Foldspace Product Page — Full Copy Draft

Track: AI-Native (primary frame) · ALG/PLG + Developer (secondary) Format: Main /product page — single long scroll Status: Draft v1 — for Mickey review Date: 2026-06-01

DESIGN NOTES FOR IMPLEMENTER

• Page reads top to bottom. Each zone escalates from category → mechanism → proof → ICP conversion.
• Do not reorder zones. The trust-building sequence matters.
• Code blocks should be syntax-highlighted (recommend Shiki).
• Architecture diagram references a visual asset to be produced from the ASCII in core/overview.md.
• All stats are verbatim from core/overview.md — do not modify without owner sign-off.
• Avoid-list compliance: no "AI-powered," "AI Copilot," "agentic transformation," "production-ready," "personalization at scale," "in-app tours."
• Modules marked [EARLY ACCESS] must be messaged forward-looking — Conversational Analytics and Trust Lab.

ZONE 1 — HERO

NAV

Product · Customers · Docs · Pricing · Blog CTA buttons: Book a demo [primary] · Read the docs [secondary]

EYEBROW

AI-Native Software

HEADLINE

The agentic interface layer. Users say what they want. The agent does it. Every interaction makes your product smarter.

SUB-HEADLINE

Foldspace sits at Layer 2 — between your users and your stack. Embed once. Ship production-grade AI-native experiences in hours, not quarters. No re-platform. No ML team required.

HERO CTAs

• [Book a demo] — primary button
• [Read the docs] — secondary link

HERO VISUAL

[Design note: 30–45 second looping product demo video. Show a real UI — user types a natural language prompt, the Product Agent executes a multi-step workflow, result appears in the live product. End frame: a Conversational Analytics dashboard showing intent signals. No voiceover needed — let the product speak.]

SOCIAL PROOF BAR (below fold line)

Logos: MixMax · Optibus · AskCody · Vidmob · Qonto · SugarCRM · Anvil · EverCommerce Caption: "Trusted by AI-native teams from Series B to public company"

ZONE 2 — CATEGORY STATEMENT

Tone: confident, declarative. No bullet points. Reads like a manifesto paragraph.

HEADLINE

AI-Native software doesn't have menus. It has agents.

BODY

The products that win the next decade won't ask users to learn a UI. They'll ask users what they want — and deliver it. That's not a chatbot bolted onto a dashboard. That's a fundamentally different interface layer: one that captures intent, executes inside the live product, and gets smarter with every conversation.

That layer is Foldspace.

We call it the agentic interface — the surface where user intent becomes product action, where every prompt is a demand signal, and where activation stops being a funnel and starts being a conversation.

ZONE 3 — THE THREE PROBLEMS

Three cards. Universal pain — before any ICP split. Short and sharp.

SECTION LABEL

Why AI-Native teams choose Foldspace

CARD 1: THE CLICK TAX

Your product is powerful. Reaching that power costs too much.

The average SaaS workflow takes 30 clicks. Your users came for the outcome, not the journey. Every empty state, settings menu, and tooltip is activation risk. The Click Tax is real — and it's killing conversion.

CARD 2: THE ACTIVATION GAP

PLG removed the salesperson. Nobody removed the complexity.

Self-serve works for simple products. For everything else, you pushed the onboarding burden onto the user, replaced the salesperson with a 7-step product tour (16% completion rate), and hoped. It didn't work. The activation gap isn't a UX problem — it's an interface problem.

CARD 3: THE BUILD TRAP

18 months of infra. Still demoing the same thing.

Every AI-Native team faces the same fork: build a production-grade agentic layer yourself, or bolt on a chatbot and call it done. Building takes quarters. Bolting on burns trust. There's a third path: an interface layer purpose-built for production that ships in hours.

ZONE 4 — HOW IT WORKS (Architecture + Code)

This zone is critical for technical trust. CTO and AI Engineers read this section carefully.

SECTION LABEL

The Platform

SECTION HEADLINE

Every AI-Native product converges on three layers. We own the top one.

ARCHITECTURE DIAGRAM

[Design note: Render this as a clean, branded visual. Use the three-layer stack from core/overview.md. Important labels below.]

┌─────────────────────────────────────────────────────────────┐
│  LAYER 2 — AGENTIC INTERFACE              [ Foldspace ]     │
│  Product Agent · Conversational Analytics · Trust Lab       │
│  Context Engine · Optimizer · Inference Gateway             │
└──────────────────────┬──────────────────────────────────────┘
           SDK ↓              ↕ A2A + MCP
┌──────────────────────┴──────────────────────────────────────┐
│  LAYER 1 — ORCHESTRATION                  [ Your IP ]       │
│  Your agent loops · Domain models · Proprietary logic       │
└──────────────────────┬──────────────────────────────────────┘
                       │
┌──────────────────────┴──────────────────────────────────────┐
│  LAYER 0 — DETERMINISTIC BACKEND          [ Unchanged ]     │
│  Your APIs · Permissions · Business logic · Audit trails    │
└─────────────────────────────────────────────────────────────┘

DIAGRAM CALLOUT

Layer 1 is your moat. We sit above it — not inside it. Optibus built their own orchestration layer. Then they bought Foldspace for the interface. Your IP stays yours.

