AI for Sales and Customer Support
Answer from approved knowledge and connect each conversation to the right next step.
BasisAssistant is an AI-powered sales and customer support assistant for website interactions. It works with the knowledge and rules approved by your organization, so users can ask questions, receive guidance, and provide the information needed to continue their journey.
It goes beyond a basic FAQ chatbot. A conversation can move into lead capture, a demo request, meeting booking, ticket creation or routing, or a controlled handoff to a human specialist when human judgment or action is required. You can start without replacing your current website or adopting the entire product ecosystem at once. BasisAssistant supports a Land–Validate–Expand approach: begin with one defined use case, review what happens in practice, and expand the scope only after the value and limitations are clear.
Many customer interactions begin with a simple question: Which product is right for me? How does this service work? Where should I submit my request? Who should I speak to next?
When every answer depends on a person being available, some requests wait unnecessarily. Sales and support specialists also spend time repeating the same basic information. Answers may vary between people or departments, and the contact details or context needed for follow-up may not be captured consistently.
BasisAssistant connects the initial response to approved organizational knowledge and gives the conversation a clearer path forward.
A useful starting point is a limited use case with an outcome you can evaluate.
If your team handles a high volume of recurring questions, support requests, or cases that may need escalation, BasisAssistant can provide the initial response from approved knowledge and move the conversation to a person when required. This gives customers a clearer first step while allowing specialists to focus on situations that genuinely need human involvement.
If the main challenge is understanding what a visitor needs and collecting useful information before sales follow-up, BasisAssistant can gather the relevant details during the conversation. The sales team receives more context for the next interaction instead of starting with an incomplete inquiry.
If customers or employees have to search across pages, files, FAQs, or guides to find an answer, BasisAssistant can provide a more direct route to approved knowledge. The result is faster access to relevant information while the organization retains control over which sources the assistant may use.
Every assistant should have a defined scope, trusted sources, and a clear next action.
Before BasisAssistant is put into use, the organization defines the problem it should handle, the sources it may use, and who is responsible for its knowledge and relevant settings.
During a conversation, the assistant works with the sources approved for that use case. When more context is required, it can ask the user for additional information before determining the next step.
If a reliable answer is not available, confidence is low, the request falls outside the approved scope, or a human decision is required, BasisAssistant should not present an unsupported answer as certain. It can ask a follow-up question, make the uncertainty clear, or move the conversation to a human specialist.
Set the topic, audience, response rules, and expected next action.
Make relevant website pages, FAQs, catalogs, documents, files, or structured information available within the approved knowledge scope.
BasisAssistant responds from those sources or asks for additional information when more context is needed.
The conversation can continue into sales, support, ticketing, or a human handoff.
Measure the use case against the KPIs selected for that specific scenario.
Response quality depends on the quality of the source, the scope, and the ownership behind it.
BasisAssistant can work with sources such as documents and files, website pages and FAQs, structured information, product and service data, and information made available through approved integrations.
The knowledge scope should not include raw passwords or secrets, private keys, complete payment-card information, or data the organization does not have the legal or internal authority to use. Personal, confidential, financial, medical, or identity information should only be processed when there is a valid business need, an appropriate legal or organizational basis, and suitable access control.
Before starting, define:
Every use case should have a defined problem, required data, next action, and success measure.
BasisAssistant can provide an initial response from approved knowledge. When the situation requires human review, judgment, or action, the conversation can move to the appropriate specialist.
Value in practice: Customers receive a more consistent first response while specialists stay focused on cases where human involvement matters.
BasisAssistant can collect contact details, the reason for the inquiry, and the initial information required by the sales process during the conversation.
Value in practice: Sales can continue with more useful context instead of starting from an incomplete lead.
Users can be guided toward a more relevant product, service, or course based on their stated needs. The guidance must remain grounded in approved product, service, or course information.
Value in practice: Users can narrow their options without manually reviewing every available page or offering.
BasisAssistant can retrieve answers from relevant approved sources. When a citable source is available, the relationship between the answer and its source can also be retained.
Value in practice: Customers and teams get a faster, more controlled route to organizational knowledge.
The conversation can collect the initial information required for a support request and pass that context into the ticketing path for continued handling.
Value in practice: The receiving team starts with better initial context, reducing the need for the user to repeat the same explanation.
For requests that require specific information or documents before review, BasisAssistant can guide the user through the intake process step by step.
Value in practice: The receiving team starts with a more complete request before evaluation begins.
Control knowledge, responses, access, handoff, and outcomes through a defined operating model.
These six capabilities form the approved core capability set for BasisAssistant. Each has a dedicated page for a more detailed evaluation.
Use sources approved by your organization as the basis for responses. This keeps the assistant connected to information selected for your business rather than relying only on general-purpose knowledge.
