A typical one-off chat
You carry the task.
- Notice the job and bring it to the chat.
- Supply the relevant context and instructions.
- Check the answer and carry the result forward.
How it works / The intended service
Give it a job. Keep the final say.
We’re designing a managed employee model: Spontalo sets up and maintains the role, while you define what good work means for your business. You should not have to build workflows or operate AI tools to keep it going.
From incoming work to an outcome
Today / on your plate
Notice it Gather context Check Chase Decide Follow through
An event arrives
The employee takes the job.
A message, document, booking or agreed check-in starts the work. It reads the context, applies your rules and handles the permitted steps.
Role + business context + agreed permissionsRoutine / within permission
Done. Recorded.
The agreed job is complete. Its history is there when you need it.
Exception / uncertainty / approval
Your decision, prepared.
A recommendation, evidence and an explicit question. Approve, edit, reject or ask for more.
You set the authority. The employee works up to that boundary and brings you what needs judgement.
From handover to everyday work
A clear agreement at the start.
This is the model being developed. There is no live employee service available yet.
Choose one job.
Name the responsibility and what counts as finished. Start with a repeatable piece of work, not a whole department.
Agree your rules.
Set permissions, preferences, evidence requirements and the actions that always need your approval.
Give work a way in.
Agree which messages, documents or events start the job and where its results should arrive. Use only the access the role needs.
Let the routine move forward.
The employee uses the business context to check, chase and complete permitted steps. It records what it does.
Receive the decision that needs you.
When a case is uncertain, unusual or needs approval, it brings a recommendation, supporting evidence and a clear question.
Correct it once. Review what carries forward.
Approved corrections become part of the role’s context. Changes to permissions or important rules remain a human decision.
The daily experience
A result where
you already work.
The intended handoff could be an email or a message in an existing business system, chosen when the role is set up. Routine cases finish within the authority you agreed. Exceptions arrive ready to decide.
A management surface would hold the rules, history, evidence and corrections. It supports oversight; it should not become another daily inbox to patrol.
Spontalo’s responsibility would include maintaining the role and checking its behaviour as the work changes. Your responsibility remains the business judgement and approval boundary.
Prompting and responsibility
Why not just use ChatGPT?
For a one-off question or a task you want to direct yourself, a general assistant can be a useful fit.
Spontalo’s intended model
The role carries the routine.
- Agreed events bring work to an assigned employee.
- Business rules and approved context are already in place.
- Permitted work advances; decisions arrive prepared.
ChatGPT also supports scheduled and event-triggered tasks. This comparison is about a managed business responsibility, not a limit on what a model can do. Read about ChatGPT automations.
Beside the systems you use
Why not another
SaaS tool?
Your accounting, booking and job management systems remain the places where the business runs. The employee would take responsibility for a small piece of work around them: receiving an event, gathering context and taking it to the agreed finish.
The aim is less manual operation, without asking you to replace your core systems. Each connection would need to be assessed for access, permissions and reliability. No specific integrations are available yet.