Automatic vs manual chargeback response on Shopify

The idea sounds wonderful until you picture it going wrong. A tool that files every chargeback response for you, while you sleep, with no one looking. Then you picture it sending a weak response on a $900 order because a tracking page was missing, and the bank's decision being final. Now you are not sure you want the tool to touch anything without you.

Both instincts are right, which is why the automatic vs manual chargeback response question has no single answer. Manual mode means a tool prepares the evidence and you approve every response before it goes out. Automatic mode means the tool files for you. The sensible position for most stores is in between: start fully manual, measure how often you change the draft, then let the tool file on its own only for the cases where your review adds less value than it costs. This guide gives you the formula, a table of cut-off order values, a staged rollout and the questions to ask any tool, so you can decide with numbers instead of nerves.

This guide is for you if:

You use a chargeback tool, or are about to, and you are deciding whether to switch on automatic filing.
You are worried about a bad response going out without your sign-off.
You are worried about the opposite problem: missing a deadline because you did not get to a dispute in time.

Last updated: 30 October 2026. Written by the ChargeFix team. Reviewed against sources on 30 October 2026.

A note on bias before we start: we make ChargeFix, which offers both modes, with manual as the default and automatic available on Starter and above. We want you to use whichever fits, and we say below where full manual is the better choice. The figures in the formula are assumptions you should replace with your own, and the tool features come from vendor pages we read on 30 October 2026.

The 30-second answer (TL;DR)

In Manual mode a tool drafts and you approve each response. In automatic mode it files without your approval.
Evidence is final once submitted, and the bank's decision is final, so a bad automatic response cannot be taken back.
If you submit nothing, Shopify sends order data on the due date, which is a weak default. Missing a deadline is a real cost of Manual mode.
Reviewing a draft is worth it when the dispute is large enough. With a 10% error rate in drafts, a reviewer who catches 80% of them and a 30 point loss in win chance per error, the cut-off is about $167 for the order plus the $15 fee.
A sensible policy: manual for the first 20 disputes, automatic for small orders with a strong evidence score after that, and manual always for high-ticket orders.
Auto refunds and auto declines are a different, riskier kind of automation than auto filing. Treat them separately.

Quick note: Automatic does not mean careless. A good tool still scores the evidence, tells you when a case is weak and follows rules you set. The question is who presses the final button, and whether that person adds value.

At a glance

	Manual	Automatic	Hybrid
Who approves	You, each time	Nobody, by rule	You for some, the tool for others
Time per dispute	About 8 minutes of review	None	8 minutes on a share
Risk of a bad response	Low if you review well	Higher, uncorrectable	Controlled by your rules
Risk of a missed deadline	Real	Low	Low
Best for	First month, high-ticket, odd cases	Small, routine disputes	Most stores after the first month

Automatic vs manual chargeback response: what is the difference?

The two words describe who presses submit.

Manual. The tool reads your order data, assembles evidence, scores it and drafts a response. Nothing is sent until you look at it and approve. In ChargeFix, Manual mode is the default, you can attach your own files, and the letter is drafted from your records only.

Automatic. The tool does the same preparation and then files the response without you approving each one. In ChargeFix, automatic mode is available on Starter and above. Other tools use different names, such as an autopilot, an AI agent or automation, and they differ in what they do before filing, so read the description.

Hybrid. You set rules, and the tool files automatically when a dispute meets them, for example an order under a set value with an evidence score above a set level. Everything else waits for you.

There is also a fourth thing that people confuse with filing: tools that act on money, such as automatic refunds of weak cases. ChargebackWiz's listing, for example, describes automatic refunds for weak cases. That is more consequential than filing a letter, because a refund cannot be undone. We cover it below.

What happens if you do nothing?

This is the baseline that both modes are competing with. When a dispute opens, Shopify gives you a response due date, typically 7 to 21 days away. If you submit nothing, Shopify sends the order data it already holds on the due date. That is a thin response without your delivery proof, your communication or your policy, and it usually carries a low chance of winning.

