monitoring) runs for the first week after going live. Stores with high chat volumes or complex agent workflows may need an additional week of testing.

Will my customers lose any conversation history when I switch chatbots?

Conversation history from your old chatbot will not transfer to the new platform automatically. Most platforms store historical transcripts separately, and customers who had previous chats with your old chatbot will start fresh in the new one. If a customer contacts you mid-migration about a previous conversation, your team can look up the history in your old platform (keep the account open for at least seven days after switching). Going forward, all conversations will be stored in the new platform.

What happens to the customers who are mid-conversation when I switch?

If a customer has an open conversation in your old chatbot at the moment you switch, that conversation will continue in the old system until it closes. Any new conversations started after the switch will go to the new chatbot. For this reason, it is best to switch during a low-traffic window when the number of active conversations is minimal.

Do I need a developer to migrate my Shopify chatbot?

For most modern AI chatbot platforms designed for Shopify, no developer is needed. Installation is done through the Shopify App Store, and setup is handled through a browser-based dashboard. The only technical requirement is connecting the app to your Shopify store (a guided process) and, if you are adding WhatsApp, completing a Meta Business account verification — which is a one-time process done through Meta's own flow.

Can I migrate mid-month or do I need to wait for the start of a billing cycle?

You can migrate at any point in the month. There is no technical reason to wait for a billing cycle. What matters is the operational timing — avoid peak trading periods, active promotions, and days when your support team is understaffed. A quiet Tuesday or Wednesday morning during a normal trading week is the safest window regardless of where you are in the billing cycle.

What if I realise after switching that the new chatbot is not working as expected?

Keep your old chatbot account paused but accessible for seven days after the switch for exactly this reason. If a critical issue appears in the first week — for example, the handoff is broken or the AI is consistently giving wrong answers — you can re-enable the old tool while you fix the problem. Before escalating to a full rollback, check whether the issue is a training gap that can be fixed quickly. Most post-launch problems are fixable in under an hour by updating the chatbot's knowledge base.

How do I migrate the FAQ content from my old chatbot to the new one?

Export your FAQ content from the old platform as a CSV, PDF, or text file before closing the account. In the new chatbot's admin dashboard, look for a knowledge base or FAQ import option. Most platforms accept bulk uploads via file or URL. If the new chatbot reads your Shopify store's published pages, you can also publish your FAQ as a page on your store and point the AI to it — which has the added benefit of keeping your FAQ content in one place rather than two.

Is it possible to run two chatbots simultaneously on the same Shopify store?

Technically yes — you can have two chatbot apps installed at the same time — but it creates a poor customer experience if both widgets are visible. Customers may see two chat icons, receive duplicate responses, or have their conversation split across two platforms. The correct approach is to keep the old chatbot active but hidden during your test phase and only show the new chatbot once it has passed your test checklist. Never have both widgets visible to customers at the same time.

Ready to scale customer support — without the chaos?

Unify all your customer messages in one place.
No prompt setup. No flow-building. Just faster replies, happier customers, and more conversions.

Ready to scale customer support — without the chaos?

Unify all your customer messages in one place.
No prompt setup. No flow-building. Just faster replies, happier customers, and more conversions.

AeroChat is an omnichannel customer communication platform that unifies chat, email, and ticketing — helping businesses respond faster, support smarter, and convert more — without the chaos.

© 2025 AeroChat. All rights reserved.

AeroChat is an omnichannel customer communication platform that unifies chat, email, and ticketing — helping businesses respond faster, support smarter, and convert more — without the chaos.

© 2025 AeroChat. All rights reserved.