Build your own agents.
Seven agents come ready to use. Build your own by choosing what it watches, describing what to look for in plain words, and ticking what it is allowed to do. No code, no developer, and your data stays where it already is.
A leave request over five days asks a skip-level manager, writes the answer to attendance, tells the requester on WhatsApp and posts the month to a board. One workflow, four departments.
Seven ready to use
Start from one of these.
Each one comes set up with what to watch and when to run. Use it as it is, change it, or build your own from scratch.
Start from one of the seven, or point a new one at any table you have.
Finds deals going quiet, stale updates, a missing party or pipeline, and files a follow-up for the owner.
Watches invoices past due and part-payments going stale, and flags the ageing total.
Reads the action items from recent meetings and files each one as a task, so nothing dies in the notes.
Spots inbound chats waiting on a human reply past the SLA, and nudges whoever owns the thread.
Weekly sweep for deals with no company, contact or pipeline, and files a fix-list so rot never piles up.
Sees approved leave starting soon and flags what needs cover before the person is out.
Categorises and prioritises incoming support tickets.
Once it is yours
Finds deals going quiet, stale updates, a missing party or pipeline, and files a follow-up for the owner.
Build an agent for anything you track.
Building one
Five things to fill in.
Every option comes from your own NearSync setup, so a record type you added last week is already on the list. Nothing here is code.
- 1What it watches
Anything you keep in NearSync - deals, invoices, tickets, leave, or a record type you created yourself. Add filters and a limit so it only looks at what matters.
- 2What else it checks
Up to three related areas to check against. This is how one agent can look at deals, tasks and the calendar together.
- 3What to look for
Described in plain words. The agent works from your real records, so it is checking facts rather than guessing.
- 4What it may do
Tick actions from a list. Anything you leave unticked can never happen, whatever the agent decides.
- 5How much, and how often
Set a schedule, and a cap on how many actions one run can take.
One agent can check several things at once.
It checks the calendar and the open tasks before it tells you a deal has gone quiet.
- Point it at deals, invoices, tickets, leave or anything else you keep
- Have it check up to three related areas at the same time
- Filter what it looks at, and cap how much it takes in
- Watch a record type you created yourself, with no setup from us
- Get an answer that knows about the meeting you already booked
Test it before it goes live.
A dry run does everything except the actions, and shows you what it would have done.
- Run a full test that changes nothing
- See exactly what it would have done
- Tick only the actions it may take; nothing else can happen
- Cap how many actions one run can take
- Switch it off in one click, with nothing to unpick
Worth knowing
Nobody signs off each action.
A live agent makes its changes without asking first. That is what makes it worth having, and it is why the controls matter: you choose exactly what it can do, you cap how much it can do in one run, you can watch a full test first, and one switch turns it off.
Access
Who can do what.
Anyone can see what an agent did. Switching on an agent that changes a department's records takes a higher role.
See every agent and what each of its runs did.
Dry run an agent and see what it would have done.
Build agents and switch them on for your department.
Decide who can switch an agent on at all.
Admin is the highest role and is not limited by these permissions, so keep it to the smallest group that genuinely needs it.