Clients rarely complain that a firm is slow. They complain that they don’t know. The message went out on Tuesday and nothing came back. By Thursday the client is guessing. Is the work under way? Did the message arrive at all? Is the polite thing now to call? A response time policy is your answer to that gap. It works when it promises the thing the client is actually missing, which is confirmation and a date. A finished answer by Wednesday is the thing no practice can promise in advance.
- The number you publish covers the acknowledgment, not the answer. The first is a staffing decision you control. The second depends on work nobody has looked at yet.
- Set the number from your worst ordinary week rather than your best day. A policy gets tested on bad weeks.
- A tax practice’s year has a shape, so one flat promise breaks in the same four weeks every year. Publish the busy-season number in advance instead of apologizing for it afterward.
- A published number you miss costs you more than no number at all. You’ve told the client exactly when to start worrying.
What the number is promising
Two clocks run on every client message. Most firms publish the wrong one.
The first is the acknowledgment clock. It stops when a person at the firm has confirmed the message arrived, said what they understood it to be, and given a date for the real reply. That reply takes a couple of minutes to write. It needs no research, no file and no partner.
The second is the resolution clock. It stops when the question is answered, the return is reviewed, the notice is dealt with. How long it runs depends on facts nobody has looked at yet, on whose desk it lands, and on what week of the year it is.
When a firm writes “we respond to all client emails within 24 hours” on its website, it has almost always published the second number by accident. That promise commits the practice to finishing unknown work on a fixed schedule. It gets kept in November and broken every day from the middle of March. The breaking is arithmetic.
The acknowledgment number is a different kind of commitment. One named person can hold it in the third week of March, in the time it takes to read what arrived. And it’s the number the client was asking for anyway. Tell a client on Tuesday morning that their question arrived, that it needs half an hour with their figures, and that the answer comes by the 14th. They experience an eight-day wait as service. A client told nothing and answered on Thursday experiences a two-day wait as neglect. What they’re grading is the silence.
Two things follow from that.
The first is that an acknowledgment carries a date. “Thanks, we’ll get back to you shortly” is a politeness, and once the client has had two of them it reads as the sound silence makes. The reply has to name a day. Naming one gives the client something they can act on, which is the standing to tell you it’s too late.
The second is that the acknowledgment is where you find out whether you have a problem. The question that comes back is when the client has to have decided, and their date is frequently earlier than yours. Finding that out on Tuesday costs one line. Finding it out on the 14th costs the engagement.
A note on wording, before any of this goes near a signed document. “We aim to” and “we will” are different commitments, and an engagement letter is where the difference gets tested. A response time written into the terms is a term. Whoever drafts your engagement letters should see the sentence before it goes in. Engagement letters by email covers the rest of that document. The point here is that the response promise becomes a clause the moment it’s in there, and a clause gets held to.
Picking a number the third week of March can survive
Your firm already has a number. It’s sitting in the sent folder, and you can find it in about twenty minutes.
Take one ordinary week from last fall and one week from the middle of the season. For twenty threads in each, find the client’s message and your firm’s first reply. Write down the gap in business hours.
Three numbers come out of that, and two of them matter. The median is the third, and it will flatter you, usually a few hours. What matters is the worst case in each week and how many of the twenty sit near it. That tail is your real policy. It’s the only part of the distribution a client ever writes a complaint about.
Then set the published number above your peak-week tail rather than above your median. A promise you keep 95 percent of the time in April is a promise. A promise you keep 95 percent of the time in November and 60 percent in April is a schedule of disappointments with a date on it.
For most firms of five to fifteen people, one shape holds all year. Confirmation by the end of the next business day, with that first reply carrying either the answer or the date the answer is coming. It’s plain, and one person on the mailbox each morning can keep it. Same business day is achievable too, and it’s worth pricing before you publish it. Somebody reads client mail at least twice a day, every day. That includes the days the owner is at a client site. It means Friday afternoon has an owner. Same business day is a staffing arrangement before it’s a sentence.
Four details decide whether the number survives contact with a real week.
