> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getvector.me/llms.txt
> Use this file to discover all available pages before exploring further.
# Meeting analysis
> Add transcripts and tickets from sales, research and support, and let AI connect what customers said to your features.
Sales calls, user research sessions and support tickets are full of signal about your product — but only if someone reads them and connects what was said to what you're building. Meeting analysis does that reading: you add the transcript, and AI finds each moment a feature came up and links it to that feature.
Open it from **Meeting analysis** in the **Track** group of the left menu. What it finds lands on your features, in the **🗣 Meeting analysis** section of each [feature dossier](/features/feature-dossier).
## Add a meeting
On the **Add new** tab:
The **team tag** is required. It tells Vector what kind of meeting this is, and which analysis to run on it:
| Team tag | What it's for |
| -------------- | ----------------------------------------------------------------------------- |
| Sales team | Demo-call transcripts — what blocks the deal and which features are asked for |
| UX Researchers | Feature-presentation transcripts — how users react to features |
| Support team | Support tickets — product pain points and feature requests |
Paste the transcript or ticket text, or drag files onto the drop area — PDF, DOCX, TXT or Markdown. Vector extracts the text.
Add **Custom tags** — your own labels, on top of the team tag. If the meeting was about one specific feature, set **About a feature**: the meeting then also shows in that feature's dossier.
If a new document looks like one you already added, Vector says so and shows how close the match is. **Delete document** drops the new copy; **Process anyway** keeps both.
## See what customers said about your features
Analysis starts as soon as a meeting is added. It finds each moment a feature came up and turns it into an insight: the quote, a short rationale, and a **context tag** for how the feature came up — **strength**, **weakness** or **need**.
Open a meeting to read its insights two ways. **Transcript** shows the text with each insight highlighted in place. **Insights** lists them in two groups:
* **Linked features** — insights matched to a feature in your [catalog](/features/features-in-domain). Each also appears as evidence on that feature.
* **Unmatched signals** — insights that didn't match any existing feature. Vector keeps them rather than forcing them onto the wrong feature, and they reach the [candidates](/features/candidates) inbox, where you decide whether they're a new feature.
To analyze a meeting again — for example after you've changed the team's instructions — click **Re-run** on its page.
## Check what AI found
Every link is a starting point, not a verdict. For each insight you can:
* **Confirm** it as correct
* change its context tag, if AI called it a weakness when it's really a need
* set its confidence to low, medium or high
* **Choose feature(s)** to link it to a different feature, or to several
* **Dismiss** it — it's removed from the feature and moves to **Dismissed** in the meeting
To review a long meeting quickly, use **Confirm all** or **Dismiss all** in the **Transcript** view: they act on every insight under the current filter, and **Confirm all** skips unmatched signals.
If AI missed something, select the text in the transcript and link it to a feature by hand.
## Find a meeting later
All meetings live in one list, on the **All meetings** tab. Filter it by **Team** or by your own **Tags**, and search it. Each row shows whether the meeting has been processed.
To keep a team's meetings organized, filter to that team, create a folder, and drag meetings into it.
## Link a meeting to a client
When a meeting is with a known customer, link it to their profile in **Clients & Leads**. Vector tries to detect the company during analysis; you can change the link, remove it, or create a new profile from the meeting.
## Change how a team's meetings are analyzed
Each team tag runs its own analysis. On **Add new**, open **AI pipeline** under the team tag to edit the instructions the analysis follows, and the template that turns an insight into the evidence note shown on a feature. **Save pipeline** applies the change to that team only.
## Where to go next
Keeping the catalog these insights attach to clean.
Acting on what customers keep asking for.