
Agents that work alongside you like colleagues.
Agents take on routine tickets, reports and whole processes. In Leadtime they work in the same system as your team, with their own account, their own role and booked time. That makes your company ready for agents, and you stay in control.
#1953 File server permissionsRheintal Logistik support · assigned by Pia Wolf
- Read the project documentation
- Wrote the answer with instructions into the ticket
- Booked 18 minutes on the contingent
- Set the ticket to done
The problem
Your team already uses AI. But nobody knows who does what with it.
Every company wants to use AI, and in most of them people already do: each with their own tool, their own account and their own data. That is sprawl. It is hard to roll out together, and even harder to control.
For agents to really take work off your hands, they have to sit deep inside your processes: in the tickets, in the projects, in time tracking. Clearly defined processes they handle themselves, such as automations, regular routines or whole processes like handling client requests. That needs a coordinated approach instead of a patchwork nobody can see.
The trick is to stay in control. That works with a system in which people and agents work together, an operating system for the whole company. There, agents are not outsiders but team members: they take on roles and processes, and it feels normal to plan them in and give them work.
That is what Leadtime is about: we make your company agentic ready.
Sprawl or team member
Everyone on their own
- Everyone uses their own tool
- Client data ends up in private accounts
- Nobody sees what an agent has done
- Agent work is on no invoice
In Leadtime
- One system for people and agents
- Every agent with its own account and role
- Every change in the history, under its name
- Agent time booked like everyone else’s
What it looks like
Agents in just under three minutes, in the real app.
The Leadtime Agent, your own agents, automations and what an agent may do. Jump straight to the part you care about.
Video transcript
Introduction
This video shows in four steps how agents work in Leadtime: the Leadtime Agent in the chat, your own agents, automations and the permissions an agent works with.
The Leadtime Agent
The Leadtime Agent is built in. You write to it in the chat, as you would to a colleague, or ask it in the search field at the top. For example: which projects need attention right now? The agent goes through open tasks, priorities and deadlines and answers with an overview. Below that it sums up what matters now.
Your own agents
You create your own agents in administration. Each agent is a user of its own, with a name and a role that decides what it may do. Describe what the agent should do, and Leadtime sets it up with you. Or you create it by hand. Besides your own agents, you can connect coding agents or run an agent environment of your own.
Automations
With automations, an agent works on a schedule or when certain events happen. You choose the trigger, for example once a week. And you describe in your own words what needs doing. You run it yourself or let the agent run it.
Role and permissions
In the settings you decide which role the agent works with when it takes on tasks from assignments or automations. In the chat it works with your permissions. That is how agents work in Leadtime, like colleagues: you assign them tasks, they work with the permissions of their role, and what they do shows up in the history under their own name, with the time they booked.
Photo: The managing director of a software company stands in the open-plan office holding his phone, looking sceptically at his team at their laptops.
Managing director of a software company
“Who here is actually using which AI, and with what data?”
How to bring order to itHow Leadtime solves it
One system where people and agents work together.
Your team, the built-in Leadtime Agent, your own agents and your tools all work with the same clients, projects, tickets and times. An example: a software company running on Leadtime.
People
Pia WolfProject lead, works in the app
Jonas MeierDevelopment, books time on tasks
Miriam LehnerClient, as a guest in the client portal
- Clients
- Projects
- Service
- Team
- Money
Leadtime Agent built in
Roles, permissions and history for everyone
Agents
- Leadtime AgentSupport agent with its own account
- Claudevia the MCP server
- Codexvia the MCP server
- CCursorin the cloud, on development tasks
Tools and processes
- SlackAgents and automations by message
- InboxGoogle, Microsoft 365 or IMAP: emails become tickets
- Automationson a schedule, an event or a webhook
- API, Zapier, n8nfor every other system
Agents on the team
An agent joins the team like a new colleague.
It gets a name, a picture and a role. The role decides what it may see and change. Then you assign it tasks, mention it in a comment or write to it in the chat.
What you get: Your team works with the agent as with any colleague: assign a task, mention it in a comment, check the result.
Tickets
A ticket goes to the agent. A solution comes back for you to check.
Pia assigns the Rheintal Logistik ticket to the Leadtime Agent. The agent reads the project documentation, replies with short instructions, books its 18 minutes and sets the ticket to done. Every step is in the history, under its name.
What you get: Routine work gets done, and you see every step under the agent’s name.
- 1Assigned
- 2The agent works
- 3Reply in the ticket
- 4Time logged, done
Agent time
It books its time. So you see what it worked on.
The agent books its time on the task, as an ordinary entry under its own name. It appears in the time list next to your team’s and counts towards the contingent or the project. Whether it is billable is set by the project, as with any other entry.
