Prism

Educate.        Engage.        Convert.

PiSrc Prism is an AI conversation platform for enterprise websites that earns engagement through knowledge rather than interception. Most AI chatbots try to capture a name and book a meeting as fast as possible, skipping the education and nurture steps that build genuine buyer interest. Prism takes a different approach. It ingests your entire knowledge base and answers the visitor's question first, with depth and citations, showing them the full spectrum of what you offer so by the time a lead reaches sales, the lead is one worth having.

Leads Worth Having

The Nurture is the Conversion

As visitors engage with your content, customer-defined agentic playbooks guide the conversation from first question to first call, collecting qualifying information naturally through dialogue rather than forms. By the time a lead reaches your sales team, it arrives with a conversation transcript, collected context, a computed lead score, and a documented trail of the visitor's journey through your content. Prism shows your visitors the full spectrum of what you offer and shows your team the full picture of who they are. The result is better-informed buyers and leads worth taking. We offer an out-of-the-box solution optimized for AEM and ElasticSearch, or customizable enterprise solutions that support alternate architectures and business requirements.
How it works

From Question to Conversation to Conversion

1) Map. Prism ingests your sitemaps, pages, KB articles, FAQs, product catalogs, legacy datasources, and external DB/API feeds into a single unified index. Your full spectrum of offerings becomes accessible to every visitor interaction.
2. Educate. When someone asks a question, Prism uses intent-based routing to assemble an on-brand, citation-backed answer from authoritative sources. Every reply includes deep links to the exact source and follows your tone and guardrail rules. The visitor gets the answer they came for, not a deflection to a sales calendar.
3) Engage. As the conversation develops, Prism activates playbooks that match visitor intent. It collects qualifying information naturally and builds a visitor profile through dialogue and background research. The visitor is learning about your offerings while Prism is learning about them.
4) Convert. When behavior signals readiness, Prism offers the right next step: a scheduled demo, a live chat with sales, or a follow-up message. The handoff includes a conversation transcript, collected data, lead score, playbook trail, and detected context like company profile and technology stack. Sales receives an intelligence package, not just contact information.
5) Improve. Analytics capture clicks, feedback, and usage patterns. These signals refine routing, boosts, and templates over time without changing your underlying systems.
Your Conversation Strategies

Agentic Playbooks

A playbook is a conversation strategy written in plain language that tells the AI how to behave when a visitor discusses a specific topic. No code required. Your marketing team writes playbooks in Markdown, publishes them, and iterates, the same authoring workflow they already know.
Prism maintains a playbook stack. When a visitor arrives, a welcome playbook activates. As the conversation shifts, Prism detects the change and pushes the right playbook onto the stack. When that topic resolves, the previous playbook resumes. A visitor mentions pricing and Prism activates a solution-consultant playbook. They ask to speak with someone and Prism switches to a handoff playbook. When escalation completes, the conversation returns to where it left off. The visitor experiences a fluid conversation while your strategies execute behind the scenes.
Playbooks also activate automatically based on conditions you define: a visitor arriving from a pricing page, a lead score crossing a threshold, a specific technology detected, or a certain number of conversation turns. Segment visitors by intent. Qualify leads by collecting budget, authority, need, and timeline through dialogue rather than forms. Trigger promotions when a visitor mentions a competitor. Control escalation paths. A/B test strategies by swapping playbook content. Each playbook carries tone and style guidelines so your brand voice stays consistent. The education is the qualification.
Two Modes, One KNowledge Base

AI Overview

Prism meets visitors in the two places intent is highest: the search bar and the conversation.

Search AI: When a visitor searches your site, Prism executes an intent analyzer to identify potential questions that might be answered with a concise, cited summary drawn from your knowledge base. This is displayed above or alongside traditional results. The visitor gets an immediate answer with links to source pages, reducing friction and surfacing content they might never have found through keyword matching alone. For many visitors, this is the first interaction with your AI layer, and it earns trust by being useful rather than intrusive. For users looking for traditional keyword search results, the intent analyzer will suppress AI generated output.

Conversational AI: When a visitor opens the chat, Prism shifts into dialogue mode. It answers questions with depth and citation, activates playbooks to collect qualifying information naturally, and hands off to sales with a full intelligence package. The conversation builds on the same content index that powers the search overview, so there is no gap between what a visitor reads in search and what the AI discusses in chat.

Both modes feed the same analytics pipeline. Search queries reveal what visitors are looking for. Conversations reveal why. Together they give your team a complete picture of visitor intent across the site.
Capabilities

Features at a Glance

AI Recommendations

Context-aware next steps

Prism recommends the next best action from user context, mixing rules, embeddings, and behavior to surface on-brand suggestions with deep links and citations. Permissions, locale, and guardrails are enforced. Data experiments and analytics keep rankings improving.