Say business days, and say the hours. One business day needs a starting line. A message that lands at 4:50 on Friday is Monday’s, and the client should be able to work that out from the policy itself. If your practice works across time zones, name yours. That’s one line, and it settles most of the arguments a response time policy ever generates.
Publish one number. Two-tier promises are tempting, because some clients genuinely do warrant a faster path. They fail for a mechanical reason. A tier you can’t see in the inbox is a tier you’ll miss. If the person opening the mailbox at 8:15 has to look something up to tell which promise applies to the message in front of them, there’s one promise, and it’s the slower one. A tier is real when the mailbox itself shows it.
Decide what the number covers. A response time policy is about client messages that ask something. Newsletters sit outside it. So do automated notices from a software vendor, and every message in a thread of forty. Saying so in the policy costs half a sentence and keeps the policy from dying of its own scope.
Make the first reply cheap. If acknowledging a message requires knowing the answer, the people best placed to keep the number are the ones who’ll miss it. The acknowledgment says what arrived, what it appears to be, what happens next and when. Somebody who has never seen the file can write it from what’s on the screen.
The year this promise has to live in
A tax practice has a season, and a flat policy is a promise written in November about the workload of a person in April.
The shape is familiar. Information returns and the January estimate payment open the year. Mid-March is the deadline for partnerships and S corporations. April 15 covers individuals and calendar-year C corporations. Quarterly estimates run through the summer. September and October take the returns that went on extension. Those dates move when they fall on a weekend or a holiday, which is worth knowing when you name them in writing. Around each one there’s a two to three week ramp where inbound volume roughly doubles and available hours stay exactly where they were.
Two designs work, and one spends credibility on schedule.
One conservative number, all year. Confirmation within two business days, every week of the year. This is the simplest thing to publish and the easiest to staff. It under-promises for eight months of the year in exchange for being right in the other four. For a small firm too small to run a seasonal shift, it’s the honest choice.
A published seasonal band. Confirmation by the end of the next business day for most of the year, and two business days between named dates. This is better where you can run it, for a reason that sits outside the numbers. The change is announced rather than experienced. A client told in January that March would be slower, who then finds March slower, has had an accurate firm. The same client, told nothing and answered in three days on March 20th, has had a firm that let them down. Both firms behaved identically.
The same number all year plus an apology every spring. This is what most practices actually run, and it’s the design that spends credibility on schedule.
Two details make the seasonal version work. Announce the peak window with the organizer, or whatever your first message of the season is. Announced on the day it starts, it reads as an excuse, and announced in reply to somebody already waiting it reads worse. And give the window a tail past the deadline. The week after April 15 is where the backlog lands, so a policy that returns to next-day confirmation on April 16 breaks on April 16. Inbox zero during busy season covers what to do with the pile itself. The point here is that the promise about the pile gets made in January.
One fast path, or four slow ones
Publish an email response time on its own, and the client who wants faster will find another door. They’ll call, then text, then leave something in the portal, then call your cell. That’s a reasonable thing for them to do. They’re choosing a speed, and you’ve priced one of them.
So the policy says what each channel is for.
Email is the channel of record, and the published number attaches to it. Say that plainly. It’s also the answer to a question clients rarely ask out loud, which is where to write so the firm has it.
The phone needs its own sentence. Either a person answers during stated hours, or messages are returned within the same window as email. Either one is fine. An undefined phone turns a client who was content into a client who is now testing you.
Text is a decision. It’s the fastest channel and the worst one for anything you might need eighteen months from now, and it arrives on a device where it gets answered standing in a checkout line. Decide whether you accept it. If you do, decide what it’s for, and the usual answer is scheduling and nothing else.
The portal needs to be honest about messages. Many document portals have a comment box, and in a great many firms nobody is watching it. If that’s true of yours, say the portal is for documents and the messages go to email. If somebody does watch it, say who, and say how often.
That leaves the auto-reply, which is the first thing most firms reach for and is worth a minute’s thought.
A permanent automatic acknowledgment on every inbound message is usually a mistake. Clients stop reading it inside a month. It fires on newsletters and on replies to your own messages. It takes the one moment where a human sentence was worth something and spends it on a form letter. Worse, it quietly satisfies the policy on paper while doing nothing for the client, so the firm stops noticing that nobody has actually looked.