After you submit, or after Shopify submits on your behalf, the evidence is final. The bank can then take up to 75 days to decide. You cannot add to the response, and you cannot refund the customer while the chargeback is open.

So doing nothing is itself an automatic response, and it is the weakest one. The real choice is between a thin default, a careful manual response and an automatic response built from your records. All three are final.

What can go wrong in automatic mode?

It is worth naming the risks plainly, because they are the reason people hesitate.

A weak or incomplete response goes out. If a tracking page is missing or an order note contradicts the story, an automatic filing can send something that undercuts you. You do not get a second try.
Wrong reason, wrong evidence. If the tool misreads the reason for the dispute, it can send proof that does not answer it.
Fighting a case you should have refunded or accepted. On a small order with no proof, fighting can cost more in effort than it recovers. A good tool scores the case and says so.
Sensitive details. A response may include customer data you would rather check first.
Surprise at volume. A burst of disputes filed overnight, with no one looking at any of them.
False confidence. Because filing is invisible, people stop paying attention to outcomes and miss the signs that something is off.

These risks are real, and they shrink with good scoring, clear rules and regular checks of outcomes. They do not shrink if you switch it on and look away.

What can go wrong in Manual mode?

Manual mode feels safe, and it has risks of its own that people tend to underplay.

Missed deadlines. You go on holiday, a busy week hits, an email lands in spam. A dispute that reaches its due date unanswered gets the thin default.
Inconsistent review. Your attention on a Friday afternoon is not your attention on Monday morning, and review quality slips when you are tired.
Time cost. Eight minutes a dispute is small, and 30 disputes a month is four hours.
Rubber-stamping. If you approve nearly everything without reading, you carry the time cost and get no protection.
Decision fatigue. Having to decide on every case, including ones where the answer is always the same.

The trade is simple. Manual mode risks a good dispute going unanswered. Automatic mode risks a weak dispute being answered badly. Which risk is larger depends on the order value and on the quality of the tool.

Is reviewing each response worth your time?

Here is the sum that settles most of it. A review is worth doing when what it saves, in expected dollars, exceeds what it costs, in your time.

Value of a review = f x r x d x V, where:

f is the share of drafts that have a real problem, such as a missing document or a wrong emphasis.
r is the share of those problems your review catches.
d is the drop in win chance that the problem causes, as a fraction.
V is the dollars at stake, which is the order value plus the $15 Shopify fee.

Cost of a review = minutes x your hourly value. At 8 minutes and $30 an hour, that is $4.

Review is worth it when f x r x d x V is greater than $4, which means V is greater than $4 divided by (f x r x d). That cut-off value, V*, tells you the size of dispute above which you should look at every response.

With r at 80%, the cut-offs look like this:

Share of drafts with a problem	Drop in win chance of 20 points	30 points	40 points
5%	$500	$333	$250
10%	$250	$167	$125
20%	$125	$83	$63

Check one cell: with f at 10%, r at 80% and d at 30 points, f x r x d is 0.024. Dividing $4 by 0.024 gives $167. A dispute where the order plus fee is above $167 is worth reviewing, and below that your eight minutes cost more than the expected benefit.

Two caveats. First, you do not know f until you measure it, and we show how below. Second, the formula puts no price on the stress of a bad outcome or on the risk of the tool's mistakes growing with volume. If you would simply sleep better with a look, that has value.

Illustration: a store with small, routine disputes

Mina sells phone cases at about $40 an order. She gets 12 disputes a month, and V is $55 including the fee.

With f at 10%, r at 80% and d at 30 points, a review is worth 0.024 times $55, which is $1.32, against a cost of $4. At 12 disputes, reviewing every one costs $48 of her time to protect about $16 of expected value.

For Mina, automatic filing on cases with a strong evidence score is reasonable, with a manual look at the weak ones. She should still check outcomes monthly. If her edit rate in the first month turns out to be 30%, not 10%, the cut-off falls to $56 and the picture changes. The numbers should follow her own data, not ours.

