The complaint that generates the most bar grievances is not bad lawyering. It is silence. A client signs, hears nothing for six weeks, calls twice, and leaves a one-star review that mentions "never returned my calls." Meanwhile the paralegal has been working the file every day.
Client communication automation is the practical fix for that gap. It is not a chatbot and it is not a newsletter. It is a small set of messages, sent at the right moment because something happened in your case management system, that tell the client where things stand without anyone having to remember to send them.
Which messages are worth automating, and which are not
Start by sorting your client communication into three piles. The first pile is procedural: welcome messages, portal invitations, document request reminders, appointment confirmations, and "we received your file" acknowledgments. Every one of these is safe to automate because the content does not change from client to client.
The second pile is status: the case moved from investigation to demand, the discovery deadline was met, the hearing date was set. These are safe to automate when the trigger is a real field change and the message says only what the field says. The third pile is judgment: settlement offers, bad news, strategy. Never automate these. A template that says "we have received an offer and your attorney will call you by Friday" is fine; a template that conveys the offer is not.
The status update problem: what clients actually want to hear
Clients do not need detail. They need to know three things: what phase the matter is in, what happens next, and whether they owe you anything (a document, a signature, a payment). A good automated status message answers those three in under a hundred words and ends with a single instruction or a single reassurance.
Write the templates yourself, in your own voice, before touching any software. Read each one aloud as if you were the client receiving it at 9 p.m. on a Tuesday. If it sounds like a system generated it, rewrite it. The automation should feel like a competent assistant who keeps in touch, not a shipping notification.
Building the triggers in Clio and Filevine
In Clio, the cleanest triggers live in two places. Clio Grow handles the pre-engagement sequence: when a lead moves to a new pipeline stage, an email template fires, and when the engagement letter is signed the contact and matter are created in Clio Manage. Our Clio Grow intake guide covers that setup step by step. After engagement, Clio Manage matter stages and custom fields become the source of truth, and a Zapier connection watching those fields can send the status message through your email or texting tool. That is the same pattern we use for client intake automation, just pointed at a later moment in the matter.
In Filevine, phases are the natural trigger. When a project moves from Treatment to Demand, a phase-triggered task can be assigned to the case manager with the message template attached, or an integration can send the message directly. Filevine's texting and client portal features keep the exchange in the project record, which matters when a client later claims they were never told.
Whichever platform you use, the trigger should be a field a human sets deliberately (a phase, a stage, a date field), not an inferred event like "no activity for 30 days." Inferred triggers are where automation starts sending the wrong message to the wrong client.
Texting, email, and portals: choosing the channel
Text messages get read. Email gets filed. Portals get ignored unless the client has been trained to use them. Use text for time-sensitive procedural messages (appointment tomorrow, signature needed today) and email for anything the client might want to find again. Reserve the portal for documents and for the record of what was sent.
Get written consent for texting during intake and store it as a field in the matter. Automated texts without consent are a compliance problem before they are a marketing problem.
Guardrails: confidentiality, tone, and the off switch
Three rules keep this safe. First, no automated message should contain facts about the case beyond the phase name and the next step; the substance stays in conversations and in the portal. Second, every sequence needs a per-matter off switch, usually a checkbox custom field, for the client who has asked for calls only or the matter that has gone sideways. Third, someone reviews the sent log monthly, because a broken trigger fails quietly.
Once the communication rhythm is reliable, the last message in the sequence is the natural place to ask for feedback. A closing-of-matter message that thanks the client and points to your Google review link is the single most effective tactic in getting more Google reviews, and it only works because the earlier messages built the goodwill. Firms that also want the intake conversation itself captured cleanly should look at AI client intake summaries, which feed the same fields these messages read from. All of this sits inside the broader law office automation work we do.
Questions we get
Will clients know the messages are automated?
Some will, and it does not hurt you if the messages are accurate and timely. What hurts is a message that is wrong, or one that arrives while the client is waiting on a promised call. Keep the templates honest about what they are ("a quick update from our office") and make replying easy.
Can we do this without Zapier or an integration platform?
Partly. Clio Grow email sequences and Filevine phase-triggered tasks run natively. Cross-tool messaging, such as sending a text when a Clio custom field changes, usually needs Zapier, Make, or Filevine Fusion. We start with the native tools and add an integration only when a specific message cannot be sent otherwise.
How many messages is too many?
A client should hear from the firm unprompted at every phase change and at least once a month in a long matter, and never more than once a week unless something is actively happening. If your sequence has more than eight templates per matter type, some of them are probably noise.
If you want to map your matter stages to a communication sequence, tell us what you are working with and we will show you where the triggers already exist in your system.