How to reduce support tickets as a founder

Most of your support tickets are a few questions asked over and over in different words. To reduce support tickets, find out which questions those are, then fix each one where it starts. That might be the product, the page where people got stuck, or an answer you write once and give wherever customers ask. Done in that order, your tickets fall and you’re no harder to reach.

Deacon is an AI customer support agent for founders who answer support themselves. It answers your customers in seconds from your own docs, so you stop typing the same replies. It also groups everything they ask, so you can see what they’re stuck on and what they want next. There’s a free plan, and it needs no card.

Count your questions before you change anything

Every fix depends on knowing which questions cost you the most, so start with a tally. Take the last month of support email and chat, or the last 60 messages if a month gives you fewer. Give each recurring question one line in a text file, in the customer’s own words, and add a mark each time it comes back.

It takes under an hour. A few lines usually collect most of the marks, and the rest are one-offs. One line will be the question you already knew about. Another will surprise you, and it’s often something small you could fix in an afternoon.

Label each line as you go, because a how-to question, a bug report, a billing problem and a feature ask each need a different fix. Customer insights covers what each kind tells you about your product. Keep questions from people who haven’t bought yet on a list of their own. Those are sales conversations, so you want them answered faster, not fewer of them.

Fix each question where it starts

Each recurring question has a best fix. They’re listed here from the fix that removes a question for good to the one that only answers it.

Change the product

If 14 people ask where the export is, the export button is in the wrong place. Moving it is the only fix that removes the question instead of answering it, and a question nobody asks costs you nothing for the life of the product. Error messages count too. An error that says what went wrong and what to do next saves a ticket every time it appears.

Fix the page that failed

A question from someone already on your pricing page means the pricing page didn’t say it. Read it as a stranger would. What’s missing is usually one sentence, and it belongs where they were already looking. The same goes for your sign-up form, your welcome email and your receipts.

Tell customers before they have to ask

A lot of tickets come from surprises, such as a price change, a setting that moved, an outage or a renewal nobody remembered. A short email or a note in your app before the change saves a message from everyone it would have caught out. Customer service automation examples shows which of those emails Stripe and a status page can send for you.

Write down what your docs don’t say

Some questions are fair ones that no page answers yet. Write each answer once, as its own help page or entry, under a heading in the words customers used. Put the answer in the first sentence and the detail after it. For a handful of short answers, an FAQ page does the job.

Answer the rest where people ask

Even with good docs, people ask mid-task and won’t go hunting through a help centre. A chat on your site that answers from your docs catches those questions as they come up, at any hour. Pick one that answers only from your content and says when it doesn’t know, because a bot that invents a refund policy makes more work than it saves. Why support bots make things up explains what separates the two.

Here’s how an invented month at a small SaaS might come out, with one fix for each question.

Helpdesk automation sets out what to automate next, and in what order.

Three messages you should never deflect

Ticket deflection means answering a question before it becomes a ticket, and past a certain point it turns into a wall. Three kinds of message should always reach a person, so put that human in the loop before you automate anything.

Money comes first. Refunds, double charges, failed cards and plans someone can’t cancel all need you, because a wrong answer can cost you a chargeback and the customer is already annoyed.

Bug reports come second. Someone who reports a bug is doing you a favour, and a canned reply teaches them not to bother next time.

Cancellations come third, and not so you can argue. The reason someone leaves is worth more than another month of their subscription, and a cancellation that only ever meets a bot leaves you with no reason at all.

Whatever you use, make the way to a person easy to find. Put an email address on your contact page, and make sure your chat hands over when someone asks for a human.

How Deacon reduces support tickets

Deacon does the counting and the answering for you, on your site, with nobody watching the chat.

It keeps the tally. Deacon groups the questions people ask into topics, ranks them by how many people asked, and labels most of them as a how-to, a bug report, a feature ask, a pricing question and so on. Each topic also shows how many of its questions your docs didn’t cover. Each conversation shows the page it was asked from, and once one page’s questions go unanswered more often than the rest, Analytics › Widget names it, which points you to the page that failed. The feature asks tell you what to build next.

It answers what your docs cover in seconds, at any hour and in the language your customer writes in. When your docs don’t cover a question, it says it doesn’t know instead of guessing, and asks for the visitor’s email. If they leave it or ask for a person, you get an email. Reply from the dashboard, and if they still have the chat open, your reply appears there. That’s human handoff, and there are no flows or decision trees to build first.

It gets better every time you answer. Each question your docs didn’t cover has an Answer this button in the dashboard. Write the answer once and it becomes part of what Deacon knows, so it can answer the next person who asks. Deacon also tells you how many other open questions your answer covers. The more you answer, the less it needs you.

The three messages that should reach you don’t get lost. Deacon never answers a question about a visitor’s own account, invoice or data, because it can’t see them. It declines and asks for their email instead, so a billing problem reaches you rather than getting a guess. Bug reports gather in topics of their own, with a count. And every conversation is saved word for word and searchable by what visitors asked, so when someone tells Deacon why they’re leaving, you can read it in their words.

The price doesn’t grow with your questions. Some automated customer service tools bill for each question they resolve, so the bill rises as the tool does its job. Deacon’s plans have a set price, nothing is billed past a plan’s answers and no plan upgrades itself. Every feature is on every plan, including the free one, which needs no card and has no end date.

How to tell it’s working

Three signs tell you whether the fixes are working.

Your top question changes. If it does, the fix worked and something else is now your biggest problem. If the same question is still on top three months later, the fix never reached the people asking.

The questions your docs can’t answer get fewer as you write the missing answers. When that count climbs again, you’ve usually shipped something and not written it down.

Tickets per 100 active customers fall, even when your support ticket volume rises. A growing product gets more questions, so compare them with the customers you have. The glossary explains how ticket volume is counted.

In Deacon, Topics puts your top question first for whichever range you pick, and the Conversations page counts the gaps left to answer.

Questions about reducing support tickets

What is ticket deflection?

It’s a question answered before it reaches a person, by a help page, a chat or a screen that explains itself. It’s good when the customer got what they needed and bad when they gave up. The deflection rate entry shows how it’s measured.

Will an AI chatbot reduce support tickets?

Yes, for the questions your docs already answer, because it answers them in seconds at any hour. It can’t fix a confusing product or answer what nobody has written down. Deacon helps with the second by keeping every question your docs didn’t cover, each with an Answer this button, so you write each answer once.

Can Deacon issue refunds or cancel plans?

No. Deacon answers questions and hands conversations to you, and it takes no actions in other systems. It won’t issue a refund, cancel a plan or look up an order, and it declines any question about a visitor’s own account. For money, that’s what you want.

What happens when a month’s answers run out?

Visitors get a short reply saying Deacon can’t look into that one right now, and are asked for their email so you can reply, unless they’ve already left it. Nothing is billed past your plan’s answers, and you can still read and reply to every conversation.