Running an inbox for two people without mixing voices

When somebody receives a reply from an address they wrote to, they are working out who they are now talking to, whether or not they think about it consciously. Is this the person I addressed? Someone speaking for them? Someone who will pass my message on? The answer changes how they read everything that follows, and it changes what they do next.

This is why the mechanics of a shared inbox matter beyond tidiness. A reply that leaves the question unanswered produces predictable friction: the sender assumes they have reached the owner, receives a considered answer, and later discovers the owner knew nothing about it. Nobody was dishonest — the assistant answered a question they were entitled to answer, and the ambiguity did the damage. Everything that follows is really one principle applied in different places: make it effortless for the reader to know who is writing. That is cheaper than it sounds, and it removes most of what goes wrong in shared correspondence, because almost every failure in this area traces back to a reader who guessed wrong about who they were dealing with.

The messages that can be handled independently are the ones where the answer is a fact rather than a position. Availability. Directions and logistics. Confirming that something was received. Sending a document that already exists. Standard information about how you work — payment terms, opening hours, what a process requires. Nothing here depends on the owner's opinion, so answering it is a service to everyone and routing it through the owner adds delay and nothing else.

Messages that need the owner are the ones where the reply expresses a view or creates an obligation: pricing outside a published list, agreeing to a deadline, responding to a complaint, anything about a relationship's direction. The boundary is not seniority of sender but whether the reply commits the owner to something. A junior person at a client asking for a discount is a commitment question; a senior person asking for the office address is not. Worth writing down explicitly, because the categories are stable and the cost of getting them wrong is asymmetric — an unnecessary forward costs a day, while an answer that commits the owner to something they would not have agreed to can cost considerably more.

There is a real distinction between writing on someone's behalf and writing as them, and the difference is disclosure rather than style. 'Priya has asked me to send you the revised schedule' is writing on her behalf: the reader knows exactly who is typing and whose authority stands behind it. The same message sent from her address in her voice with no indication of authorship is impersonation, even when she approved it, because the reader has been given a false impression about something they would care about.

The practical approach is to match the register without adopting the identity. If the owner writes briefly and without formality, write briefly and without formality — a jarring shift in tone from the same address is itself confusing. But keep the authorship visible in a way that requires no effort to notice, which normally means the first line or the signature rather than a note at the bottom nobody reads. Where you genuinely are relaying a decision, say so: 'Priya has confirmed we can hold the price until the end of the month' is clear about both who decided and who is typing, and it protects everyone if the detail turns out to be wrong.

The most reliable mechanism is structural rather than a matter of remembering to explain. A signature that names the actual sender and their relationship to the owner answers the reader's question every time, without depending on anyone's attention on a busy day: 'Anjali, assistant to Priya Nair' at the foot of every message she sends is enough, and it is enough permanently.

Where the mailbox is a shared one rather than an individual's, the same principle applies with the addition that consistency matters across the team — if one person signs personally and another does not, the reader cannot use the signature as a signal, and it stops being informative. A few details are worth getting right. Put the name near the top for anything where the reader might act on the message immediately, since the foot of the email is read last or not at all. Avoid the temptation to leave signatures off short replies, because those are precisely the ones most likely to be misread as coming from the owner. And where a message genuinely is from the owner, having it look visibly different from an assistant's message is what gives the distinction any value.

Escalation fails when it depends on judgement at the moment a difficult message arrives. What works is a short list of triggers recognisable without interpretation: anything mentioning legal action or a lawyer, anything about money above a stated threshold, anything from a named list of important contacts, anything expressing anger, and anything you have not seen before.

That last trigger is the one that carries the most weight and is the easiest to leave out. A rule that says 'escalate anything unfamiliar' covers the situations nobody anticipated, which is where the real risk sits — a rule set enumerating known categories will always be incomplete. Escalation should also mean a specific action rather than a vague hand-off: forwarded with a one-line summary of what is being asked and by when, so the owner can act without reading a thread from the beginning. And it needs a holding reply to the sender, because escalation without acknowledgement produces silence at exactly the moment the sender is most anxious. 'I'm passing this to Priya today and she'll come back to you' costs nothing and prevents the follow-up that otherwise arrives, usually more sharply worded.

The question underneath all of this is which messages can create an obligation, and it deserves treating carefully rather than confidently. In ordinary practice a business is likely to be held to what its representatives appear authorised to agree, which means an assistant's confirmation of a price or a date may well be relied upon by the other side regardless of internal rules. Whether a particular exchange creates a binding commitment in a specific case is a question for a lawyer who can look at the facts, and it is not something to settle from a general article or a rule of thumb.

What follows practically, though, is clear enough without resolving the legal question. Anything that looks like agreement to a price, a deadline, a scope of work or a contractual term should come from the person with actual authority, or be explicitly relayed as their decision. It is also worth the owner stating the internal boundary in writing, so that both people can point at the same understanding later. That does not settle what a third party may be entitled to rely on, and pretending otherwise would be the mistake. It does remove the ambiguity between the two people sharing the inbox, which is the part within your control.