Background Intelligence

Know who you are talking to

When a visitor arrives on your site, Prism runs background research automatically. IP lookup resolves company identity. Website analysis detects technology stack, industry, and company profile. These insights fill data slots silently, enabling the AI to tailor both the content it surfaces and the engagement strategy it follows.

Audio, Speech, Multi-Languages

Every language, every market

Text-to-Speech with per-locale voices. Speech-to-Text with push-to-talk or auto-detect. Pronunciation hints for product names and acronyms. Locale-aware search and routing adapt to regional terminology and dialects while preserving brand voice. Built-in accessibility includes captions, transcripts, and keyboard controls aligned with WCAG guidance.

Escalation and Scheduling

Connecting when it counts

Prism presents available time slots as interactive calendar components within the chat, and because the visitor has already been educated through conversation, the meeting starts with shared context rather than a cold introduction. When a visitor needs a person now, Prism checks agent availability and routes to live text chat. Integration with Salesforce Lightning Scheduler and Google Calendar.

AI Conversations

Guided nurture

Deploy Prism as a conversational chatbot on your website. Visitors ask questions, explore your products and solutions, qualify themselves through natural dialogue, and connect with your team through a guided experience with product cards, carousels, calendar pickers, and action buttons. Search overview and chatbot draw from the same unified content index, so answers are consistent.

Analytics and Optimization

Close the loop

Track intents, click-through rates, and trending content by product, locale, or audience. Drill into guardrail triggers and escalations to refine policy. Test prompts, templates, and ranking weights. Capture feedback to surface content gaps and prioritize fixes.

Lead Qualifcation and CRM

Conversation to pipeline

Collect visitor information through natural dialogue and syncs to your CRM. Lead scoring updates in real time using configurable rules tied to signals like company size, budget indicators, and engagement depth. Every lead includes a sales brief of the conversation summary, user interests, lead score, and playbook trail. Integration with Salesforce is included out of the box.

Privacy and Control

Secure by design

PII masking and deny-lists at the edge before any storage. Permission-aware tracking with privacy compliance in mind. Configurable guardrails for topics, tone, and redaction. Audit logs showing exactly which prompts and sources were used. Dashboards for tokens, cost, and latency.

AI Overview

Meet vistors at the moment of intent

Search AI Overview (Site Search Integration): Prism generates cited AI summaries alongside your existing search results, meeting visitors at the moment of highest intent. See the Search AI section above for details.

Platforms

AEM & Elastic Optimized and Beyond*

Prism connects to AEM and ElasticSearch for fast, precise, low-latency results. It indexes pages, fragments, assets, and metadata with language-aware analyzers, blends keyword (BM25) with vector/semantic search, and returns stable IDs with deep links for trustworthy citations. Filters, facets, and per-section boosts align results to your site structure.

Core Indexing

  • AEM-aware indexing: pages, fragments, assets, metadata; custom analyzers (language, stemming, synonyms).
  • Hybrid retrieval: keyword + vector/semantic blending with pinning/demotion when needed.
  • Structure-aligned ranking: filters, facets, and section-level boosts that mirror how your site is organized.
  • Trustworthy citations: every result returns a stable ID and deep link.

Realtime Delivery

  • Real-time mode: make updates immediately available for high-priority pages and campaigns.
  • Priority queues: fast-track critical sections while the full pipeline catches up.

Crawler & Data Pipelines

  • Elastic-native crawling: works out of the box; extend with custom parsers, sitemaps, and recrawl rules.
  • Scheduled data syncs: continuously validate freshness and keep AEM, databases, shared drives, and APIs aligned.
  • Data-aware profiling: analyze structure, metadata, and semantic signals to optimize context engineering.

Flexible by Design

  • One brain for search + chat: a single source of truth powering both experiences.
  • Stack-agnostic: connect additional sources and backends as your architecture evolves.
Why Prism?

Substance Over Spectacle

Some AI platforms put a blinking, lip-syncing avatar on your website and call it engagement. The novelty lasts about ten seconds before the uncanny valley sets in and the visitor realizes they are watching a video presentation that cannot quite match its own speech. Behind the spectacle, the conversation is shallow: every exchange steers toward "make an appointment" or "chat with sales now," regardless of whether the visitor is ready or whether the meeting will be productive. Sales time is expensive. It should be earned through substance, not spent on curiosity clicks. Prism does not perform. It informs. Visitors engage with your actual knowledge base, explore your products with depth and citation, and arrive at a conversation with sales only when the interaction has produced genuine mutual understanding. No avatars, no theater, no wasted meetings. Just informed buyers and the leads your sales team will actually want to take.

Plans & Pricing

Enterprise

Custom Solution For Scale & Control
Custom
  • All Quickstart features + deeper tuning
  • Full ElasticSearch + hybrid search integration
  • Structured ingestion: APIs, PDFs, databases, etc.
  • Fine-tuning, guardrails, and policy controls
  • Dedicated deployment, optimization, and support
  • Stack-agnostic: AEM, headless CMS, custom sites
Get in Touch