AI Features
AI Features
senSEi includes AI-powered features that work with your transcripts and deal data to save time and surface insights. All AI analysis in the webapp runs through Claude via your Claude Code MCP connection — so you control when it runs and what context it uses.
Running AI analysis
On any opportunity, click the Run button (or open the ✦ AI panel) to trigger analysis. Analysis does not run automatically — you decide when. This keeps token costs predictable and gives you control over what gets sent to the model.
The AI uses all available context for that deal: meeting transcripts, Gong summaries, linked Slack messages, email threads, Google Drive docs, attachments, and SFDC fields.
Meeting analysis
On any meeting node, click Analyse to extract structure from the transcript. The AI extracts: summary, problem statements, action items, next steps, and key quotes. If an opportunity is linked, it also extracts Command of the Message fields and writes them back to the opportunity.
Analysis runs automatically on meetings recorded via the mobile app.
Meeting Intelligence signals
Each analysed meeting produces structured Meeting Intelligence Signals: stakeholders with sentiment scores, objections, buying signals, technical requirements, competitive mentions, verbatim pain quotes, and open questions. These appear in the Summary → Meeting Intelligence sub-tab and are fed into all opportunity-level analyses automatically.
AI panel on opportunities
Click the ✦ AI button on any opportunity to open the AI panel. Actions available:
- Synthesize COM — Reads all meeting transcripts and fills the 7 Okta Force Management Mantra fields (Challenges, PBOs, Required Capabilities, Metrics, How We Do It, How We Do It Better, Proof Points) with verbatim evidence quotes. Results appear in Sales → COM.
- Generate BV Slides — Creates 7–9 executive Business Value slides from your CoM data and transcripts. Results appear in Sales → BV Slides.
- Technical Qualification — Synthesises the TechQual scorecard across all transcripts, populating Presales → TechQual.
- Generate Current/Future State — Builds Mermaid architecture diagrams and HTML visual diagrams for the customer's current and future state. Results appear in Presales → State Analysis.
- Analyze Presales Fields — Fills Okta SFDC presales fields (Stage, Confidence, Tech Diff, Concerns, Risks/Gaps, POC Status, Next Steps) in Presales → Presales. Writes leadership-quality meeting blurbs.
- Research Account — Generates 6–8 sales intelligence bullets about the company.
- Map Stakeholders — Identifies roles and relationships from transcripts.
- Suggest Next Actions — Recommends deal advancement steps as Kanban card suggestions.
Meeting prep
Click ✦ Prep on any upcoming meeting node to generate: prep bullets (specific talking points based on deal context and prior meeting summaries), a personalised icebreaker, and industry insights relevant to the call. The Meeting Prep agent also runs automatically 60–90 minutes before a calendar-synced meeting and surfaces the notes in the Agent Inbox.
Attachment analysis
Attach PDFs, slide decks, Word docs, screenshots, or diagrams to any opportunity or meeting via the Attachments panel. Click ✦ Analyze on any uploaded file to extract AI insights directly — results are stored as a note under the attachment and fed into subsequent analysis runs. Useful for competitive briefs, RFx documents, customer architecture diagrams, or vendor questionnaires.
Follow-up emails
After analysis, click Generate Follow-up to draft a professional client follow-up email. The Follow-up Agent (see Agent Inbox) also does this automatically after meetings.
AI chat assistant
Press ⌘J / Ctrl+J to open the AI chat panel. Ask questions about accounts, deals, or meetings — the assistant has full context from your workspace.
Health score
Every opportunity has an AI-computed health score (0–100) based on: recency of last meeting, VBC Mantra completeness, close date risk, and activity velocity. Below 40 = red; 40–70 = yellow; 70+ = green.
Real-time coaching (mobile)
During a live recording on the mobile app, AI analyses the last 2 minutes of transcript every 30 seconds. When it detects a Mantra element — a challenge, PBO, required capability, metric, or competitor mention — a non-intrusive banner appears. Each signal maps to a VBC Mantra field you can capture immediately after the call.