Run every corporate meeting with an analyst in the room.

Schedule, roster, engagement, financial documents, action items and source files live together on the meeting, so the full picture is one click away.

Press ⌘J anywhere and ask “which meetings went over budget and why?” or “who still hasn’t RSVP’d?”. The assistant reads the data behind the screen you are on and answers with the numbers.

Event managers run the logistics, finance partners track spend, executives get briefings, and attendees participate from the companion iOS app.

Each client’s data is isolated by design, auditable, and never used to train models. Enterprise clients can run in a fully dedicated environment.

Filter by status and type, search titles, and see attendance, engagement and financial position at a glance. The dashboard shows a date-ordered list of upcoming and recent meetings, live meetings in progress, and six-month trends.

Sessions, breaks and activities with times, rooms and speakers. Engagement is tracked per session, so you can see exactly where attention peaked or dropped.

Drop an agenda PDF, a run-of-show spreadsheet or a calendar export. Sessions are extracted, resolved to the meeting’s dates and time zone, reviewed, then added. Multi-day agendas and parallel tracks are handled.

Follow-ups with owner, due date and status, tracked per meeting and across the organization, with overdue items highlighted and counted into the engagement follow-through score.

Every action item keeps a complete, permanent timeline: who created it, every change to status, owner, due date and details (with before and after values), notes from the team, AI assistance and deletion. History is recorded by the database itself, so changes from any screen, the AI or the mobile app are captured, and it cannot be edited or deleted.

Suggest the follow-ups a meeting still needs from its questions, feedback, notes, roster and financials, and add the ones you pick. On any item, the AI drafts a follow-up to the owner, plans next steps, or proposes a status, due date or owner change you can apply in one click. Each result and its reasoning is saved to the item’s history.

Add people by email, individually or in bulk. Existing users are linked automatically, and the companion app shows them the meeting once they sign in.

Upload a registration export, spreadsheet or a photo of a sign-in sheet. People are de-duplicated by email with roles and RSVPs inferred; rows missing an email are flagged for completion.

Everyone who has been on a roster, plus prospects not yet invited, with how many meetings they were invited to, accepted and attended. Your own team (app users) is kept out, so the directory is only about the people your meetings are for.

Track potential customers toward future meetings on a board of stages: lead, contacted, qualified, proposal, won and lost. Each entry carries company, estimated value, owner, target meeting, next step with due date and notes; drag cards between stages, and overdue next steps are highlighted. Past attendees can be added in one click, and anyone can be invited to an upcoming meeting from their profile.

Attendees check in and out, ask questions, answer polls, react, rate sessions and leave post-event feedback from their phone. Everything they do feeds the engagement model in real time. See the Mobile app section.

Show rate against accepted RSVPs, punctuality, and retention: how far through the meeting people were still present, from check-outs or their last interaction.

Breadth (the share of attendees who did anything at all) and intensity (weighted questions, polls, reactions and notes per attendee-hour). Breadth prevents a vocal few from masking a passive room.

Session and post-event ratings, including NPS when a 0–10 scale is used, blended with AI-classified sentiment of free-text feedback and questions.

Action items closed by their due date and the post-event response rate, so a meeting is judged on what it caused, not only how it felt.

Every score is shown against the organization’s median for that meeting type and against the previous instance (this quarter’s all-hands vs last quarter’s), with a 15-minute participation timeline annotated by session.

One click classifies feedback and questions into sentiment and a fixed vocabulary of themes (pacing, audio/visual, content quality, strategy, logistics…), stored on each comment so they are queryable, not just summarised.

Scan a QR code at the door or tap a geofenced prompt on arrival. Check-in and check-out stamp the roster in real time and feed attendance, punctuality and retention without a registration desk.

The meeting’s schedule as organizers maintain it, with rooms, speakers and tracks. Attendees build a personal schedule, get reminders before a session starts, and see changes the moment organizers make them.

Ask a question in the current session, upvote others, and see which ones were answered. Moderators can mark, merge and answer questions from the web app, and unanswered questions roll into follow-ups.

