Personal, executive, company and department AI assistants configured around the knowledge, permissions and recurring work of a specific role.
An assistant becomes useful when it knows what it is responsible for, which sources it may use and which actions it may take. We define those boundaries around the person or function it supports. Broad access and an open-ended instruction are not substitutes for a working role.
We connect the appropriate knowledge and task workflows, then design how someone asks for work, reviews the response and corrects the context. The assistant should make its limits understandable. Where information is missing or an action requires authority, it asks or escalates.
Prepare briefs, organise tasks, research a defined question or assemble information before a meeting. Scheduling and communications require specific permissions; the assistant does not assume authority from the fact that it can prepare a draft.
Help colleagues retrieve approved information and understand the next step in an internal process. Define source ownership, access boundaries and the route for correcting out-of-date or conflicting guidance.
Configure assistance for a function such as operations, sales or delivery. Connect the relevant records and recurring tasks, with rules for how work is created, reviewed and handed to the accountable person.
The sequence is shaped around Virtual Assistants, with scope and responsibilities agreed before implementation.
Define the responsibilities, users and boundaries. Distinguish answering questions, preparing work and executing actions. Prioritise a small set of recurring tasks that the role can support consistently.
Select sources, resolve obvious contradictions and define who maintains them. Configure the access appropriate to the role. Retrieval should not expose information simply because it exists somewhere in the company.
Integrate the agreed actions and make the review point explicit. Drafts, record updates and outbound actions may each need different levels of confirmation. Design for interruption and correction.
Test representative requests and inspect whether the assistant uses the right source, recognises uncertainty and stays within its role. Document how to update knowledge and change responsibilities.
Select a role to see how responsibility, knowledge and authority differ.
Organise a defined set of tasks and prepare research or reminders. Use the information the individual has authorised. Ask before an action changes an external record or communication outside the agreed automatic scope.
These are the delivery components we define together. The depth, access and handover arrangements follow the agreed engagement.
Responsibilities, supported tasks and explicit action boundaries.
Agreed sources, access rules and maintenance responsibilities.
A usable interaction and the approved task connections.
Representative checks and instructions for changing the role.
The opportunity is to make recurring preparation and retrieval reusable, so specialist attention can remain on decisions and relationships. We assess the work the assistant is expected to support, the time needed to review it and the cost of maintaining its knowledge. A narrowly defined role is easier to evaluate than an assistant presented as capable of handling everything. Wider responsibilities follow evidence from actual use.
A chatbot centres on a conversation. An assistant has an ongoing responsibility and may prepare or coordinate defined tasks. Both can use approved knowledge, but the action boundaries differ.
The design defines who can use each source or action. The exact access model depends on the selected tools and the systems connected during implementation.
Assign source owners and a knowledge update process. Changes to responsibilities or permissions need review as well as changes to the underlying documents.
Capture the objective, current process and first boundary. Download a working brief or copy it for a project discussion.
This tool runs in your browser. It does not submit or store your information.