Something narrower works. A seasonal auto-reply with real dates in it, running only during the named peak window, saying something the client can use. That mail is being read. That confirmation is taking two business days at the moment. And what to do if there’s a date involved.
An away message is a different thing again. A good one has three parts and no apology. When you’re back. Who to write to in the meantime, by name and address. And what actually justifies using that name. A substitute needs access to the mailbox as well as a mention in the message, and clients find out which one they got faster than firms do.
The mail that arrived with its own deadline
Some client mail is carrying a clock your policy can’t move.
An IRS notice usually prints a response window on the first page, often 30 days from the date of the letter rather than from the date the client got around to forwarding it. A statutory notice of deficiency gives a taxpayer in the United States 90 days to petition the Tax Court, and that window is fixed for everyone, your firm included. State agencies run their own windows. Then there’s the mail with no agency behind it at all. A lender who needs a return before a closing date. A buyer’s accountant with a diligence deadline. A payroll filing that runs ahead of your ordinary rotation.
Every one of these arrives looking exactly like ordinary mail. The dangerous version is a forwarded PDF with no covering note, where the only date in the entire message sits inside the attachment. The deadlines that hide in ordinary client mail is the guide to spotting them. What belongs in your policy is the escape hatch, and the design problem is defining urgent so it stays rare.
Define it by the object. “Call us if it is urgent” fails immediately. Everything is urgent to the person holding it, and the phrase asks a client to rate their own anxiety about their own money, which is a thing nobody can do.
Something in this shape. If you’ve had a letter from the IRS or a state agency, or if someone outside has given you a date, forward it and put the date in the subject line. A client can apply that test without judgment. It names things that exist rather than states of mind. It produces the one fact you need, in a place you can see it. And it stays clear of the question about whether to buy the truck before year end, which is the whole point.
One thing is worth saying inside the firm rather than to clients. An urgent path stays credible while it stays faster. If the flagged messages sit in the same queue as everything else, clients learn that within about two rounds, and then the flag is on everything. Whoever owns the mailbox in the morning owns the flagged ones first. Important versus urgent sets out the general version of that weighing.
Where the promise gets written, and in what words
Four places, in the order they do work.
The onboarding message. This is the one that matters, because it’s read. The engagement letter is signed and then sits in a folder. The welcome email is read closely, once, by a client who is paying attention to how this firm operates. Put the policy there in plain language. Onboarding a client by email covers everything else that message has to carry.
Something in this shape, adjusted to your own numbers:
How we handle email. Write to accounts@example.com and somebody will confirm we have it by the end of the next business day. If your question needs work, that first reply tells you when to expect the answer. Between March 1 and April 22, confirmation can take two business days. If you’ve had a letter from the IRS or a state agency, or somebody outside has given you a date, forward it and put the date in the subject line, and we’ll treat it as urgent. The phone is 555-0100, and messages are returned inside the same window. Documents go in the portal. Messages go to email, because that’s where we’ll see them.
The engagement letter, in the firm’s own contract language, saying the same thing the onboarding message says. If the two disagree, the client will remember the friendly one and you’ll be held to the signed one.
The email signature. One line, on every message the firm sends: “We confirm client email by the end of the next business day, and within two business days between March 1 and April 22.” It’s the cheapest reminder in the practice. It’s seen hundreds of times a year, and it does more to reduce “just checking you got this” mail than any other single change.
The contact page, if you have a website that clients ever land on. Same sentence.
Then there’s the harder case, which is the clients you already have. Their expectations come from whatever you happened to do for the last four years, and those are usually faster than anything you’re about to publish. Reset them once, in an ordinary message, and follow three rules. Give it a start date, so it’s a policy and not a mood. Frame it as what the client gets, which is confirmation and a date every time. And leave out the paragraph about how busy the firm has become. It tells the client their accountant is overloaded, and that turns a piece of operational news into a warning sign. Keeping the warmth when you are under pressure is where the tone of that message belongs.