Move the conversation to a person when the user requests it, confidence is low, the topic is sensitive, the request falls outside policy, or the next step requires human judgment or action.
When an approved citable source is available, the response can retain a connection to that source. This makes it easier for authorized users to understand where an answer came from, while restricted sources remain subject to access controls.
Collect contact details, the user’s needs, and the initial information required for follow-up during the conversation. Sales receives a lead with more useful context for the next step.
Control access to knowledge sources, settings, reports, and management areas based on organizational roles. This helps separate responsibilities and restrict access to information according to the user’s role.
Review assistant usage and the outcome of a use case against the metrics defined for that scenario. The purpose is to connect usage to an observable business result without promising the same improvement for every customer.
BasisAssistant should work within the knowledge, roles, and policies defined by your organization.
Each organization works within its own organizational space, and user access can be controlled according to role. Customer data, documents, and source materials remain customer-owned. Customer data and content are not used by default to train public AI models.
Public connections and APIs use encrypted communication. Stored data is also protected through encryption controls at the storage layer. The exact control set can vary according to the deployment model and contractual scope.
If a reliable answer is unavailable, BasisAssistant should not manufacture certainty. A direct request for human help, low confidence, a sensitive or out-of-policy topic, a need for human action, or an unsuccessful response path can lead to Human Handoff.
Fit depends more on the problem, the available knowledge, and execution readiness than on industry alone.
| Higher Fit | Lower Fit |
|---|---|
| Meaningful volume of recurring questions, requests, leads, or support interactions | Limited usage with no defined recurring problem |
| Reliable sources such as FAQs, catalogs, website content, or internal guidance | Unreliable information with no clear owner |
| A defined goal such as response, lead qualification, ticketing, intake, or guidance | General interest in AI without a defined outcome |
| A manager or internal owner responsible for decisions and execution | No accountable person for the initiative |
| Willingness to start with a limited scope and measure the result | Expectation of a certain result without sufficient data or scope |
| Acceptance of staged expansion | Expectation of unlimited customization from the beginning |
These five industries are identified as priorities in the approved product and site architecture.
There is no single KPI that applies to every BasisAssistant deployment.
Success criteria should be selected before the use case begins and should match the actual business problem. The product architecture provides examples of useful KPIs by scenario, but the target value for any metric should be determined for the scope of the individual project.
| Scenario | Example Measures |
|---|---|
| Customer response and Human Handoff | First-response time, resolution rate, handoff rate |
| Lead capture and qualification | Qualified leads, completion rate, meeting rate |
| Product or course guidance | Guidance rate, answer usefulness, conversion |
| Ticket capture and routing | Ticket creation time, classification accuracy |
| Intake and document collection | Request completeness, reduction in follow-up loops |
| Recruitment and coordination | Screening time, resume completeness, interview scheduling |
None of these measures is presented on this page with a guaranteed target or improvement percentage.
Choose one limited problem, such as customer response, lead qualification, or enterprise knowledge access. Define the approved sources, response rules, Human Handoff conditions, and KPIs for that scenario. Starting with BasisAssistant does not require a full website replacement.
Review the quality of the responses and the practical outcome of the selected use case. The purpose of this stage is to understand the real value and limitations before increasing the scope.
After the first use case has been validated, the project can expand into approved integrations or related business processes. The exact scope and data flow of each connection should be reviewed for that specific integration.
No. Answering recurring questions is one use case, but BasisAssistant can also guide users from approved organizational knowledge, collect the information needed for the next step, and connect the conversation to an action or a human specialist.
No. Its role is to increase response capacity and reduce repetitive work. Requests that require investigation, experience, judgment, or human action can be handed to the appropriate person.
No. BasisAssistant is defined as an independent entry point and does not require a full website migration before you can start. The deployment method should be selected according to the scope and architecture of the individual project.
Approved sources can include website pages, FAQs, documents, files, structured information, and other sources the organization makes available for the use case. The organization should define which sources are trusted and who is responsible for managing them.
BasisAssistant should not present an unsupported answer as certain. It can make the uncertainty clear, ask a follow-up question, suggest an appropriate source, or transfer the conversation to a human specialist.
No. According to the approved security guidance, customer data and content are not used by default to train public AI models. Uses beyond direct service delivery must remain subject to the applicable customer settings or contractual controls.
The customer owns the input data, documents, and source materials. Their use must remain within the applicable permissions, contractual terms, and third-party rights.
Yes. Each organization has its own organizational space, and access to knowledge, settings, and reporting areas can be separated according to user roles. Some advanced controls may depend on the deployment model or contractual scope.
Choose one sales, support, knowledge, or request-intake scenario and see how BasisAssistant handles that specific path.
