AI agents
Less repetitive work.
More time to grow the business.
Tabilo's AI agents fit into the processes you already have: they answer customers, manage tasks, follow up on leads and keep decision-makers informed. They work side by side with your team — not in its place.
They are not chatbots you have to set up on your own: they are agents that rely on your company memory and work with the tools you already use.
“Assign Mark the task ‘call Johnson back about the roof quote’, high priority, by Friday”
Task created and assigned to Mark, due Friday. Notification sent to his phone — he can reassign or edit it anytime.
Tools involved
The agent takes the action — a person always sees the confirmation.
The problem
It's not just that documents
are hard to find.
- Only one person knows the right procedure — and today they're on vacation, out sick, or gone for good.
- Repetitive work (a quote, a follow-up, a status update) still goes through someone, even when the answer is always the same.
- You already have a CRM, a ticketing tool and a project management tool: each one does its part, but none of them talk to each other — connecting them is still your job.
It's an operational problem, not just a document problem: the work depends on who knows how to do it, not on a system that knows how to do it alongside them. And buying another standalone tool doesn't close the loop — it adds one more piece to connect by hand.
What you get
Agents with a clear job, not one more piece of software.
Each agent has a clear responsibility, not a vague scope to interpret. You always know who uses it in your company, what it actually does, which tools it touches and what it delivers — just like a new hire.
Sales AI
who uses itYour sales team and whoever handles customers
what it doesFollows up on leads and quotes, keeps the CRM up to date
toolsCRM, email, calendar
Customer Care AI
who uses itWhoever answers customers every day
what it doesAnswers recurring requests, hands complex cases to a person
toolsPhone or web, company memory
Marketing AI
who uses itWhoever handles communication and content
what it doesDrafts blog posts, newsletters and social posts for you to approve and publish
toolsBlog, newsletter, social
Project Manager AI
who uses itWhoever coordinates projects, deadlines and people
what it doesCreates and assigns tasks, tracks priorities and progress, prepares reports
toolsTask management/kanban, calendar
CEO Assistant AI
who uses itWhoever runs the company and needs the big picture
what it doesGathers KPIs and scattered data, prepares summaries and priorities
toolsCompany memory, CRM, tasks
The setup
An agent dedicated
to your company.
Each agent works in an environment reserved for your company, with its own memory, skills and connections. Your team finds it on the shared board, next to their coworkers, and on their phone, at any hour.
Setting up the first agent, with memory, board and phone access, is included from the Tabilo AI System plan up.
What we set up for each agent
Always on
The agent runs on a server dedicated to your company, on 24/7 and monitored by us. Nothing to install on your computers.
Memory and skills
Your company memory and skills written around your processes: the agent knows how you work and remembers what happens.
Two ways to work
Your team hands work to the agent from the shared board or messages it from their phone: same company context, different entry point.
A separate space for each company
Operational data, memory and settings stay within each company's own space: no customer sees or uses another company's space.
AI roles
Five roles, with real examples.
These aren't sweeping promises: they're examples of how each role works once it's set up around your company's processes. Some capabilities are ready now, others are built as a custom project — each card tells you which.
AI role
ConfigurableSales AI
Your sales team and whoever handles customers
Reads incoming leads, prepares follow-ups, drafts quotes and updates the CRM as each deal moves forward.
No forgotten leads, quotes ready sooner, a pipeline that is always up to date.
Example: A form filled out on your website becomes a lead in the CRM within minutes: the agent drafts the first reply and sets a three-day follow-up reminder in case nobody has answered yet.
AI role
ConfigurableCustomer Care AI
Whoever answers customers every day
Answers recurring requests with replies that fit the context — it takes the customer's history into account and checks the company memory — and hands a person the cases that need a decision.
Faster, more consistent answers, and coverage outside normal business hours too, when the system is properly hosted and monitored.
Example: A customer asks about the status of their request late in the evening: the agent checks the history and answers right away; if a decision is needed, the case is ready for the next morning.
AI role
ConfigurableMarketing AI
Whoever handles communication and content
Drafts blog articles, newsletters and social posts, with the keywords that help your website in mind; your team reviews and publishes.
More regular publishing, less time spent writing every piece from scratch.
Example: From a case discussed on an internal call, the agent drafts an article and the matching newsletter: your team reviews them and publishes in a couple of steps.
AI role
ConfigurableProject Manager AI
Whoever coordinates projects, deadlines and people
Creates and assigns tasks, keeps track of priorities and progress, and prepares regular reports on project status.
Fewer manual reminders, and project status always visible to the people who need it.
Example: At the end of the week the agent prepares the progress report for the Baker project and sends it to the project lead before Monday's meeting.
AI role
Custom projectCEO Assistant AI
Whoever runs the company and needs the big picture
Gathers KPIs and operational data scattered across your business systems, prepares regular summaries and highlights the week's priorities.
Faster decisions, less time spent pulling together numbers from different sources.