Illustration: a store with a mix

Dario sells mid-range audio gear. In a month he has 15 disputes: ten on $50 orders and five on $400 orders.

Group	Count	V per dispute	Value of a review at the base case	Cost of a review
$50 orders	10	$65	0.024 x $65 = $1.56	$4
$400 orders	5	$415	0.024 x $415 = $9.96	$4

For the small group, review loses money, and for the large group it wins. Reviewing the five large disputes costs 5 times $4, which is $20, and protects an expected 5 times $9.96, which is $49.80. Reviewing all 15 would cost $60 in time and protect about $65.40 in total. The first policy returns $29.80 net, and the second returns $5.40. A hybrid, with the $400 orders held for manual approval and the $50 orders filed automatically when the evidence score is strong, is better than either extreme.

Illustration: a high-ticket store

Joan sells furniture at $1,500 an order and gets 4 disputes a month. V is $1,515.

At the base case, a review is worth 0.024 times $1,515, which is $36.36, against $4 of her time. Even at a 5% problem rate it would be worth $18.18. At that scale she should also spend more than eight minutes: 20 minutes at $30 an hour costs $10 and is still comfortably justified. For Joan, every response should be manual, and she should add her own files, such as signed delivery paperwork or photos, to the draft. She should treat the tool as a fast assistant and herself as the decision maker.

What is a sensible policy by evidence score?

Many tools give an evidence strength score, and ChargeFix gives a score plus a verdict on whether a case is worth fighting. A score lets you combine order value and evidence quality into one rule. This is a template, not a rule from any vendor, so adjust the cut-offs to your own data.

Evidence score	Order plus fee under your cut-off	Order plus fee over your cut-off
Strong	Automatic	Manual review, then file
Medium	Manual review	Manual review, add your own files
Weak	Decide: accept or refund, do not fight automatically	Decide by hand; fight only if the odds justify it

The weak row matters most. A weak case is where automatic filing does the most harm, because it spends effort and, on some plans, a filing from your allowance, on a case that is unlikely to win. In ChargeFix, a dispute counts as one filing the first time its evidence is saved or filed, and usage resets on the 1st of each month, so a plan's allowance is worth protecting from low-value fights.

How do you roll automatic filing out safely?

Do not flip the switch on day one. Move through four stages and let your own data decide when to advance.

Stage 1: fully manual for the first 20 disputes or 30 days. Review every draft. For each one, record whether you changed it, and why. This gives you your edit rate, which is the closest measure you will have of f.

Stage 2: compute your edit rate and your cut-off. Divide the drafts you changed in a meaningful way by the total. If you changed 2 of 20, f is 10%. Put it in the formula with a realistic r and d, and find V*.

Stage 3: automatic for the safest slice. Turn on automatic filing only for disputes below V* with a strong evidence score. Keep everything else manual.

Stage 4: sample and audit. Each month, pull a random one in five of the automatic filings and read them as if you were the bank. Compare win rates between automatic and manual cases, and watch for drift. If you see a pattern of weak filings, tighten the rule.

At any stage you can go back. A policy is a setting, not a commitment.

What should you measure to know it is working?

Four numbers tell you most of what you need.

Measure	How to calculate it	What it tells you
Edit rate	Drafts you changed / drafts reviewed	How often review adds value (f)
Win rate by mode	Wins / decided cases, split automatic and manual	Whether automatic is holding up
Missed deadlines	Disputes that hit their due date unanswered	The cost of manual delay
Net recovered per dispute	(Wins x average order minus fees) / disputes filed	The overall result

Compare win rate by mode carefully. Automatic cases are usually the easy, small ones, so a higher automatic win rate may reflect case mix and not quality. The fairer test is automatic against manual on the same order-value band and score.

What about missed deadlines in Manual mode?

The cost of a missed deadline is not just the thin default. It is the chance of winning you gave up. If your usual win chance is 40% on a $115 case, an unanswered dispute loses an expected $46 of value. If 3% of your manual disputes slip, that is $1.38 per dispute on average, which adds to the case for automatic filing on routine cases.