Organizers launch a poll from the web app and results appear on screen as answers arrive. Lightweight reactions let a room respond without interrupting the speaker.

A one-tap rating when a session ends, and a short survey with a 0–10 recommendation question and free-text feedback after the event. These power the sentiment dimension and net promoter score.

Slides, PDFs and handouts attached to a session open right in the app. Attendees highlight, draw and add notes on their own copy, which stays private to them and syncs across their devices, and can share a marked-up page with the presenter or a colleague. Organizers upload materials once and control whether a file can be downloaded or only viewed in the app.

Room changes, meal times and emergency notices pushed to attendees, with read receipts so organizers know the message landed.

Check-ins, the question queue, poll results and session-by-session participation update live on the meeting’s Engagement section while the event is running.

Agendas and personal schedules are cached, and interactions queue when venue connectivity drops, syncing when the device reconnects.

The app talks to the same data environment as the web application. Attendees sign in with the email they were invited with, see only the meetings they are on the roster for, can update only their own RSVP and interactions, and never see financials or other attendees’ private data. Access is enforced by the same row-level security policies and audited the same way.

The app records only the interactions an attendee chooses to make. No location tracking beyond an optional arrival prompt, no contact scraping, no advertising identifiers. Organizations can turn individual features off per meeting, and attendees can request their interaction history or its deletion.

A pilot build is planned for selected client events, followed by App Store release. Enterprise clients can distribute through their mobile device management. Branding per organization (name, colours, logo) is on the roadmap.

Document type, category, vendor, reference number, total, amount paid, currency, status (draft, pending, approved, rejected, paid, cancelled), issue, due and paid dates, notes and terms, a link to the original, and itemised line items.

Drop an AV proposal, hotel contract, catering order or invoice. The vendor, totals, dates, terms and line items are extracted for review before saving, and the original file is kept with the meeting.

Approved budget, forecast (all live documents), committed (approved contracts, POs, invoices and deposits), paid, and projected variance, with a breakdown by category and documents awaiting a decision.

Spend versus approved budgets rolls up to the dashboard and the meetings list, and the assistant can compare meetings, vendors and categories across the organization.

The original is kept in a private, org-scoped bucket and recorded with its type, size and who uploaded it. Links to originals are signed and expire.

Each document kind has a schema. The model fills it from the file and reports its confidence and anything it had to assume.

You see the extracted rows in an editable table before anything is created. Every created row keeps a link to its source document.

Each meeting lists every original with its status and what it populated. Deleting a document never deletes the rows made from it.

It knows which screen you are on. On a meeting page, “are we over budget?” means that meeting. On the dashboard it reasons across the portfolio.

Nine read-only tools cover meetings, attendance, engagement, financials, action items and their history, people and the pipeline. Answers cite the numbers they used and flag what is missing instead of guessing.

It calls out variances, pending proposals, overdue items, unanswered questions and low RSVP rates when they are relevant to the question.

Threads are saved per user and can be reopened and continued from the AI History page.

Powered by a leading enterprise AI model under a commercial agreement. Prompts and data are never used to train models. Zero-data-retention can be enabled per client, and the assistant can be turned off for a tenant entirely.

The assistant never changes data. It explains which screen to use for edits.

Pick a tool, drop in a document or type notes, optionally choose a client profile and the meeting it concerns, and watch the result stream in. Results are kept per user, can be reopened, copied, or downloaded as a branded PDF, and keep running if you navigate away.

A skill is versioned Markdown guidance with optional reference files (checklists, templates). Admins edit skills by hand or ask the AI to propose a revision, review the diff summary, and save a new version that can be restored at any time.

Describe a new tool in a sentence and the AI proposes its inputs, picks the right skills or drafts a new one, and fills in the configuration for review. Or configure it manually: file and text inputs, model, effort and output limits.

Reusable context about a customer, such as contacts, event history, preferences and compliance rules, applied to any run so results reflect who the work is for.

Members see only the tools an admin has granted them; org admins see every enabled tool. Grants, tools, skills and client profiles are per organization and covered by the same row-level security and audit log as everything else.