Example: Every Monday morning a summary lands on the channel you chose: the week's revenue, overdue tasks, hot leads to follow up on.
How it works
A written request.
A traceable action.
You don't have to give up the tools you already use: your team hands off work from the shared board or writes from their phone. The agent keeps your company's context, picks the tools it's allowed to use and confirms what it did — always under the supervision of the people who actually work in your company.
How a request travels
Your team writes the request
From the shared board or from a phone, in plain language — the way they'd write it to a coworker.
The agent interprets it
It works out what's needed and decides which authorized tools to use to get it done.
It checks, then acts
Company memory, tasks, email, calendar, CRM, documents: it uses what it needs, then gets to work.
It confirms, under supervision
It reports what it did. Your staff review and correct routine actions; the most sensitive operations still require approval in advance.
Company memory
Everything your company knows,
in one place.
Procedures, SOPs, FAQs, manuals, contracts, price lists, meeting notes, uploaded transcripts, technical documentation: the company memory is where all of this material — uploaded or connected by your company — turns into answers that fit the context, always with their sources cited and visible only to people with access to that kind of information.
It's the easiest place to start, even before choosing which AI agent you need: putting order into what your company already knows.
Answers with verifiable sources: available now. Connections to more specialized business tools: assessed during the project.
“What's the procedure for a return after 30 days?”
Above €200 you need approval from the sales manager; below that amount the return is accepted directly, as set out in Returns SOP v3.
Sources — from company memory
Answers from what your company knows — with separate access by role.
Workflows
Four examples, taken from real work.
Not abstract demos: situations that come up every week at a small or mid-sized business, and how AI agents get them done.
Handling a task
A manager writes “assign the contract review to Emma by Thursday”: the agent creates the task, assigns it and confirms in the chat.
Looking up a procedure
A team member asks how to handle a return after 30 days: the agent checks the company memory and answers, citing the right SOP.
New lead from the website
A form filled out on the website is assessed, turned into a task in the CRM and assigned to whoever handles that type of customer.
Daily report
Every morning the agent gathers the previous day's operational data and sends a summary to the channel your company chose.
Security and control
The agent's autonomy has limits.
An AI agent works within clear rules: what it can do on its own, what it needs approval for, and what is always logged.
Separate data for each company
Operational data, documents, company memory, settings and the AI agent's environment are kept separate for each company: no company can access another's space.
Role-based access
Access is checked by account and role: each person sees and uses only what they're responsible for, following the principle of least privilege.
Traceable actions
The AI agent's actions are logged — what was done, when, and at whose request — so you can check them whenever in doubt.
Human control
The most sensitive operations still require approval: the agent proposes, a person confirms.
Backups and continuity
Backups and business continuity procedures, designed to limit the impact of a failure or a mistake.
Onboarding
Eight steps, from mapping to ongoing improvement.
It's not an installation: it's a process we go through together, starting from how your company really works.
Process review
Together we look at how your company works today: where time gets lost, which requests keep coming back, who does what.
Gathering documentation
Procedures, SOPs, price lists, contracts, FAQs, technical material: everything that today is scattered across people and folders.
Building the company memory
The material we gather becomes the body of knowledge AI agents rely on to answer and decide.
Setting up the system
We set up the AI roles your company needs, with rules and permissions that fit your situation.
Connecting your business tools
CRM, task management, calendar, email: we connect the AI agents to what you already use.
Real-world testing
We test the system on real cases before putting it in front of customers or your whole team.
Staff training
Your team learns to work with the new agents: how to write requests, how to read the answers.
Monitoring and ongoing improvement
We follow how the system performs in the first months and fine-tune rules and answers based on what actually happens.
The offer
Every company starts from a different place.
AI roles are set up around your company’s specific needs: we work them out and size them on a call, starting from Tabilo’s plans, not from an off-the-shelf price list. Where your data lives and what the agent does on its own is explained in security and data.
- If you want to put order into your company's knowledge, start with the company memory on its own: it's the easiest place to start, even before choosing which AI role you need.
- If you have a specific process to lighten — customer requests, quotes, your team's tasks — pick the first AI role that helps with that process.
- If you want several roles working together, build a larger team of AI agents, one step at a time.
What we define on the call
- Dedicated server, always on
- Shared board, with accounts for your team
- Private access from your phone
- Company memory
- A first AI role or a team of AI roles
- Integrations with the tools we agree on
- Onboarding and team training
- Hosting, monitoring and support, as set out in the contract
We don't publish a price list for AI agents: the mix of roles, integrations and components is defined on a call, starting from Tabilo's plans.
Let's find the first agent that's useful to your company.
In half an hour we look at your company's processes and find the spot where an AI agent would make a difference fastest — company memory included.
What to prepare before the call
- A couple of tasks that repeat every week
- The tools you already use (CRM, tasks, email, calendar)
- Who in your company would follow the project