You can reduce the risk without full automation.

Set a reminder for the day a dispute arrives and for three days before its due date.
Use a tool that shows due dates in one list. Our dispute deadline countdown tool does this.
Ask whether your tool can file automatically if a dispute has gone unreviewed until a set day. We have not assumed any vendor offers this, so ask.
Name a backup person for holidays and busy weeks.

A deadline fallback is the best of both worlds: you review what you can, and the rest does not fall off the cliff.

What are auto refunds and auto declines, and why are they different?

Some tools go beyond filing. ChargebackWiz's listing describes automatic refunds of weak cases. Fraud and alert tools may decline or refund orders by rule. These are not the same risk as an automatic letter.

A refund is permanent. It sends money back and cannot be recalled. Filing a letter might win, and refunding cannot.
A decline loses a sale. A good customer turned away does not return.
They change your ratios differently. A refund before a dispute can prevent it. A declined order never happens.

If you turn on any feature that moves money or blocks orders, treat it as its own decision with its own test. Run it in a mode that only recommends before you let it act, and check that its rules match your policy. Our guide to chargeback insurance versus a chargeback app covers the cost of declines, and our Shopify chargeback fees guide explains when an early refund beats a fight.

What should you ask any tool about its automatic mode?

What exactly does automatic mode do: draft, file, refund, decline or all of them?
Can I set rules by order value, reason code or evidence score?
Can I leave some disputes manual while others are automatic?
What happens to a weak case: is it filed, flagged or held?
Can I see every automatic response afterwards, exactly as sent?
Can I turn automatic mode off instantly, and does that affect disputes already in progress?
Does automatic filing count against my plan allowance, and how?
What customer data is read and stored in the process?
Is there a deadline fallback, such as filing on a set day if I have not reviewed?

A vendor that answers these in plain words, in writing, is confident about the product. Our chargeback automation guide goes deeper on what automation can and cannot do.

What does a normal week look like in each mode?

Numbers help, and so does picturing the routine. Take a store with 8 disputes a week, six of them small and routine and two of them larger.

Fully manual. On Monday morning you open the dispute list, sort by due date and work through each case. Each one takes about 8 minutes of review, so the whole batch is roughly an hour, and you reply to anything the tool flagged as weak. The routine stays in your calendar every week, and the risk is that a sick day or a busy launch pushes it back.

Fully automatic. The tool prepares and files as disputes arrive. Your weekly job is to read a short report of what went out, check any case marked weak or held and read two or three responses at random. That takes perhaps 15 minutes. The risk is that you stop reading the report and a drift goes unnoticed.

Hybrid. The six small disputes with strong evidence go out on their own. The two larger ones wait for you. You spend about 16 minutes reviewing them, add your own files if you have signed delivery paperwork or photos, and approve. You still read the weekly report. This takes about 30 minutes and protects the cases where a mistake costs the most.

Over a month, the hybrid routine saves about 2 hours against fully manual, and it keeps a human on every dispute large enough to matter. Most stores end up here.

How do you cover disputes while you are away?

Holidays and peak weeks are where Manual mode fails, because disputes keep arriving and their due dates do not pause. Plan for them before you leave.

Check the calendar for due dates. Look at every open dispute and note which fall inside your absence. The typical window is 7 to 21 days, so something opened today may come due while you are gone.
Clear what you can before you go. Review and file the open disputes in the days before you leave.
Loosen your automatic rules for the trip, not tighten them. If a dispute would otherwise miss its deadline, a reasonable automatic response beats a thin default. Raise your value cut-off for the week, but keep high-ticket orders on manual with a named backup.
Name a backup. A colleague, partner or VA with access to the approvals screen can cover the manual ones. Give them the decision rule in writing.
Set an alert for any dispute within three days of its due date. Make sure the notification goes to someone who will see it.
Review everything on your return. Read what went out automatically while you were away and note anything you would have done differently, because that updates your edit rate.

A plan like this turns the biggest weakness of Manual mode into something you can manage, and it makes automatic filing a tool you use on purpose and not a setting you forgot about.

