AI reads client and lead activity, prioritizes who needs attention and drafts personalized messages for your team and automations.
01
Maya reads the data, so you don't have to.
Review.
Categorize.
Draft.
Every open, reply, visit and invoice already lives in Lattice. An AI agent reads all of it and tells you who needs attention and what to say.
Maya reviews client activity, sorts clients into lists such as "likely to cancel appointment", "likely to pay" and "needs more follow up", then drafts a personal message for each one. Your automations send it, or you call.
Step 1. The data Maya reads
Client
Last activity
Visit history
Invoice
Next appointment
Dana R.
Opened reminder, no reply
Rescheduled twice
Paid on time
Thu 10:00
Marcus T.
Viewed quote 4 times
First visit
Quote sent, unpaid
None booked
Priya S.
Replied "see you then"
6 visits this year
Paid on time
Fri 2:30
Kevin L.
No opens in 21 days
Missed last visit
Overdue 12 days
Mon 9:00
Alicia M.
Clicked pricing page
Consult done
Deposit pending
None booked
Jon B.
Asked about weekend hours
2 visits
Paid on time
None booked
Step 2. The lists Maya builds
Likely to cancel appointment
2 clients
Dana R.
Kevin L.
Repeated reschedules, unopened reminders and a missed last visit.
Likely to pay
2 clients
Marcus T.
Alicia M.
Repeat quote and pricing views right after a consult.
Needs more follow up
2 clients
Jon B.
Kevin L.
Open question with no answer, or silence after an overdue invoice.
Healthy
1 client
Priya S.
Confirmed, paid and booked. No action needed.
Step 3. The result: drafted follow up
To Dana R. · Likely to cancel · Sent by automation, Wed 6:00 PM
"Hi Dana, just checking that Thursday at 10 still works. If the week got busy, reply with a better time and I'll hold it for you."
To Marcus T. · Likely to pay · Sent by automation, today
"Hi Marcus, I saw you had a look at the quote. Happy to walk through any line on it, or I can lock in the start date now."
For Lorenzo · Needs more follow up · Call task
Call Kevin L. today. Missed his last visit and his invoice is 12 days overdue. Suggested opener: ask how he's doing before mentioning the balance.
Example data for illustration. Maya drafts, your automations send, and you make the calls only a human should make.
Builds the relationship and makes the judgment calls.
Reads client activity, spots patterns and drafts the next message.
Sends the agreed message at the agreed moment.
What it reads
The client in front of them.
Opens, replies, visits, quotes, invoices and appointments.
The trigger you set: a booking, a reply, a missed visit, an overdue invoice.
What it produces
Decisions, calls and the tone of the business.
Client lists such as likely to cancel, likely to pay and needs more follow up, plus drafted follow up.
Bookings, reminders, follow ups and review requests, sent on time.
Who approves it
You, always.
You review the drafts, or you approve a list once and it runs.
You set the rules once, then it runs without editing.
When it works alone
Never. That is the point of the other two.
Rarely. A draft still needs a person or a rule to send it.
Yes, for anything already agreed: confirmations, reminders, review asks.
Each case study below follows the same shape: what the human did, what the AI found, and what the automation sent.
Case study 01
Appointments: getting clients to show up
Human
Manual call actions. Nancy, the secretary, has booked appointments for years. Clients love her and she loves them, and her role is not changing.
AI
Runs a report on all upcoming appointments and sorts clients by priority: A, likely to cancel (no contact); B, 50/50 chance (keep warm and active); C, likely to show (spend less time on these). Then uses the universal speech prompt to engage clients with personalized, drafted outreach based on their profiles. The only goal: showed appointments.
Automation
Appointment booked, then wait until a set time before the appointment. If confirmed: wait. If rescheduled: wait. If cancelled: send a reschedule link.
Case study 02
Upsells and cross-sells: the offer after the appointment
Human
Reaches out manually by text after the appointment to make an offer. Before Lattice there was no system in place at all.
AI
Drafts a personalized outreach message from the client's Lattice profile and finds the best upsell product for them. The solution: AI and automation connected, built on a human language prompt.
Automation
When an appointment is marked complete, send a text and email to the client. If Do Not Disturb is on for both, create a task to call the client in one day. If they reply, notify the business owner. If there is no reply after 3 days, send the owner a task to call the client.
Case study 03
Sales: from cold calls to postcards that convert
Human
Before: outbound calling to scraped phone numbers that annoyed people, plus networking events and direct mail campaigns.
AI
Finds home addresses in the local city to mail, drafts postcards per city and zone, then fires the automation. Afterward it tracks conversion rates and learns which cities, populations and postcard designs convert best.
Automation
Lead added to the pipeline. When the stage moves to send postcard, the postcard goes out. Wait 10 days: no response or profile activity moves the lead to cold call, a response moves it to hot leads. Leads move through the sales process with conversion rates marked at each step.
Case study 04
Retargeting and repeat buyers: the offer after the roof
Human
Re-approaches previous clients for a new offer. Sets the discount code that is applied at checkout, then books the appointment and does the work once the client buys.
AI
An AI agent is already in place to draft outreach. It reads the database for likely targets: clients who are in touch, responsive, opening emails, or opted into email and text campaigns. It drafts personalized outreach to those valued clients. For the rest, it finds the criteria for the offer and adds them to a unique automation that targets those specific audiences.
Automation
Waits 30 days after the roof is serviced and asks if window washing is of interest, with an exclusive offer for clients who replaced roofs with us. When the client checks out, they book an appointment and the human does the work.
Case study 05
Marketing my business: from agencies to a working system
Human
Posts on social media, pays for Google ads, and hires a web developer through networking or searching, or builds the site on their own. That part is up to the owner. Before: a web design agency and a content agency, money going out with no return on investment yet.
AI
Generates content for the business, researches top performing times and trending Google keywords, and drafts or creates blog posts for the website. Then it publishes the blogs and posts, and drafts outreach around the business's unique selling points.
Automation
Once clients or leads are added, sends an introductory offer email. Posts automatically 5 times a week on social media, fed content by the human or the AI.
Case study 06
Relationships: staying in touch all year
Human
Remains in touch, sets reminders to text on birthdays, and calls with occasional offers.
AI
Researches individual client profiles and drafts the messages that go straight to clients on Lattice. It sends communications out through the API, honoring opt-out requests and compliance. The human supplies the offers, and the AI hands each one to the right audience.
Automation
Every month, a Lattice email goes out. On birthdays, holidays, Mother's Day and Father's Day, texts are sent when the client's tags meet the right criteria: a client tagged MOTHER gets the Mother's Day text, FATHER gets the Father's Day text, and a client with no birthday on file gets no birthday text.