How We Built a Daily Briefing That Runs Itself

Our practice manager used to start every morning the same way. Arrive at the clinic, sit down, and begin the rounds. Open the practice management software — check today's appointments. Switch to Gmail — scan for anything urgent. Open the lab tracker — see what's due back. Pull up the roster — confirm who's in. Check for overnight cancellations. Maybe glance at whether any patients have birthdays today.

Twenty to thirty minutes of clicking through systems, every single morning, before she could actually start managing the practice. It wasn't difficult work. It was just tedious, repetitive, and spread across too many places.

So we automated it. Now, at 7am every morning, a briefing lands in her inbox and is available through Spark, our voice assistant. By the time she walks in, the day is already laid out. No clicking. No logging into five systems. Just a clean summary of everything she needs to know.

We think of it as her AI chief of staff — it does the morning prep so she can focus on decisions, not data gathering.

What the briefing actually includes

We spent a couple of weeks observing what our practice manager actually checked each morning. Not what we thought she checked — what she actually did, in what order, and what she wrote down. That observation shaped what goes into the briefing.

Today's appointments with notes. Not just a raw list of names and times. The briefing highlights anything the team should be aware of — a patient with a medical alert, someone who's been anxious at previous visits, a complex procedure that might run over time, or a new patient who needs extra onboarding time. It pulls these notes from the practice management system so nothing gets missed.

Overnight cancellations and changes. Patients cancel via the online portal or by email outside office hours. The briefing catches all of those and presents them upfront — "two cancellations overnight, leaving gaps at 10:30am and 2:15pm." The team can start working the waitlist before the day even begins.

Outstanding lab cases and expected returns. Crown coming back today? Denture repair overdue by two days? The briefing tracks what's out with the lab and flags anything that's late or expected for today's patients. Nothing worse than a patient arriving for a crown fit and the crown isn't back yet — that used to happen when things slipped through the cracks.

Priority emails flagged by the AI inbox. We already use AI to manage our email, so the briefing taps into that system. It pulls out anything flagged as high-priority overnight — a health fund policy change, a patient complaint, an urgent supplier notice — and summarises it in a sentence or two. The full email is still there to read, but the briefing gives you the headlines.

Staff roster for the day. Who's in, who's off, any notes. If someone called in sick yesterday afternoon, the briefing flags it. If a locum is covering, it mentions that too. Simple, but it prevents the "wait, I thought Sarah was in today?" moment at 8:45am.

Birthdays and special occasions. This one surprised us with how valuable it became. The briefing flags any patients coming in today who have a birthday this week. It takes two seconds for the receptionist to say "Happy birthday, by the way!" and it makes a genuine impression. Small touches like that build loyalty, and they only happen if someone remembers to check. Now nobody has to remember — it's just there.

How it works

The system is simpler than you might expect. There's no magic involved — just a scheduled task and some well-connected plumbing.

Step one: a scheduled task runs at 7am. Every morning, a script kicks off automatically on our local server. No human involvement. It runs the same way whether it's Monday or Saturday, whether someone's in the office or not.

Step two: it pulls data from multiple sources. The script connects to our practice management system for appointments, patient notes, and the roster. It checks the email system for flagged messages. It queries the lab case tracker for outstanding work. Each of these is a separate data source, and the script knows how to talk to each one.

Step three: an AI summarises everything. The raw data from those systems is messy — appointment codes, internal reference numbers, technical labels. A language model takes all of that raw data and turns it into a clean, readable summary. It highlights what's important, skips what's routine, and writes it in plain English. The AI runs locally on our own hardware, so no patient data leaves our network.

Step four: it delivers the briefing. The finished summary gets sent via email and made available through Spark. The practice manager can read it on her phone on the way to work, or ask Spark to read it aloud when she arrives. Same information, different format — whatever suits the moment.

The whole process takes about 30 seconds to run. By 7:01am, the briefing is ready and waiting.