Common mistakes with automatic and manual modes

Switching on automatic on day one. You have no data on how often the tool is wrong for your store.
Reviewing everything forever. Review that never changes a draft costs time and protects nothing.
Never looking at automatic filings. Drift goes unnoticed.
Comparing win rates across different case mixes. Easy cases make automatic look better than it is.
Fighting weak cases automatically. Effort and plan allowance go on cases that rarely win.
Treating refunds and declines like letters. They cannot be undone.
Ignoring deadlines in Manual mode. The default response is the weakest one.
Treating the formula's inputs as facts. They are estimates until you measure your own.

What we checked, and what you should check

We read the vendor pages and Shopify App Store listings for the tools named in this series, Shopify's help center pages on the chargeback process and ChargeFix's own plan and mode descriptions on 30 October 2026. The cut-off formula and its inputs are our own model, built to help you reason about the trade-off. They are not vendor figures or Shopify statistics.

What you should check: your own edit rate, your win rate by mode, your missed deadlines and your tool's exact automatic behaviour. Use the chargeback recovery calculator and the should I fight a chargeback calculator to test your cases.

Frequently asked questions

What is the difference between automatic and manual chargeback response?

In Manual mode a tool prepares the evidence and you approve each response before it is sent. In automatic mode the tool files without your approval, usually under rules you set. Either way, evidence is final once submitted, so what matters is who checks the response and when.

Is automatic chargeback filing safe?

It can be, for routine disputes, if the tool scores evidence honestly and you set rules and check results. It is riskier for large or unusual cases, because a bad response cannot be taken back. Start manual, measure how often you change drafts, then automate the safest slice.

What happens if I do not respond to a Shopify chargeback?

Shopify gives you a due date, typically 7 to 21 days after the dispute opens. If you submit nothing, Shopify sends the order data it holds on the due date, which is a thin response with a low chance of winning. After that, the bank decides, which can take up to 75 days.

When should I use Manual mode?

For your first month with a tool, for high-ticket orders, for weak or unusual cases, and whenever you want to add your own files. Manual review is worth the time when the dollars at stake are above a cut-off you can compute from your own edit rate, often around $100 to $250.

When should I use automatic mode?

For small, routine disputes where the tool's evidence score is strong and your own edit rate is low. Keep high-value and weak cases manual, sample automatic filings every month and switch back if the results slip. A hybrid policy suits most stores after the first month.

Does automatic mode use up my plan allowance faster?

It can, because every dispute the tool handles counts. In ChargeFix, a dispute counts as one filing the first time its evidence is saved or filed, and usage resets on the 1st. Automatic rules that skip weak, low-value cases help you protect your allowance.

Can I switch from automatic back to manual?

Usually yes, and you should know how before you start. Ask the vendor whether the switch is instant, whether it affects disputes in progress and whether past filings can be exported. A policy is a setting, so treat automatic mode as something you can tighten, loosen or turn off.

Is an auto refund the same as automatic filing?

No. Filing sends a letter that might win. A refund returns money and cannot be recalled, and an automatic decline turns away a sale. Treat anything that moves money or blocks orders as a separate decision, and test it in a recommend-only mode before letting it act.

How do I know if automatic mode is working?

Track your edit rate, win rate by mode, missed deadlines and net recovered per dispute. Compare automatic and manual on the same order-value band and score, because automatic cases are usually the easy ones. Read a random sample of automatic filings each month as if you were the bank.

Start free trial with ChargeFix

If you want a tool that starts in Manual mode and lets you move to automatic filing only when your own results say it is safe, ChargeFix starts free and keeps manual as the default. Automatic mode is available on Starter and above, and it never takes a cut of what you win. It checks your evidence item by item, gives each case an evidence strength score and a worth-fighting verdict, and drafts a letter from your real records. Nothing goes to the bank until you say so while you are in Manual mode, and paid plans include a 14-day free trial.

Start free trial

Automatic vs manual chargeback response on Shopify: when to let the tool file and when to review