Procedures sit in SharePoint, rules in Bitrix24 or in emails, and material from earlier projects in Jira and in colleagues' heads, so staff ask the manager or do it their own way. The assistant answers in Teams from your documents with a link to the source, and a new hire gets up to speed in days.
Built for ISO 27001NIS2 DORA
In a professional services firm, knowledge piles up faster than it is organised. Procedures are written in SharePoint, but some updates were sent by email and never moved into the document. Internal rules and announcements sit in Bitrix24, if you use it, service descriptions exist in several Word versions, and decisions and notes from earlier projects are in Jira or Asana tasks whose structure only their creators understand. A new employee spends the first weeks asking colleagues in Teams and learning from answers that do not always agree. A veteran checks which rule is current before replying to a client, or answers the way they did last time.
The assistant brings together procedures and service descriptions from SharePoint, rules from Bitrix24 and earlier project material from Jira or Asana in one place, which staff ask in Teams in their own words: how to document a change to a service, what the validity period of a proposal is, whether we have done a similar project and who ran it. The answer comes with a link to the document in SharePoint, the Bitrix24 entry or the Jira task, and if the documents contain several versions, the assistant says so and points to the latest by date. What is not in the documents it does not invent. It suggests what to ask and what is worth writing down. We implement it one of two ways, depending on your environment: a ChatGPT Custom GPT on the OpenAI platform, where your documents are not used for model training, or Claude through the Anthropic API or Amazon Bedrock in an EU region with your SharePoint. We integrate with Bitrix24, Jira or Asana through an API, file export or an email flow, depending on your plan. Access rights are inherited from the existing ones, so an employee sees answers only from the documents and projects they have access to. At the start we also tidy the document structure and remove outdated versions, because answer quality depends on it.
Four steps from input to a record in your system, with human approval wherever there is risk.
Together we review where knowledge lives: SharePoint, Bitrix24, shared drives, email, Teams channels, Jira or Asana projects. We put current documents into an agreed structure in SharePoint, archive outdated versions, and point the assistant only to sources it may rely on.
The employee asks a question in Teams or the browser. The assistant searches only your documents in SharePoint, Bitrix24 entries and Jira or Asana tasks that the employee is allowed to see, and assembles an answer from what it finds. If there are several versions, it notes the date and points to the latest.
The answer is short, with a quote and a link to the document in SharePoint, the entry in Bitrix24 or the task in Jira. The employee can open the original and check it. Actions that affect a client or money the employee carries out personally under the procedure, and the assistant only shows what that procedure is.
When there is no answer, the assistant says so and suggests what to ask. Each month a list of unanswered questions shows which procedures are worth writing down or updating in SharePoint. When a document is updated or a Jira task is closed, the assistant answers from the new version with no extra steps.
A review of existing knowledge sources in SharePoint, Bitrix24, Jira or Asana, selection of current documents, an agreed structure, and a rule for who updates it and how.
A ChatGPT Custom GPT or Claude with your SharePoint documents and Jira or Asana project material, available in Teams or the browser, in your company's tone and with your access rights.
A set of questions and documents for the first days, so a new employee can find procedures in SharePoint, templates and earlier projects in Jira or Asana on their own.
A monthly report on what staff asked in Teams and what the assistant could not answer, serving as a plan for filling in the documentation.
ChatGPT Custom GPT or Claude + SharePoint documents + Bitrix24 rules + Jira or Asana project material via API or export. We choose the integration method after checking your plans. For more on this service and its limits, see the service page.
Search returns a list of documents or tasks, while the assistant answers the question and points to where in the document it is written. It understands the question in your own words, connects a SharePoint document with a Jira task and tells you when versions contradict each other. Search stays, but most questions get resolved faster.
Yes. Bitrix24, Jira and Asana have public APIs, through which the assistant reads rules, tasks and comments with the existing access rights. Where your plan limits the API, we use a periodic export to SharePoint or an email flow. We choose the integration method after a check at the start of the project, and SharePoint and Teams connect directly inside the Microsoft 365 tenant.
If your documents are in a Microsoft 365 tenant and it matters that data stays in an EU region, we usually suggest Claude through Amazon Bedrock with a SharePoint connector. If the team already uses ChatGPT and there are not many documents, a Custom GPT is a faster start. Both options work without training on your data, and we make the decision together based on your policy.
A proposal or contract is prepared by copying the last Word document and changing the details, and a mistake in a name or an amount is noticed…
Staff fill in Clockify or Harvest on Friday from memory, so some client hours are never invoiced.
After a client meeting, the account manager has to record what was agreed in HubSpot or Pipedrive, but usually does it in the evening from memory,…
Everything lands in the shared info mailbox in Outlook and the website form: pricing questions, document requests, complaints and serious enquiries.
This service is delivered by specialists with valid international certifications. Their expertise is confirmed by independent organisations, not by us.
AI systems are deployed with protection against data leakage and model abuse.
The AI security programme and its controls are managed following the ISACA methodology.
AI risks are identified and managed like any other IT risk.
We reply within 1 business day