The time maths

Let's be conservative. Our practice manager was spending 20 to 30 minutes every morning on this manual check. Call it 25 minutes on average.

That's over two hours a week. About nine hours a month. Over a hundred hours a year.

A hundred hours of a skilled practice manager's time, spent doing something that a script can do in 30 seconds. At a loaded cost of around $50 per hour, that's roughly $5,000 a year in staff time — on morning prep alone.

But the real value isn't just the time saved. It's the consistency. A human doing this manually will occasionally miss things. They're tired on a Monday morning. They get interrupted halfway through. They forget to check the lab tracker because they got pulled into a phone call. The automated briefing doesn't get tired, doesn't get interrupted, and never forgets a step. Every morning, without exception, the full picture is there.

What changed in how we work

The most noticeable change wasn't the time saved — it was the confidence. Our practice manager walks in already knowing the shape of the day. She's thinking about how to handle the afternoon gap from a cancellation, not still discovering that the cancellation happened.

The morning team huddle got sharper too. Instead of everyone gathering around a screen while the manager clicks through systems, she arrives with the briefing already read and runs through the key points in two minutes. "Two cancellations, the Johnson crown is due back today, we've got three new patients, and it's Mrs Cooper's birthday. Any questions?" Done.

The birthday thing became a favourite with the whole team. Patients genuinely light up when you remember. One patient told us she'd never had a dentist wish her happy birthday before. It's a tiny gesture that costs nothing, and it only happens reliably because the system flags it automatically.

This works for more than dental practices

We built this for ourselves, but the pattern is universal. Any business where someone spends their first half hour checking systems and getting up to speed is a candidate for an automated briefing.

• Restaurants — a morning briefing with tonight's bookings, any large parties, dietary requirements flagged, staff roster, and whether key stock items need reordering. The head chef and floor manager get the same picture before the kitchen even fires up.
• Trades companies — today's job schedule, weather forecast for outdoor work, expected supplier deliveries, any customer messages that came in overnight. The office manager sends the crews out informed instead of scrambling.
• Professional services — an accountant or lawyer starting the day with a summary of deadlines, new client enquiries, upcoming meetings, and anything flagged from yesterday. No more opening six tabs and three apps before the first coffee.
• Retail — today's staff, expected deliveries, online orders to fulfil, any customer complaints pending, and stock alerts. The store manager hits the floor ready to go.

The common thread is the same everywhere: information scattered across multiple systems, and a person who has to manually stitch it together every morning. That stitching is exactly what AI is good at — gathering data from different sources, summarising it, and presenting it clearly.

What you need to make it work

This isn't a plug-and-play product you download from an app store. It's a custom system built around your specific tools and workflows. That said, the requirements are surprisingly modest:

• Your systems need to be accessible via an API or database. Most modern practice management, booking, and email systems have APIs. If your software has an integration or export option, it can probably be connected.
• You need to define what matters. The briefing is only as good as the questions it answers. We spent time watching how our manager actually started her day before deciding what to include. That observation step is essential.
• Someone needs to build and maintain it. The initial build is straightforward. Ongoing maintenance is minimal — mainly adding new data sources or adjusting what gets highlighted as your workflows evolve.

The technical complexity is low. The value comes from understanding the workflow and connecting the right dots.

Your AI chief of staff

We started calling the briefing our "AI chief of staff" as a joke, but it stuck because it's accurate. A good chief of staff doesn't make decisions for you — they make sure you have everything you need to make decisions quickly. They've read the briefs, checked the schedule, flagged the problems, and laid it all out before you walk in the door.

That's exactly what this system does. It doesn't run the practice. It just makes sure the person running the practice never starts the day in the dark.

If your mornings start with 20 minutes of logging into systems and piecing together what's ahead, that time is recoverable. We've done it for ourselves and we can help you do the same. Get in touch — we'll walk through your morning workflow and show you what an automated briefing could look like for your business.