How OpenChair Handles Client Messages Across SMS, Email and Instagram
See which OpenChair replies can Auto-send on supported SMS and Instagram, which stay as drafts, and which booking actions need review.

On this page
A client asks about availability while the team is busy with appointments.
Another sends an Instagram DM asking which service to book.
An email arrives with a question that needs someone at the venue to make a decision.
Those messages can all reach OpenChair, but OpenChair does not treat every channel or every question the same way.
On Pro, eligible Smart Replies can Auto-send on supported SMS and connected Instagram conversations when the owner has enabled that authority and the reply meets the required confidence and safety rules.
Email replies stay as drafts.
Booking actions stay review-first.
That distinction is the useful starting point.
SMS, email and Instagram have different rules#
OpenChair's Inbox brings supported client conversations into one working area, but the automation available depends on the channel.
| Channel | What OpenChair can do |
|---|---|
| SMS | On Pro, eligible high-confidence replies can Auto-send through a dedicated Australian or UK number when configured. Other supported replies can remain Draft. |
| Instagram DMs | On Pro, eligible replies can Auto-send from a connected Instagram account where configured, or remain Draft for review. |
| AI-assisted replies remain draft-only. A person reviews and sends them. | |
| Booking actions | Proposed booking creation, rescheduling or cancellation remains behind operator review rather than being completed autonomously by Smart Replies. |
The owner controls which supported reply intents have permission to Auto-send.
Auto-send is therefore not a single switch that gives AI unrestricted control of the inbox.
SMS Auto-send requires a supported dedicated number#
OpenChair Pro supports dedicated SMS numbers in Australia and the United Kingdom.
When a venue has the supported SMS setup in place, Smart Replies can respond to eligible inbound messages according to the authority the owner has configured.
A routine question might be suitable for Auto-send if:
- the intent is one the venue has authorised;
- OpenChair has enough venue context to prepare the answer;
- the response meets the required confidence threshold;
- the applicable safety and validation rules pass;
- sending is allowed under the configured communication controls.
A message that does not meet those conditions can move to Draft or escalation instead.
SMS is not currently offered as an OpenChair channel in New Zealand, so the Australian and UK SMS behaviour should not be generalised to every OpenChair market.
Instagram can use the same authority model where connected#
A connected Instagram account can bring supported DMs into the OpenChair Inbox on Pro.
The owner can decide how eligible Smart Replies behave for supported intents.
That can mean:
Off: OpenChair does not prepare or send the Smart Reply for that intent.
Draft: OpenChair prepares the response, but a person reviews it before sending.
Auto-send: an eligible high-confidence response can send automatically when the configured rules are satisfied.
This can be useful for routine questions that arrive while the team is with clients.
For example, a client might ask:
Can I book a cut and blow dry online?
If the venue has authorised that supported intent and the answer is available from the venue's configured information, an eligible reply can point the client towards the appropriate booking route.
That does not mean OpenChair books the appointment on the client's behalf.
The client can continue through the booking flow, or the conversation can move back to the team when a decision needs human input.
Email stays draft-only#
Email has a different boundary.
OpenChair can prepare supported email reply drafts, but those replies do not Auto-send.
A person reviews the draft before it goes to the client.
This is useful when AI can do the first pass of the writing without the venue giving up control of the final response.
A draft can save someone from starting with an empty composer.
It can still be edited, rejected or replaced before anything is sent.
The distinction should be clear anywhere OpenChair describes automated messaging:
SMS and connected Instagram can Auto-send eligible replies on Pro when configured. Email remains draft-only.
Smart Replies and OpenChair Assistant are different workflows#
OpenChair Assistant is primarily for the operator.
It works with supported venue context to answer questions, prepare drafts and surface supported actions.
Smart Replies operates within supported client-message workflows.
They can use related information, but they do not have the same authority.
For example, an owner might ask Assistant for help preparing a response or understanding something about the venue.
An inbound SMS or Instagram message might instead enter the Smart Replies workflow, where the venue's configured intent rules determine whether the response stays as a Draft or is eligible to Auto-send.
Keeping those workflows separate matters because “AI can draft this” and “AI has authority to send this” are different permissions.
Booking Replies do not make booking autonomous#
Client messages often turn into booking requests.
Someone may want to:
- make an appointment;
- move an existing appointment;
- cancel;
- ask for another time;
- clarify which service to choose.
OpenChair can prepare replies and supported proposed booking actions, but booking actions remain review-first.
That means a Smart Reply being able to answer a routine question does not give it unrestricted authority to create, move or cancel appointments.
A reply can point a client towards a supported booking or management route.
Where an operator action is required, OpenChair keeps that action behind the relevant review step.
This boundary is intentional.
A message and a booking mutation are not the same thing.
Venue context makes a reply more useful#
A useful response needs information about the actual venue.
Depending on the supported workflow, OpenChair can use configured venue context such as services, booking information, policies and other relevant business information when preparing a reply.
That is more useful than asking a generic writing model to invent an answer from the client's message alone.
It also means the quality of the response depends partly on the information the venue keeps current.
If a service, policy or booking route changes, the underlying venue information should be updated rather than expecting an AI reply to know something that has never been recorded.
AI should work from the business information available to it, not fill factual gaps with guesses.
