What does Skai Control do for a sales department?+
The system helps analyse calls and chats, check compliance with standards, identify missed actions and prepare feedback for sales representatives. It focuses on specific criteria as well as the overall score: needs discovery, solution presentation, objections and agreement on the next step.
Can we use our own checklists and sales scripts?+
Yes. Criteria are tailored to your processes, stages and enquiry types. Before launch, define what satisfies each criterion, how partial completion is scored and which exceptions apply. We validate the configuration on an agreed sample.
How is a conversation score explained?+
The review combines a transcript, criterion-level scores and explanations. A manager can compare each conclusion with the conversation. The final percentage is calculated from the configured scale; it does not measure AI accuracy or the probability of a successful sale.
Can chats be analysed alongside calls?+
The product supports calls and chats. For each source, we check how history is retrieved, message completeness and links to the employee or deal. Written communication criteria are adapted separately from telephone conversation criteria.
Which CRM and telephony systems are supported?+
The original system links sales representatives to Bitrix24 and includes leads, deals and funnel sections. For your configuration, we clarify permissions, access to recordings and messages, limits and exchange formats. We assess other CRM and telephony systems based on their interfaces and available data.
How should different sales representatives' results be compared?+
Comparison is meaningful when criteria, customer types, stages and periods are comparable. Sample size and exceptions must also be considered. A summary helps choose a review topic but does not explain the causes of differences on its own.
What if AI misunderstands a conversation?+
These cases go to a responsible employee for review. Causes may include recording quality, an ambiguous criterion or context absent from the communication. During the pilot, we agree review procedures and use discrepancies to refine evaluation.
What data is needed for a pilot and how is pricing determined?+
We need a process description, a sample checklist and an agreed communication sample. Access and processing arrangements are defined separately. Cost depends on call and chat volume, sources, configuration depth, integrations and hosting requirements. We prepare a proposal once the scope is clear.
Can we create quality criteria ourselves?+
Yes. Each criterion has a name, a description of what counts as fulfilment, an evaluation method and a weight. Checks can use a language model, keyword rules or a combination. Criteria can be edited or disabled. Before applying them to the team, it is useful to agree their interpretation using conversation examples.
How do rule-based, LLM and hybrid checks differ?+
A rule suits a formal condition, such as a required word or phrase. A language model evaluates meaning and context: whether needs were identified, terms explained and the next step agreed. A hybrid combines formal checks with content evaluation. The choice depends on what needs to be checked.
Can we build our own sales quality dashboard?+
Yes. Views can be created for specific roles: leaders need trends and team comparisons, quality teams need deviations by criterion, and sales representatives need their conversations and development areas. Widgets, periods and breakdowns are selected for the task. Builder requests are written in ordinary language; available metrics come from connected and prepared data.
How do checklist changes affect comparisons between periods?+
When criteria or weights change, it is important to know which rules produced a score. Period comparisons use comparable rules and conversation groups. We define historical reassessment and checklist versioning procedures when configuring the quality control process.