# HelpAlive

> HelpAlive is an AI agent that watches every user inside a SaaS product, intervenes the moment they're stuck with step-by-step in-app walkthroughs, and answers their questions in real time with full session context. One product replaces an analytics tool, an in-app chatbot, and a guide builder, so customers supervise outcomes instead of operating tools.

HelpAlive is built for B2B SaaS teams of 10–150 people who care about adoption and retention but don't have a dedicated product analyst. Setup takes minutes — drop one script tag into your app, call `identify()` after login, and HelpAlive starts capturing user behavior automatically and powering an AI Agent that's grounded in your documentation. Personal data is redacted twice — first in the user's browser, then again on our servers — so you get adoption insight without the privacy headache.

## How the dashboard is organized

The dashboard has three main sections, plus settings:

- **Agent** — train the AI assistant on your docs, customize its tone, design, and behavior, then publish. Five tabs: Train, Personalization, Test, Appearance, Configuration.
- **Studio** — work the Topics inbox of unanswered questions, manage the pages your Agent runs on, ship in-app guides. Tabs: Pages, Topics, Guides (plus Features and Configuration coming soon).
- **Insight** — browse workspaces and users, see adoption reports, monitor Agent sentiment. Tabs: Audience, Reports, Sentiment.
- **Settings** — Team, Integration & API Key, Security & Access, Billing, Privacy Management.

## What's available today

- **One-script SDK** — auto-captures clicks, pageviews, form submits, rage clicks, and DOM errors. No event tagging required. Works with React, Vue, Angular, plain HTML, or via Google Tag Manager. The only required code call is `HelpAlive.identify({ userId, tenantId })` after login.
- **AI Agent** — a context-aware assistant inside your product. Trained on your docs (URL crawl or file upload), aware of the user's current page, identity, and recent activity.
- **Topics inbox** — every question the Agent can't answer well, grouped into topics, ranked by volume and severity. Your team's working backlog of doc improvements.
- **Audience analytics** — per-user, per-workspace activity views with friction signals (rage clicks, errors).
- **In-app guide runtime** — multi-step walkthroughs with spotlight highlights on real elements in your product. Authoring UI is in active development.
- **Privacy & security** — two-layer PII redaction, per-customer data isolation, DSAR-ready async deletion jobs with audit trail, consent modes (auto / required) with runtime toggle.

## What's in active development

- Dashboard guide editor with record-mode authoring (click through your product to capture a guide).
- Agent-launched guides — the Agent identifies a goal during chat and runs the matching guide.
- Auto-suggested guides based on what your users actually try to do.
- Automatic feature discovery and feature-health scores.
- Proactive in-product recommendations driven by behavioral pattern analysis.

## Documentation index

### Get started
- [Introduction](/introduction) — what HelpAlive does and who it's for
- [Quickstart](/quickstart) — install in 5 minutes
- [Installation](/installation) — JavaScript script tag, Google Tag Manager, framework examples (React, Vue, Angular, plain HTML)
- [Use cases](/use-cases) — activation, churn, ticket deflection, upsell, visibility
- [Compare](/compare) — vs Pendo, Intercom Fin, WalkMe, FullStory, Userpilot, Appcues

### SDK
- [SDK overview](/sdk/overview) — what's auto-captured and the public API at a glance
- [Identify](/sdk/identify) — `identify({ userId, tenantId, ... })` — both required
- [Consent](/sdk/consent) — auto / required modes, runtime toggle

### Agent
- [Overview](/agent/overview) — what makes the Agent context-aware
- [Train](/agent/train) — load your knowledge: URL crawl, PDF / DOCX / Markdown upload, pasted text
- [Personalization](/agent/personalization) — tone, custom responses, guidance templates
- [Test](/agent/test) — preview draft answers, batch tests, simulate user context
- [Appearance](/agent/appearance) — colors, layout, welcome message, color scheme
- [Configuration](/agent/configuration) — visibility on/off, rate limits, danger zone (reset / remove)
- [Publishing](/agent/publishing) — atomic draft → preview → publish workflow

### Studio
- [Overview](/studio/overview) — Studio at a glance
- [Pages](/studio/pages) — where the Agent is active and how each page is performing
- [Topics](/studio/topics) — inbox of questions the Agent can't answer well
- [Guides](/studio/guides) — in-app walkthroughs, what works today, what's in development

### Insight
- [Overview](/insight/overview) — adoption analytics layer
- [Audience](/insight/audience) — workspace and per-user analytics with full activity timeline
- [Reports](/insight/reports) — date-range summaries (live: users today, total users, sentiment trend; coming soon: churn, onboarding, recommendation impact, coverage gaps)

### Settings
- [Account](/settings/account) — profile, password, sessions, notifications, dashboard light/dark mode, account deletion
- [Team](/settings/team) — invite teammates, manage roles, deactivate members
- [Integration & API key](/settings/integration) — install snippet, integration health, API key management
- [Security & Access](/settings/security) — email domain allowlist, session policy
- [Billing](/settings/billing) — seat usage, plan limits

### Privacy & security
- [Privacy overview](/privacy/overview) — guarantees, what's captured, what's not
- [Redaction](/privacy/redaction) — two-layer PII redaction details
- [Data deletion](/privacy/dsar) — async DSAR jobs with audit trail

### Plans
- [Plans](/plans/overview) — per-session + AI-usage pricing model, early-access program
- [FAQ](/plans/faq) — trial options, payment, what counts toward usage

## What customers get, in detail

- **In-app AI Agent** that answers user questions in context — knows the page, identity, plan, and recent activity. Reduces routine tickets and shortens time-to-answer for everything else.
- **Step-by-step in-app Guides** that walk users through real tasks. The Guide highlights the next element to click, advances when they actually click it, and ends with the task *completed* — not just read about. Common uses: first-day onboarding, surfacing under-used features, complex multi-step flows, error recovery, release announcements, plan-aware paths.
- **Topics inbox** — every question the Agent couldn't answer well becomes a row in your team's working backlog. Grouped by topic, ranked by volume × gap score, status-tracked (identified / addressed / ignored).
- **Audience analytics** — per-user and per-workspace activity timelines, friction signals (rage clicks, errors, abandoned forms) surfaced inline so support can investigate any user incident in one place.
- **Privacy controls** — two-layer PII redaction, async DSAR deletion jobs with audit trail, consent modes (auto / required), per-customer data isolation.

## Outcomes teams target

- Lower churn — catch silent struggle before it becomes cancellation
- Fewer support tickets — answer in-product, with full context
- Faster activation — guide users to first value without docs
- Less spend on user training
- A real, always-current map of how your product is used (no manual tagging)
- Clearer upsell triggers — see plan-restricted features being attempted

## Pricing model

HelpAlive is priced per end-user session captured by the SDK + AI usage on top, with a per-resolution charge layered in once outcome data exists. The model aligns customer outcomes with revenue — customers pay more only when more friction is being resolved.

In early access — pricing tiers are calibrated with design partners. Apply for early access at https://helpalive.com/#early-access for a pilot quote.