One more thing belongs in the same note, and it’s easy to leave out because it’s uncomfortable. The clock runs both ways. A policy that commits your firm to a turnaround and stays quiet about what you need from the client is half a policy, and it’s the half that produces April. Say what you need and by when, in the same message. The mechanics of getting it are a separate job, and they live in following up on a missing document and writing a deadline reminder that works.
Keeping it on the week that goes wrong
A response time policy is an operating arrangement, and it fails in the same four places every time. Each of them has a name and a fix.
Nobody owns the mailbox. If the answer to “who reads client mail before 9am” is “whoever gets in”, the answer on the bad day is nobody. Put a name against each morning, including the mornings the owner is at a client site or on a plane. Make sure that name has access as well as good intentions. Running a firm inbox is the fuller version of this, and covering for each other is the version for a practice too small to keep a roster.
The person acknowledging is the person answering. Keep those two jobs apart. Confirming that a message arrived and saying when the answer is coming takes the screen and nothing else. Split them, and the first job stops waiting behind the second, which is where nearly all missed windows come from.
The cutoff is undefined. Decide what time of day closes the window, write it in the policy, and hold it. Mail after the cutoff is tomorrow’s, and a client who knows that is settled.
Nobody checks. A policy that goes unmeasured drifts within a quarter, and the firm hears about it from a client. Measuring it takes twenty minutes and no software. Pull twenty client threads at random from last month, find the first reply, look at the gap. Do it once a quarter and once in the middle of the season, and read the worst few rather than the average.
Two situations need a decision rather than more effort.
When you miss the window on a specific message, deal with it inside the reply, briefly, and leave the season out of it. When you miss it on something that matters, the miss needs a message of its own, and telling a client about a delay is the guide to writing it.
When you miss it every week, the number is the thing to change. A number a well-run firm can’t hold is the wrong number or the wrong staffing. Change one of those and republish. It costs one email, and it buys back the credibility of every other thing the firm has promised.
What Point does with the clock
Firms kept response promises on a call sheet and a wall chart long before any of this existed. Software only changes who does the noticing, what holds the promise once it’s made, and how much of the reply you type. Choosing the number, staffing the morning and judging whether a client’s date is real all stay where they are.
Point works through the mailbox ahead of you. Point ranks what needs you today against what can wait, so a client sitting inside your published window stays above a software renewal at 8:15. How triage decides what needs you explains the weighing. Each thread carries a plain summary in a line, and that’s what makes an acknowledgment cheap. The person keeping the window sees what arrived from the line itself, rather than reading four paragraphs and two attachments to find out.
The part that fits this piece most directly is what happens to the promise once you’ve made it. An ask sitting in an email turns into a dated item on its own, with nobody retyping it. The date you named in the acknowledgment is the date it comes back to you. So the commitment lives inside the conversation, rather than on a list beside a thread that looks finished because somebody read it. Open loops come back before they go cold, which covers the other half: the messages where you’re the one waiting.
Replies come back drafted in your own voice, so the acknowledgment is an edit rather than a composition, and the send is yours. A forwarded notice can be asked directly what date is on it, and the answer points back to the line it was read from. That’s the practical repair for a deadline that lives inside a PDF and nowhere in the sentence above it. A standing watch, set in plain words, covers the mail that has to move today. A short briefing gathers what needs you today, and a digest after a day away covers the morning nobody was rostered.
How far Point goes on its own is set for each kind of action separately, and it starts at review out of the box. Point prepares the work and holds it for you. The autonomy dial explains that properly, and this is a subject where the line is worth drawing carefully. Acknowledgment is the class worth raising, because it’s the cheap half of the promise and the half that gets missed. An answer that turns on facts the message doesn’t contain stays at review, at any setting. Whatever is done for you lands in a log, in order and with times against it, and you can approve, refuse or reverse it from the line that recorded it. One thing sits outside that, as it does in every mail product ever built. A message that has already reached the other side stays sent.
You start by signing in to the Gmail or Microsoft 365 mailbox the firm already has. Nothing migrates, and clients go on writing to the same address.