What you get: You see what an agent delivers, and its time can go onto the invoice like any other.
In a plain ticketing tool
- The agent resolves the ticket.
- Its time isn’t recorded anywhere.
- The retainer knows nothing about it.
- Nobody sees what it cost and what it earned.
In Leadtime
- The agent resolves the ticket.
- Its time is in the time list, under its own name.
- It counts towards the retainer or the project.
- Billable hours on tickets land in invoice review.
Leadtime Agent
Say what you need. Leadtime gets it done.
The Leadtime Agent is built in and does whatever its role allows: create tasks and tickets, book time, write reports, change projects. You type or speak, it carries it out, and if something is unclear, it asks. You reach it as its own page, in the sidebar, via ⌘K, in the editor and in Slack.
What you get: No hunting through menus, no forms to fill in. You state the task, the agent carries it out and shows what it did.
Today
Create a ticket for Rheintal: the printer in hall 2 isn’t printing. Give it to Jonas, high priority.
Agent worked for 6s
Done. The ticket is in the project Rheintal Logistik support:
What else is open at Rheintal?
Agent worked for 4s
Three tickets, two of them with Jonas:
- #1958 Printer in hall 2 isn’t printing · Jonas
- #1956 Check the backup report · Jonas
- #1953 File server permissions · Leadtime Agent, done, Pia is checking
Ways of working
Five ways agents work with Leadtime.
Depending on the task, a job reaches Leadtime in one of five ways: from the built-in chat, from your own AI system, from an automation, from an agent in the cloud or via the API. Choose a way.
What you get: You use the AI that suits your team, and every way ends in the same system, with role and history.
You, in the chat or via ⌘K
“Create a ticket for Rheintal: printer in hall 2.”
Leadtime Agent, with your role
create_task →Leadtime
Ticket #1958 in the Rheintal support project
Your AI system: Claude, Codex, Cursor and others
“Which Rheintal tickets are open?”
MCP server, signed in with Leadtime
list_tasks →Leadtime
Three open tickets, with links into the app
Schedule: every Monday, 8 am
“Summarise the week’s open tickets and send it to the support channel.”
Automation, run by an agent
run_automation →Leadtime and Slack
Summary in Slack, the run in the log
Task assigned to Cursor or Claude
#2134 Order overview for dealers
Agent in the cloud, in the project’s repository
open_pull_request →Leadtime
Comment, status and link to the pull request
Your script or a self-hosted agent
Hermes, OpenClaw or your own program
Public API, token, signed webhooks
POST /tasks →Leadtime
Data in Leadtime, with the token’s permissions
leadtimeread and maintain dataleadtime-setupfrom an empty workspace to ready to goleadtime-coachsetup, import, migrationleadtime-public-apiscripts, Zapier, n8nproject-documentationdocument projects
Coding agents
A development task goes to Cursor. The result comes back, with a pull request where needed.
Assign the task to Cursor in the cloud or mention it. It works in the project’s repository, reports back with a comment and a status change, with the link if there is a pull request, and hands the task back to you.
What you get: Development work goes through the same planning as everything else, and you see on the ticket what the agent delivered.
- A repo per projectYou set the default repository and settings once on the project.
- Claude as a managed agentLeadtime sets up the agent, its environment and a vault for credentials, with skills, your own MCP servers, GitHub repos and a log of every session.
- How much of Leadtime it seesWith “Basic” it reads the task, comments and sets the status. Or you give it full access to Leadtime.
Automations
What runs the same way every week takes care of itself.
An automation is a job in your own words plus someone who carries it out: you or an agent. It starts on a schedule, when something is created, on a Slack message or a webhook. You see in a test run beforehand whether it does what it should.
What you get: Recurring work runs to plan, and every run is in the log.
Photo: The operations lead of an IT service provider sits at her laptop, half turned away, looking to the side with a testing expression.
Head of operations at an IT service provider
“And who checks what an agent like that actually does?”
How you stay in controlSecurity and control
You decide what an agent may do. And see what it has done.
An agent gets exactly the permissions of its role, like everyone on the team. It cannot sign in to the app; it works only through the API and MCP. And when you no longer need it, one click locks it out.
What you get: You use agents without giving up control. Every change can be looked up.
Permissions: the “Agent” role
| Works on it | |
|---|---|
| Tasks and tickets | yes |
| Projects and organisations | yes |
| Times | yes |
| Invoices and insights | yes |
| Administration and settings | no |
You can also give an agent any other role. Who may chat with agents is a separate permission.
Log: the trail under its name
- Leadtime Agent took on #1953, started by Pia Wolf
- Leadtime Agent wrote a comment with instructions
- Leadtime Agent booked 18 minutes on the contingent
- Leadtime Agent set the status to done
Pia Wolf read the result
- No app sign-inAgents work only through the API and MCP.
