Productivity Blueprint
Safety checkedFind who is free next week and book the work in Hive
Avoid assigning work to someone who is on leave or already over capacity. See daily free hours, review proposed project bookings, then approve changes in Hive and verify the updated schedule.
What problem does this solve?
A resource manager asks the same question each week: who is free, and who is full? The answer is spread over a resource plan, leave records and a holiday calendar. Putting it together by hand takes time, and it is old by the next day. Work then lands on a person who is on leave, or on one who is already over their hours.
This Blueprint takes the numbers from Hive. Hive has already taken booked work, time off and holidays out of each person's day. You get one answer in one chat.
How does it work?
- Set the window. You give the dates, the people or team, and, if you have it, the work to place. Claude reads the member list and the placeholders from Hive. A placeholder is a role, such as "Designer", that you can book before you have a name.
- Read free hours. For each person and each day, Claude reads capacity, booked hours, time off and the hours left. A day with no capacity is a weekend, a holiday, or a person with no schedule. Claude checks the cause.
- Read bookings. Claude lists the assignments in the window, with project names. It sees time off as a booking of its own type.
- Propose. You get a table of free hours for each person, a list of anyone overbooked, and up to five moves. Each move names the person, the project, the dates and the hours for each day. Claude is told not to propose more than a person's free hours on any day. Nothing is written yet.
- Book. After your yes for a move, Claude creates the assignment in Hive. To change one, it finds the assignment first and changes only the fields that differ.
- Check. Claude reads the assignments and the free hours again and reports what changed.
Example from our run. We asked Claude to place 12 hours of work on a test project, on a placeholder, to finish by Wednesday. Hive showed 8 free hours a day. Claude proposed 4 hours a day on Monday, Tuesday and Wednesday, with an end date of Thursday. Hive stores the end date of a booking as the day after the last working day. After the booking, Hive showed 12 booked hours, not confirmed. Free hours fell from 8 to 4 on each of the three days, and Thursday stayed at 8.
What's the biggest win?
You get an answer from Hive's own capacity numbers. They already count booked work, time off and holidays. You can show the plan to a colleague before you approve it. Every write needs your yes.
What's required to run this?
- A Hive workspace with Team resourcing. Hive lists it as an add-on to the Teams plan (+ $5 for each user each month on hive.com/pricing).
- The Hive connector in Claude. Set-up steps: https://hive.com/integrations/claude?utm_source=g2&utm_medium=blueprint&utm_campaign=resource-planning
- Working hours for each person in Hive, and a holiday calendar. Free hours are only as right as these settings.
- Leave entered as time off in Resourcing, so that it shows.
- A project for the work you want to place.
- The skill file from this Blueprint, added to Claude as a skill or pasted as the first message.
What are the constraints?
- Hours only. Hive reports hours for each day, with no time zone. Dates are whole days.
- It shows work capacity. It does not find a time for a meeting. We never saw Hive return a meeting block, so do not rely on meetings being counted.
- One availability call covers at most 91 days, counting both end days. Keep the window to a few weeks: a large range returns a lot of data.
- Hive stores the end date of a booking as the day after the last working day. A Monday-to-Thursday job has an end date of Friday. Claude says this back to you. Check the dates in the plan view.
- Hive does not stop you from booking a person above their free hours. In a test it accepted 6 more hours on a day that already had 4 of 8 booked. Claude's rule is the only guard.
- A zero can mean a weekend, a holiday, leave, a full day, or no schedule. Claude checks before it acts. A holiday only affects the people on its calendar.
- It reads up to 50 workspace members in one run. For a larger team, name the people or a team. We did not test more than one page.
- Claude balances hours only. It does not weigh skills, cost or preference. Capacity can differ by day (7 hours instead of 8); Claude uses the number Hive gives.
- Claude saves each booking as not confirmed. Hive does not tell the booked person about a booking that is not confirmed. When someone sets it to confirmed, Hive tells the booked person. Claude asks you before it books anyone but you.
- Claude does not send messages. You tell the team. Hive's own notices still apply: if you change a booking that is already confirmed, Hive tells the booked person.
- If you ask for time off, Hive may tell the people who assigned that person's cards in those dates, and that can reach them by email. Claude asks you first. We confirmed these notice rules in a test environment with two users. We did not test time off that people book for themselves, and we did not test the email.
- We tested it in Claude, in workspaces with one member. We tested bookings for a second person, through the same connector calls, in a test environment with two users. It runs when you start it.
Tools in this Blueprint
About This Blueprint
- Industry
- Advertising & Marketing
More Blueprints to explore
Move a team's spreadsheet into a Hive project in one conversation
Avoid dropped tasks and a day-one flood of assignment alerts when moving a spreadsheet into Hive. Review the proposed sections, cards, and unmapped rows first, then add tasks with owners unassigned until you approve notifications.
Send Weekly Product Defect Reports
Catch product defects before scattered support tickets obscure a rising trend. Weekly reports group faults by product and type, compare them with the prior four weeks, and include customer photos and quotes for faster quality investigations.
Urska Blagojevic
GTM Lead at Yuma AI (YC W23)
Fix Wrong AI Agent Answers Before They Recur
Stop the same support mistake from frustrating customers again: trace incorrect Yuma replies to the Handbook, missing order data, failed actions, or bugs, then test a proposed fix before publishing it. A weekly report highlights recurring causes and the next tickets to investigate.
Urska Blagojevic
GTM Lead at Yuma AI (YC W23)