Import four ready-made tools and their skills (Event Contract Summary, Vendor Proposal Review, Event Brief Generator, Client Email Drafter) and adapt them to your voice.

Owners and admins invite colleagues by email with a role. Existing accounts are added immediately; others join automatically when they sign up with that email. Roles can be changed, members removed, invitations cancelled, and AI tool access granted per person.

Owner, admin and member for the organization; organizer, presenter, attendee and guest on each meeting. Financials are never visible to plain participants.

Rename the workspace, find the workspace ID for API integrations, and keep your name, title and department current.

Every create, update and delete on meetings, attendees, financials, documents, sessions and action items is recorded per organization with who did it and what changed. The log is append-only.

Server-rendered React. Server functions validate every input and run under the caller’s session; a JSON API serves the iOS app with bearer tokens.

Each tenant is one data environment: a relational database, authentication, private file storage and backups. Authorization lives in database policies, not application code, so the web app, the iOS app and the AI tools all obey the same rules.

A single AI integration powers the assistant (tool use, streaming), document standardization (structured outputs) and engagement classification. The server orchestrates; the model receives only the data the current user may see.

Stateless application servers deploy to any Node host or edge platform. Database changes ship as reviewed SQL; the schema and policies are version-controlled with the product.

By default organizations share a hardened pool project, isolated by forced row-level security on every table. Clients that need physical isolation get a dedicated project with their own database, storage, authentication users, API keys, backups and region. Moving a client between the two is a directory change, not a migration project.

Every record carries its organization, stamped by the database and locked so rows cannot be moved between organizations. Policies check it directly. Anonymous access to the data API is disabled; even service accounts are subject to the policies.

TLS for all traffic, AES-256 encryption at rest for databases, backups and files. Document links are signed and short-lived. Cookies are HTTP-only and secure.

Email and password or magic link today; SAML single sign-on per dedicated client, with SCIM-style provisioning on the roadmap. Mobile tokens are routed only to the project that issued them.

Append-only per-organization audit log of data changes; point-in-time recovery per project; structured server logs without personal data.

An automated isolation test creates two organizations and proves that every table, view, the audit log, file storage and anonymous access refuse cross-tenant reads and writes. It runs against staging on every schema change.

Only the fields needed for an answer are sent to the model, under the user’s own permissions. No training on customer data. Zero-data-retention and a per-tenant AI off switch are available.

In progress. Controls for access, change management, availability, confidentiality and incident response are implemented in the platform (RLS, audit log, reviewed schema changes, least-privilege keys). Our hosting and AI subprocessors hold SOC 2 Type II reports, available under NDA.

Data processing agreement available; EU data residency through dedicated projects hosted in EU regions; export and deletion on request; subprocessor list published.

Access, deletion and do-not-sell obligations supported. Customer data is never sold or shared for advertising.

Planned after SOC 2 Type II. Subprocessors are ISO 27001 certified.

WCAG 2.1 AA target for the web application: keyboard navigation, screen-reader labels, sufficient contrast and large type.

MeetingIntel.ai is not intended for protected health information and is not offered under HIPAA business associate agreements.

Full export of all organization data (meetings, rosters, engagement events, financial documents, originals, action items, audit log) in open formats on request, or continuously through the API.

Data is retained while the contract is active. Backups follow the project’s point-in-time recovery window. AI conversation history can be cleared by users at any time.

On termination, the organization (or the whole dedicated project) is deleted on request, with backups expiring on the documented schedule and written confirmation provided.

Dedicated clients choose the hosting region for their project at provisioning time.

We provision your organization in the pool or a dedicated project, and you invite your team from Settings → Team.

Create the meeting, then import its agenda and attendee list. The schedule and roster are ready for review in minutes.

Add the proposals and contracts you already have. Forecast and committed totals appear immediately against the approved budget.

Attendees participate from the iOS app; organizers watch check-ins, questions and polls live and log anything captured elsewhere.

After the meeting, analyse feedback, compare the engagement score to your benchmark, and ask the assistant what to change next time.