- Tool scopeA coding agent sees only its task, or you give it full access.
- QuestionsInstead of guessing, the agent asks and lets you choose.
- Tokens that expireRead or write, revocable at any time. Webhooks are signed.
- Deactivated instantlyOne click, and the agent can no longer authenticate.
What you save
A monthly report in 1 hour instead of 6.
Example: a service provider with 20 employees. The agent compiles the management report from projects, times and invoices, and your team reviews it. Its time is booked on the project.*
60fewer hours a year spent on the monthly report
6×faster per report6 hours today, 1 hour with Leadtime, per month
€3,000less in staff costs a yearapproximately, at €75,000 per full-time role and 1,600 hours a year
* Model calculation for the “management report” process with our calculator and mid-range assumptions, not a promise of results. In the calculator you also see routine tickets and quotes taken on by an agent.
What it costs
The Leadtime Agent is included in the user price. For more, there are packages.
Usage is billed in credits; 100 credits equal €1. All credits sit in one shared balance, and what a job actually uses is deducted. Agents that run inside Leadtime use credits and don’t need a seat on the team. Agents you bring, such as Claude, Codex or Cursor, run on your account with the provider.
Included
300 credits per user and month
In the user price. Enough for normal use, about 20 routine tickets or 100 questions per user.
Package S
€20 2,000 credits
about 150 routine tickets or 40 draft quotes
Package M
€50 5,000 credits
about 400 routine tickets or 100 draft quotes
Package L
€100 10,000 credits
about 800 routine tickets or 200 draft quotes
Agents you bring
Your account with the provider
Claude, Codex, Cursor and others. You pay the provider directly and use no credits.
Packages come as a monthly package or a one-off top-up.
All pricesEverything at a glance
Every feature for agents.
Each feature comes with a sentence on what it does for you.
Create an agent
Enter a name, picture and role by hand, or describe the agent in one sentence. You try it out straight away in the test chat.
The “Agent” role
Access to tasks, projects, organisations, time, invoices and insights, but not to admin and settings. You can assign any other role too.
Assign and mention
Assign the agent a task or mention it in a comment. Every assignment and every mention is a session of its own.
Sessions on the task
You can see when an agent is working on the task. The timeline shows every session from start to result, with the trigger and who started it.
Log time
The agent logs its time on the task as a normal time entry under its own name. It shows up in the time lists like your team’s.
Chat with history
As a page of its own, in the sidebar or as a floating window. You pin the chats that matter.
Everything its role allows
Create tasks and tickets, log time, edit projects and invoices. It also opens the right page in the app.
Bulk changes
Change many tasks at once. If anything is unclear, it asks first.
Voice and files
Dictate your request or attach a file. Reports come back as a file to download.
Your choice of model
Models from Anthropic, OpenAI, Google and others, or “Auto”. You set your favourites.
Skills
Built-in skills such as report, what to do today or project documentation, plus skills of your own for your workspace.
MCP server
Claude, Codex, Cursor and others sign in with Leadtime, as an agent if you like, and reach the whole public API.
Access tokens
Read only or read and write, with an expiry date and revocable at any time.
Plugins with skills
For Claude, Codex and Cursor: keep data up to date, set up a workspace, move over from another system, write scripts, document projects.
Self-hosted agents
Through signed webhooks and the session API, with setup guides for Hermes and OpenClaw.
Coding agents on tasks
Cursor in the cloud or Claude as a managed agent work in the project’s repo and report back with a comment and a pull request.
Prompt and runner
Every automation is a request in your own words, carried out by you or by an agent.
Schedule
Daily, weekly, monthly or on your own cron expression. Recurring tasks run this way too.
Events, Slack and webhooks
When an organisation, project or task is created, on a message in a Slack channel or an incoming webhook.
Test run and log
Try an automation out before it runs for real. Every run is logged afterwards.
No app login
Agents work through the API and MCP only. They can’t log in to the interface.
A trail under its own name
Every change is in the activity log under the agent’s name. You see separately what a person did and what an agent did.
Who may chat
“Chat with agents” is a permission of its own that you grant per role.
Deactivate instantly
One click locks the agent out everywhere, via the API, MCP and all tokens.
Next step
Let us show you on one
of your processes.
30 minutes, one process from your day-to-day work that an agent could take on. You see what it looks like in Leadtime and where you stay in control.
ClientsOpportunities, quotes, product catalogue
ProjectsPlanning, tasks, sign-off
ServiceHelpdesk, tickets, retainers, assets
TeamPeople, absence, time
MoneyInvoices, e-invoicing, insights
AgentsLeadtime Agent, your own agents, automation


