Conversation analysis
Produces the transcript of a conversation and extracts a summary and customer experience (CX) insights from it.
- Speaker-separated transcript
- Executive summary
- Customer experience insights
VexCore Teknoloji product · AI-assisted meeting assistant
VexCoreSI
From meeting straight to backlog: joins your meetings as an AI participant, understands the conversation and manages your backlog on your behalf.
What is VexCoreSI?
It joins the meeting and carries the work discussed into your tracker itself.
Teams spend hours in meetings every week, and a large share of the work that comes out of them is lost. Notes stay incomplete, "who was going to do this?" is asked again, and decisions reach the tracker late or never.
VexCoreSI joins the meeting like a participant, listens, and extracts the summary, decisions and tasks. It then opens the items in Jira or Azure DevOps assigned to the right person with a due date, updates the status of existing items and drops the links into the meeting chat.
- 01Join
The meeting link is pasted into the portal; the assistant joins as a guest and introduces itself in the chat.
- 02Listen
Speech is recorded and speakers are distinguished.
- 03Analyse
Summary, decisions, tasks and status suggestions are extracted.
- 04Approval mode
Fully autonomous or human-approved; by chat command or one click in the portal.
- 05Transfer
Items are created or updated; links land in the chat.
The problem
Why does work from meetings get lost?
The same five situations repeat in most teams.
- 01
Incomplete notes
The note-taker cannot take part in the discussion; notes stay incomplete or vary from person to person.
- 02
Ownerless tasks
"Who was going to do this?" is asked again after the meeting.
- 03
Late, context-free decisions
Decisions reach the tracker days later, detached from their context, or not at all.
- 04
Stale statuses
"This is done" or "it's in testing" is said in the meeting but never updated in the system.
- 05
The absentee's burden
Someone who missed the meeting has to watch the whole recording to find out what was discussed.
What it does
Three capabilities, one analysis engine
Conversation analysis and the smart meeting participant share the same engine; backlog management is what sets the product apart.
Smart meeting participant
Takes notes on your behalf, finds the points you missed and analyses agenda adherence, deviations and silent speakers.
- Notes and missed points
- Agenda adherence and deviations
- Speaking-time distribution
Backlog management
Creates items in Jira or Azure DevOps, assigns them to team members, sets dates and updates the status of existing items. Drops the links into the meeting chat; no need to open the portal.
- Create and assign to the right person
- Due dates and status updates
- Links in the meeting chat
The difference
How an item is created and updated
Summarising a transcript is a common capability. What sets VexCoreSI apart is that it carries the work discussed into the tracker itself.
- Create itemOpens a commitment made in the meeting as a new task in Jira or Azure DevOps."I'll fix the payment screen bug by Thursday" → a task is created.
- Assign to the right personMatches names mentioned in the meeting to team members and assigns the task directly. Team members with a recorded voice signature are also recognised by voice.A name mentioned in the meeting → assigned to that team member.
- Set due dateTurns the deadline mentioned in conversation into the item's due date."By Thursday" → due date Thursday.
- Update existing itemFinds the open item in the system and updates its status; it does not only create new tasks."The reporting module is done" → the item moves to Done.
- Evidence quoteEvery task carries a quote with the moment it was said and the speaker. Clicking the quote jumps to that exact point in the recording.[12:34] Speaker: "I'll finish this by Friday"
- Link sharingDrops the links of created or changed items into the meeting chat.The team reaches the item from the chat in one click.
Four outputs at the end of the meeting
- 01Executive summary
- What was discussed, in a few paragraphs.
- 02Decisions taken
- Item by item.
- 03Tasks
- Who does what, by when.
- 04Status updates
- Which existing work progressed or was completed.
You stay in control
The organisation sets the level of autonomy
Human approval is not mandatory. VexCoreSI can run fully autonomously or leave every action to the team's approval.
Fully autonomous
Creates, assigns, dates and updates the items discussed directly.
- Who decides?
- The assistant; results appear in the chat with links.
- How?
- No approval step.
- Who is it for?
- Teams that want speed for low-risk, routine backlog work.
Human-approved
Posts suggestions as a numbered list to the meeting chat and the portal, then waits for the team's decision.
- Who decides?
- Authorised people; approvers are defined per team.
- How?
- By typing /approve 1,3, /approve all or /reject 2 in the chat, or with one click in the portal.
- Who is it for?