Low confidence should mean less authority, not more#
Not every message should be answered automatically.
A client may send an ambiguous request.
They may ask something outside the information OpenChair has.
The message may contain a complaint or something that requires the team to make a judgement.
In those cases, the useful behaviour is to reduce automation.
OpenChair's supported Smart Replies workflows include confidence, safety and escalation controls that can move a response away from Auto-send.
Complaints escalate rather than being treated as routine Auto-send conversations.
Low-confidence or safety-triggered messages can move to Draft or escalation depending on the workflow.
That means automation has an exit.
The system does not need to pretend it knows the answer to every message.
The owner decides which routine replies have authority#
Auto-send should reflect the venue's appetite for automation.
One salon may be comfortable automatically answering straightforward questions about booking links.
Another may want every client-facing response held for review.
A third may Auto-send a small set of routine intents while keeping everything else in Draft.
OpenChair lets the owner configure supported reply authority by intent.
That makes the useful setup exercise:
- Identify the questions the venue receives repeatedly.
- Decide which have stable, factual answers.
- Keep the underlying venue information current.
- Decide which supported intents can safely Auto-send.
- Leave judgement-heavy conversations in Draft or escalation.
- Review the setup when services, policies or team processes change.
The objective is not the highest possible automation rate.
It is choosing where automation removes repetitive work without taking a decision away from the person responsible for it.
An Instagram enquiry and a booking request are not the same thing#
Consider two messages.
The first client writes:
Can I book a cut and blow dry online?
That may be a straightforward enquiry. A supported reply can provide the venue's booking route if the configured rules allow it.
Another client writes:
I booked a cut tomorrow but I actually want a full colour. Can you change it to 2pm?
That request affects the booking itself.
Changing services or time can affect duration, staff availability, price and other booking conditions.
The reply and the booking change therefore belong to different layers of authority.
OpenChair can help prepare the response and proposed action, but the booking action remains review-first.
That is why describing OpenChair as an autonomous receptionist would be inaccurate.
A draft is still useful#
Automation does not have to mean Auto-send.
For messages that need review, a useful draft can still remove some of the repetitive work from the team's day.
The operator can open the conversation, see the proposed response, check it against the message and venue context, edit it if needed and send it.
That can be particularly useful during a busy day when the alternative is writing the same type of response from the beginning each time.
The human step remains visible.
The point is to shorten the work, not hide who is responsible for the final decision.
Keep client communication preferences in the workflow#
A reply system also has to respect the client's communication choices and the venue's channel setup.
OpenChair has channel-specific controls, opt-out handling and quiet-hour rules across supported communication workflows.
Those controls matter because “the system can generate a reply” does not automatically mean “the system should send it now”.
Transactional communication, marketing messages, reconnect activity and direct replies can have different purposes and controls.
When reviewing your messaging setup, check the actual workflow rather than assuming one permission applies to every type of communication.
Test Smart Replies with real questions from your venue#
Before enabling Auto-send for an eligible intent, test the questions clients genuinely ask.
Use fictional client details, then try examples such as:
- “Where can I book?”
- “Do you offer this service?”
- “Is there parking nearby?”
- “Can I book with Sarah?”
- “Can you move my appointment?”
- “I'm unhappy with my appointment.”
- an unclear message with missing information.
For each one, check:
- Does OpenChair understand the request?
- Is the venue information behind the answer current?
- Does the reply contain anything the system cannot substantiate?
- Does it stay in Draft when staff judgement is needed?
- Does a complaint escalate?
- Does a booking change remain behind review?
- Is Auto-send available only on the channel and intent you authorised?
Testing the difficult examples is more useful than testing only the questions you know the system can answer.
Integration history is not the same as reply authority#
OpenChair's late-April 2026 release record includes Instagram DM integration, booking-enquiry SMS in Inbox and improvements to conversation context. Later work continued to improve message handling.
That history should not be read as a claim that every reply permission described in this guide existed in April. Connecting a channel, drafting a response, Auto-sending an eligible reply and approving a booking action are separate capabilities.
For a lash and brow studio, a routine booking-link question may be suitable for a configured reply. A request to move an appointment still requires the relevant booking review. A complaint needs escalation rather than a routine sales response.
On Pro, eligible high-confidence Smart Replies can Auto-send through supported dedicated Australian or UK SMS numbers and connected Instagram DMs where configured. Email stays draft-only. Booking actions remain review-first, and confidence, validation, quiet hours and pause controls can limit reply authority.
Where the Inbox fits#
The Inbox gives the team one place to work with supported conversations rather than treating SMS, email and Instagram as three unrelated tasks.
The value is not that every channel becomes identical.
It is that the team can see the conversation and use the appropriate response workflow without forgetting which rules apply.
For Pro Smart Replies:
- eligible SMS replies can Auto-send through a dedicated Australian or UK number when configured;
- eligible connected Instagram replies can Auto-send when configured;
- both channels can use Draft where the owner wants review;
- email remains draft-only;
- complaints and low-confidence or safety-triggered messages can move to review or escalation;
- booking actions remain review-first.
That is the automation boundary OpenChair is designed around.
The repetitive reply can move faster where the rules are clear. The decisions that need the operator still come back to the operator.