Tangensys • Plan, deliver, measure
Custom AI assistants
Make the right information available to the right person. Retrieval, permissions and source updates are central to a useful business assistant.
The work to plan together
A scope shaped around this requirement.
- Content preparationScope
- Retrieval setupReview
- Assistant interfaceCheck
- Approved informationWhat the tool is allowed to use
- Routine stepHandled automatically
- Human hand-overYour team takes over when needed
- Regular reviewOutputs checked and corrected
In short
What does Custom AI assistants involve?
Custom AI assistants answer questions from your own documents and data, for staff or customers, using retrieval (RAG) so answers stay grounded.
When Custom AI assistants is a useful fit
Support teams
Use approved support material and maintain an accountable handover route.
Internal knowledge bases
Respect access permissions and document which source material each answer can use.
Product documentation
Keep versions and compatibility information current so outdated instructions are not returned.
Problems worth addressing first
Use these situations to identify the work that deserves attention before expanding the scope.
- Staff cannot find information
- Customers ask questions docs already answer
A practical example
An internal assistant can answer a policy question from the current approved document and link back to it, while restricting documents the staff member cannot access.
This is an illustrative scenario to explain the approach, not a claimed client project.
The work we can include
The final quote identifies the deliverables, quantities, approvals and exclusions. Select the work that serves your goal rather than assuming every item is required.
- Content preparation
- Retrieval setup
- Assistant interface
- Access control
Where we usually begin
Retrieval does not eliminate incorrect answers. Document preparation, access control, model charges and evaluation are included only as specified.
Practical detail
An answer should lead back to its permitted source
An illustrative internal assistant answers a leave-policy question from an approved document. It must use the right version and respect the requester’s access.
| Part of the work | Example output | How to review it |
|---|---|---|
| Source set | Approved policy, version date and document owner | Superseded or unapproved documents are identified. |
| Answer check | A response with the relevant source reference | The cited section supports the actual answer. |
| Access test | A user requests a restricted document | The assistant does not retrieve or reveal information outside that user’s permissions. |
| Unknown answer | The source does not resolve the request | The assistant states the limit and routes the question to the owner. |
The handover includes source updates, evaluation questions and access responsibilities. A citation can be wrong, so its relevance is part of the test rather than proof by itself.
Build around an approved knowledge and action boundary
A custom assistant can help staff find information or carry out a small, clearly defined task. Define which sources it may use and which systems it may change. Access to a document library does not automatically justify access to every customer record or operational action.
Use representative questions and an evaluation set before expanding the feature list. Retrieval, model behaviour and the interface each influence the result. A useful first version may answer one internal process question with sources rather than attempt a general-purpose agent.
Check unsupported requests, stale sources and permission boundaries. Record which actions require human approval and how a user reports an incorrect answer. Deployment depends on confirmed technical capability and an agreed operating owner.
Free AI project review
Start with a clear scope and a practical next step.
Tell us what you want to improve and share the relevant website or project details. We will review the starting point, recommend a suitable scope and provide a custom quote.
The initial review is free. Delivery, third-party charges and ongoing support are quoted separately.
Process
How the work is delivered
Review the starting point
Review the current setup, examples of the problem and the information needed to agree the work.
Agree scope and sign-off
We start with real examples of the task and approved information. A small, clearly defined pilot tests usefulness, failure cases and the level of human oversight before wider rollout.
Deliver and review
Evaluation includes inaccurate answers, unavailable tools and uncertain requests. Provider charges, monitoring and updates are defined separately; the business keeps an accountable human owner.
Hand over the next steps
Review the agreed outputs together. Confirm what your team maintains and what needs a separate ongoing arrangement.
What progress should mean
Measures that match the work
Test answer quality, source relevance, permission enforcement and unresolved questions. Retest when documents or model configurations change.
A record of delivery
See what was completed, reviewed and still dependent on access or approvals. Project delivery and business results are reported separately.
Clear limits in the data
We explain what can be verified and where a result is only an estimate. Contact clicks, conversations and sales are different actions.
Scope and responsibilities
You keep ownership of your business accounts and approved deliverables under the agreed contract. Any third-party licences, subscriptions or usage fees are identified before you commit.
What makes a custom assistant different from a generic chat tool?
The difference is the approved source access, workflow and controls built around the task. A model alone does not supply those. The proposal should identify integrations, evaluation, permissions and maintenance rather than only the chosen model name.
Can one assistant use documents from several departments?
Yes where the architecture and access rules support it. Document ownership and permissions must remain clear across sources.
Does adding more documents always improve answers?
No. Conflicting, stale or poorly organised sources can make retrieval harder. Use approved information and test the questions people actually ask.
What is RAG?
Retrieval-augmented generation: the assistant looks up your documents before answering.
How is Custom AI assistants priced?
We quote against your goals, current setup and agreed scope. The quote separates delivery from third-party costs and ongoing support. You receive a custom quote within 2 working days.
How long does the work take?
The timeline follows the size of the agreed scope, access, content and approvals. We confirm milestones in the quote; a simple change and a full implementation need different schedules.
How will we communicate?
Your point of contact agrees a review rhythm with you. We use email, WhatsApp and scheduled calls, with decisions and scope changes recorded clearly.
What do you need to get started?
Share your current website or system, your main goal and examples of what is not working. We then agree the access and information needed for discovery.
Is everything included in the free review?
No. The free initial review helps define the next step. Detailed research, design, implementation and ongoing support are included only when stated in your quote.
Explore the related services
Let’s agree the right next step.
Tell us your goals and current setup. We’ll recommend a practical starting point and provide a custom quote.
Custom quote within 2 working days