What your firm should promise stays your call. Point doesn’t know your staffing, can’t tell you whether two business days is honest for your April, and holds no opinion on whether a client’s stated deadline is a real one. That judgment is the actual work of this subject. The benefits page is the full inventory, and AI email for accountants takes the buying question apart for a practice. If the question you’re holding is really about whether any tool should be reading client mail, that one comes first, and is it safe to use AI with client financial data is where it gets settled.
Common questions
What is a reasonable email response time for an accounting firm?
For the acknowledgment, the end of the next business day is reasonable. Most firms of five to fifteen people can hold it, and it ends almost all of the anxiety a client feels while waiting. The answer itself has no reasonable universal number, because it depends on work nobody has scoped yet, which is why the answer should carry a date. If you want same business day, price it first. Client mail gets read at least twice a day, every day, including the days the owner is out. Friday afternoon has an owner by name.
Should the response time go in the engagement letter?
Yes, and it should also go somewhere the client will actually read, because the engagement letter is signed once and then sits in a folder. The wording matters more in the letter than anywhere else. “We aim to” and “we will” are different commitments, and whoever drafts your engagement letters should see the sentence before it goes in. Make sure the letter and the friendly onboarding version say the same thing. When they disagree, the client remembers the friendly one and the firm is held to the signed one.
Should we set up an automatic reply confirming we received an email?
A permanent one on every inbound message is usually a mistake. Clients stop reading it within a month. It fires on newsletters and on replies to your own messages, and it can quietly satisfy the policy on paper while nobody has actually looked at anything. A seasonal auto-reply is different and does work. Run it only during your named busy window. Put real dates in it. Say that confirmation is taking two business days at the moment, and say what to do if there’s a deadline involved. An away message is different again, and it needs three things and no apology: when you’re back, who to write to by name and address, and what justifies using that name.
How do we tell clients that busy season is slower without sounding like we are making excuses?
Timing does almost all of the work. Announce the busy-season window before it starts, with the organizer or whatever your first message of the season is, and name the dates it runs between. Told in January, it’s a firm that plans. Told on March 20th to somebody who has already been waiting four days, the identical sentence is an excuse. Keep the reason short or leave it out. Give the window a tail past the deadline, so April 16 finds you still keeping it. And describe what the client still gets, which is confirmation and a date every time.
A client expects an answer within the hour. What do we do?
Answer the acknowledgment. Reply quickly with what you have, what you understood the question to be, and the date the real answer is coming, then ask when they need to have decided. Most of the time that date is later than the hour, and the expectation resolves itself without a conversation about policy. Where it genuinely is that urgent, it’s usually the urgent path in disguise: somebody outside has given them a date. Where a client is like this on everything, the container is the problem rather than the response time. A standing slot of twenty minutes a month is easier to offer than a bill, and it usually reads to the client as an upgrade. Stopping client questions from piling up works that through.
How do we know whether we are actually meeting the number?
Pull twenty client threads at random from last month. Find the client’s message and your firm’s first reply, and write down the gap in business hours. It takes twenty minutes and no tooling. Do it once a quarter and once in the middle of the season, and read the worst few rather than the average. The average flatters every practice, and no client ever complained about a median. If the tail sits well past your published number, the fix is a different number or a different morning assignment.
The short version
A response time policy is a promise about acknowledgment, and firms get into trouble by publishing a promise about answers instead. Find your real number by reading twenty threads out of last March’s sent folder and setting the published figure above the worst of them rather than above the median. Say business days and say the hours. Publish one number, because a tier you can’t see in the inbox is a tier you’ll miss. Give the peak weeks their own number, announced in January, with a tail that runs past the deadline. Name the channels, so the client who wants faster has one door you control. Define urgent by the object, because a letter with a printed deadline is a fact and a bad morning is a feeling. Write the policy into the onboarding message, the engagement letter and the email signature, and reset existing clients once with a start date and no complaint about how busy you are. Then staff it: a name against each morning, the acknowledgment split from the answer, a cutoff you hold, and a twenty-minute check every quarter. When you miss the number on one message, say so in the reply. When you miss it every week, change the number.