SDK EMBED — CODE BLOCK

Label: "Live in your product in under an hour"

import { FoldspaceAgent } from '@foldspace/sdk';

FoldspaceAgent.init({
  token: 'fs_live_••••••••',
  actions: ['navigate', 'filter', 'export', 'create'],
  context: {
    userId: user.id,
    role: user.role,
    plan: user.plan,
    screenState: getCurrentScreenContext()
  }
});

Caption: "Single SDK. Connects to your existing APIs and permissions — no new backend required."

EMBED PROOF POINT

Hours, not quarters. The average Foldspace customer ships their first production agentic experience in under 8 hours of engineering time.

ZONE 5 — THE THREE PILLARS

Each pillar gets its own full section. AI-Native framing throughout. Ordered: Agentic Interface → Analytics → Trust Lab.

PILLAR 1: AGENTIC INTERFACE

EYEBROW

Product Agent

HEADLINE

The interface AI-Native software was always supposed to have.

BODY

The Product Agent is the surface where your users live. They type or speak what they want — "set up my Q3 pipeline," "pull last month's report for enterprise accounts," "onboard this new user with the same config as Sarah" — and the agent does it, inside your live product, bound by the same APIs and permissions as your UI.

This is not a chatbot in the corner. This is the primary interface.

Every response the Product Agent generates is assembled from a live, four-dimensional picture of the user and the product state:

• User memory — preferences, prior intents, prior outcomes, persistent across sessions
• Behavioral signals — what this user, and users like them, actually do in the product
• Knowledge — grounded against your help docs, policies, and product content via RAG
• Live screen state — what the user is looking at right now, their role, their permissions, their in-flight workflow

[Design note: show a 4-quadrant visual of these context inputs feeding into a central "Agent Response" — like a signal map]

This context engine is the moat. It's what makes the Product Agent accurate, low-latency, and cost-sustainable at scale — and it's what separates Foldspace from a prompt wrapper.

MODALITIES

The agent meets users where they are:

STAT CALLOUT

Day-1 retention: 57% → 90% MixMax deployed the Foldspace Product Agent. Day-1 retention moved from 57% to 90%. Day-7 retention moved from 40% to 60%. Time to first outcome: under 60 seconds.

SECONDARY STATS (smaller, beneath)

• +344% daily AI-driven actions (AskCody)
• Feature activation 8% → 16% (Qonto)
• 3–5× trial-to-paid lift across deployments

CTA

[See the Product Agent in action →]

PILLAR 2: CONVERSATIONAL ANALYTICS

EYEBROW

Conversational Analytics · Early Access

HEADLINE

AI-Native analytics captures intent, not clicks.

BODY

Your existing analytics tells you what users clicked. Conversational Analytics tells you what users wanted — including the things they tried to do and couldn't.

Every prompt your Product Agent receives is a real demand signal. Every unresolved prompt is a feature you haven't built yet. Every cost-per-action trend is a model decision waiting to be made. Conversational Analytics surfaces all of it in one dashboard, in real time.

FEATURE LIST

Intent signals See what users are actually trying to accomplish — not the buttons they pressed to get there. Cluster by intent type, user segment, product area, and time period.

Unresolved prompt tracking The prompts your agent couldn't complete are your highest-priority roadmap input. Surface them ranked by frequency and user segment.

Cost per action Every agentic interaction has a token cost. Track it by action type, model, and user cohort. The CFO line item becomes a product decision.

Model performance Response quality scores, latency distributions, and eval pass rates — live, not in a spreadsheet.

Sentiment and friction signals Know before users churn. Conversational signals detect frustration, confusion, and drop-off earlier than any click-based heuristic.

CONTRAST CALLOUT (visual — two columns)

QUOTE PLACEHOLDER

[Mickey to supply: a customer quote on how Conversational Analytics changed their roadmap or churn signal]

CTA

[Join the early access program →]

PILLAR 3: TRUST LAB

EYEBROW

Trust Lab · Early Access

HEADLINE

Ship agents the way SREs ship infrastructure.

BODY

AI-Native teams don't ship agents the way they used to ship features — with a staging environment and a prayer. They ship them the way great engineering orgs ship infrastructure: with golden sets, shadow mode, canary rollouts, and full-loop evals that see both the response and the resulting API call.

Trust Lab is that system — built into Foldspace, not bolted on.