- Teams that keep the final say for process, audit or habit reasons.
In both modes
- 01Even if the same item is approved from the chat and the portal at the same time, it is created only once.
- 02Every task and status update carries a quote and timestamp from the meeting; why it was created is always clear.
- 03Jira and Azure DevOps credentials are never shown in the interface.
Privacy and deployment
Meeting data can stay inside your organisation
Speech recognition and AI models can run on the organisation's own server. Cloud models can be enabled when higher accuracy or speed is wanted.
- Fully on-premises
- Speech recognition and AI models run on the company's own server. In this mode no meeting data leaves the company.
- Optional cloud
- When higher accuracy or speed is wanted, OpenAI, Anthropic or Google models can be enabled with one click in the portal.
- No meeting setup
- The meeting owner does not need to configure anything; the assistant joins as a guest and also works in meetings hosted outside the organisation.
- Credentials protected
- Jira and Azure DevOps credentials are never shown in the interface.
When the assistant joins, it introduces itself in the chat; participants know the meeting is being recorded.
Platforms and integrations
Supported today, and the roadmap
Only integrations that work today are listed as supported; planned ones are listed separately.
Today
- Meeting platforms
- Microsoft Teams, Google Meet
- Work trackers
- Jira, Azure DevOps
- AI models
- On-premises (local) models or OpenAI, Anthropic, Google Gemini
- Language
- Turkish-first
- Teams
- Each team's meetings, members, approvers and project are managed separately
Roadmap
- Live analysis during the meeting: instant summary and candidate tasks
- Zoom and Webex support
- Advanced speaker recognition using the "active speaker" signal from the meeting screen
- Enterprise management: user sign-in, role-based authorisation and audit logs
No dates are given for the roadmap; priorities are set with customer feedback.
Who it is for
Every team whose meetings produce work
- Software and product teams: sprint planning, daily stand-ups and architecture reviews
- Project managers and Scrum Masters who want to automate action tracking
- Managers who want to reach the summary and decisions of meetings they missed
- Organisations that care about data privacy and do not want meeting content to leave the company
What your team gains
What changes after the meeting
- 01Note-taking and entering tasks afterwards disappear.
- 02Commitments made in a meeting are recorded; every task has an owner and a source.
- 03The tracker stays up to date in step with the meeting.
- 04People who missed the meeting can access the summary and the task list.
Example scenario
A sprint planning meeting
Both modes end with the same result; only the approval step differs.
- 01
Commitment
A developer says "I'll fix the payment screen bug by Thursday"; someone else says last week's reporting work is done.
- 02
Suggestion
VexCoreSI posts two actions to the chat: a new task for the payment screen bug with owner and due date, and a "Done" status update for the reporting item.
- 03
Decision
In fully autonomous mode the actions run immediately. In human-approved mode the team lead types /approve all in the chat.
- 04
Result
A new assigned task appears in Jira and the reporting item moves to Done; the links are shared in the chat.
- 05
Evidence
Both records carry the relevant quote and timestamp from the meeting.
Frequently asked questions
About VexCoreSI
Does our data leave the company?
Not in on-premises mode: speech recognition and AI models run on the company's own server and no meeting data leaves the company. If a cloud model is selected, processing uses that provider's model; the switch is one click in the portal.
What if it creates the wrong item?
In human-approved mode no item is created without approval. Every item carries a quote, the speaker and a timestamp from the meeting; clicking the quote jumps to that exact point in the recording. The organisation chooses the mode according to its risk.
Does the meeting owner need to do anything?
No. No special setup is needed; the assistant joins as a guest and introduces itself in the chat. It also works in meetings hosted outside the organisation.
Do we have to use the portal?
No. The assistant accesses the meeting chat directly: it writes which items it will create or change, performs the action and drops the links into the chat. The portal is used to add the meeting link, approve items, choose models and review records.
Does it update existing work too?
Yes. It finds open items in the system and updates their status; it does not only create new tasks.
How does it know who said what?
It distinguishes speakers by voice. Team members with a recorded voice signature are recognised by name, and names mentioned in the meeting are matched to team members.
We use Zoom. Does it work?
Microsoft Teams and Google Meet are supported today. Zoom and Webex are on the roadmap.
Does it work in Turkish?
Yes, the product is designed Turkish-first.
Let no commitment made in a meeting be lost.
Let us plan a short demo around your team's meeting and tracking setup.