TLDR:
AI workflows can sit alongside your existing PR software rather than replacing it. Connect your calendars and timesheets to Claude or ChatGPT and account team capacity, cover notes, handovers and new business staffing arrive drafted.
In a PR agency, people are the product and they are also the constraint: every account leader sits across several clients, and the honest answer to who has room next month usually comes from asking around. Your calendars and timesheets already contain that answer. This guide covers how to connect them to Claude or ChatGPT so capacity, cover and handovers arrive drafted.
It covers the team side of our PR agency automation guide. Retainer hours and margins are in PR agency billing automation; client approvals and status updates are in how to automate PR client admin.
Connecting AI to your agency's calendars and timesheets
The inputs here are ones you already have switched on. Google Calendar and Gmail, or Outlook and Teams, connect to both Claude and ChatGPT as a setting. Timesheet data arrives the same way it does for billing — an API where your tracker offers one, a scheduled export where it doesn't.
Same boundary as everywhere else: the AI reads and drafts, a person decides. It does not reassign anyone, move a meeting, or send a note to your team until someone has approved it.
In staffing that line is not a technicality. Who covers which account is a development decision as much as a scheduling one, and an allocation that looks efficient on hours can be the wrong call for the person doing the work.
Using AI to see account team capacity
The question agencies ask constantly and answer badly is who has room. It is answerable from records, and the useful version is specific rather than a utilization percentage.
Committed time by person. Hours already logged plus what is on the calendar for the next two weeks, per person, against the accounts they sit on. A name at full commitment across four accounts is visible before they tell you.
The uneven split. Where one person is carrying significantly more than peers at the same level, surfaced as a short note. In a business that competes hard for experienced account leads, this is the number worth reading every week, and it is almost never the one on a dashboard.
What is coming, not just what has happened. Launch dates, reporting deadlines and client reviews already on the calendar, mapped against who is attached to them. Two launches landing in the same week on the same team is a staffing problem while it is still three weeks out, and a scramble the week it arrives.
If you have ever found out about a team's workload from an exit conversation, Unprompted can build this against the calendars your agency already runs.
Using AI for cover when someone is out
Planned leave and unplanned absence are the same operational job with different notice, and most of the cost is in the handover rather than the absence.
- The handover note. For each account the person leads: what is in flight, what is awaiting client approval, which reporters are mid-conversation, what is due while they are away, and who the client's day-to-day contacts are. Drafted from the tracker and the inbox instead of written at eleven the night before.
- Who can realistically cover. People with existing context on that client or that beat, with their current commitment shown, so cover goes to someone with room rather than to whoever answers first.
- The client-facing note. A short message telling the client who to contact and until when, drafted for the account lead to approve.
- The return brief. What moved while they were out, waiting on the morning they are back.
Using AI to staff new business without guessing
New business is where staffing decisions get made fastest and with the least information, usually in the week between a win and a kickoff.
Against a signed scope, the work is readable: what the account needs in media relations, content and reporting hours, which existing accounts resemble it closely enough to be a useful comparison, and what those accounts actually consumed in their first three months rather than what they were scoped for.
Put beside current commitments, that produces a staffing option with its tradeoffs visible — who would need to come off what, which accounts would be near capacity as a result, and where a freelancer is the better answer than stretching a team. The recommendation is yours; the arithmetic underneath it stops being guesswork.
The same records make the pitch stage more honest. An agency that knows what a comparable account really consumed is in a better position to scope the next one properly, which is the cheapest way to avoid the overservicing problem before it starts.
What stays human
Every decision about a person stays with people. Who is ready to lead an account, who needs a lighter month for reasons that are not on a timesheet, who should be in the room for a pitch because of what it will do for their career — the records do not hold any of that.
The same goes for the conversation itself. Being told you are overloaded by someone who already knew is very different from a system noticing it. The reading and the drafting can be handed over; the noticing that matters is still a manager's job, done better when the arithmetic arrives without being asked for.
If you find out about workload problems too late to do anything about them, talk to us and we'll show you the capacity view built from your own calendars.
Want this running inside the tools you already use?
Book a callFAQ
Can AI tell me who has capacity on my team?
It can show committed time per person — hours already logged plus what is on the calendar — against the accounts they sit on. That turns who has room into a number you can read weekly instead of a question you ask around.
Can AI write a handover note?
Yes. For each account a person leads it drafts what is in flight, what is awaiting client approval, which reporters are mid-conversation and what falls due while they are away, built from the tracker and inbox rather than written the night before.
Does this make staffing decisions for us?
No. It prepares the arithmetic — commitments, comparable accounts, what those accounts actually consumed — and lays out the tradeoffs. Who is ready to lead an account, and who needs a lighter month, stays a manager's call.
What data does it need to see capacity?
Calendars and timesheets. Google Calendar or Outlook connect as a setting; timesheet data comes through your tracker's API or a scheduled export. Read-only access is enough.