HOW IT WORKS (step flow — visual)

STEP 1: SHADOW MODE
Run your new agent configuration alongside production.
Compare outputs without exposing users to changes.

         ↓

STEP 2: GOLDEN SET VALIDATION
Test against real production conversations captured
automatically from live sessions.

         ↓

STEP 3: CANARY ROLLOUT
Release to a defined user cohort. Monitor eval pass
rate, cost-per-action, and latency in real time.

         ↓

STEP 4: FULL PRODUCTION
Ship with confidence. Trust Lab keeps running.
Every conversation is a new eval candidate.

FEATURE CALLOUTS

Full-loop evals Evaluate both the agent response and the resulting API call or state change. A response that looks right but triggers the wrong action fails. Trust Lab catches both.

Model switch validation Switching from GPT-4o to Claude Sonnet 4? Run your golden set against both before you touch production. Trust Lab shows cost, latency, and quality delta side by side.

Prompt and context change testing Every change to your system prompt or context configuration runs through your golden set before it ships. No silent regressions.

A2A agent testing Trust Lab tests the Foldspace agent — which can call your own agents via A2A — so you can validate complex, multi-step agentic chains end-to-end.

SECURITY + COMPLIANCE BLOCK

[Design note: render as a clean grid of badges/chips — not a wall of text]

Caption: "Foldspace-managed today. LangChain-compatible on the roadmap."

TECHNICAL CALLOUT (for CTOs)

The eval story that holds up under scrutiny. Trust Lab's golden sets are built from real production conversations — not synthetic benchmarks. You own the golden set. You define the pass criteria. The methodology is yours to audit.

CTA

[Join the early access program →]

ZONE 6 — PLATFORM CAPABILITIES

Shorter section — two platform-level features that underpin all three pillars.

SECTION HEADLINE

The platform underneath the platform.

OPTIMIZER

Self-learning. Self-tuning. Always on.

The Optimizer runs continuously across every conversation — adjusting routing, context selection, and model choices to improve accuracy, reduce latency, and cut token cost. Paired with the inference gateway, it learns what to route where and gets better with every interaction. This is the data flywheel in practice.

Metric: Customers see 20–35% token cost reduction within the first 90 days.

A2A + MCP

Your existing AI investments stay yours.

Foldspace connects to your own autonomous agent loops, domain models, and voice agents via Agent-to-Agent protocol. It's fully MCP-compatible — so your MCP servers, your tool ecosystem, and any ChatGPT App integrations you've already shipped all work natively. Foldspace integrates with your AI stack. It does not replace it.

ZONE 7 — ICP SECTIONS

Sequential scroll — not tabs. Each section feels written for one buyer, without excluding others.

FOR PRODUCT & GROWTH TEAMS

EYEBROW

For Product & Growth

HEADLINE

Stop building activation funnels. Ship activation.

BODY

PLG gave your users a sign-up form and a 7-step tour. The tour completes at 16%. The funnel leaks. The board wants to know why the AI initiative isn't moving the metrics yet.

Agent-Led Growth is the answer — and it's not a rebrand. It's a concrete change to how your product handles the moment between "user signed up" and "user got value." The Product Agent collapses that moment from days to under 60 seconds.

You don't need an ML team. You don't need to rebuild your stack. Agent Studio is the no-code surface where your team defines actions, schemas, and agentic experiences — and the Remote Plugin means a PM can stand up a live agentic demo without touching engineering.

KEY FEATURES FOR THIS BUYER

• Product Agent — executes complete workflows on behalf of the user
• Agent Studio — no-code action and schema builder for PMs and growth teams
• Remote Plugin — deploy agentic experiences without engineering
• Conversational Analytics — unresolved prompts write your roadmap; intent signals replace click funnels

PROOF POINT

MixMax — Day-1 retention 57% → 90% Day-7 retention 40% → 60%. Time to first outcome: under 60 seconds. "We replaced a 15-minute onboarding flow with a single prompt." — [Mickey to confirm quote + attribution]

CTA

[See how it works for Product teams →] (links to /for/product-growth)

FOR ENGINEERING & AI TEAMS

EYEBROW

For CTOs & AI Engineering

HEADLINE

Your Layer 1 stays yours. We own Layer 2.

BODY

You've done the hard work. The orchestration layer is built. The evals are running. You have agents in production — or close enough. The question you're now getting from the CEO is: "Why does the demo still look like a demo?"

The missing piece isn't another framework or another model. It's the interface layer — the surface your users actually see, the analytics that turns their prompts into roadmap signal, and the eval infrastructure that lets you ship changes to the agentic stack without a prayer.

That's Layer 2. That's Foldspace.

ADDRESSING THE REAL OBJECTIONS

[Design note: render as an accordion or three expandable callout cards]

"This is a thin wrapper." The Product Agent's context engine fuses user memory, behavioral signals, RAG-grounded knowledge, and live screen state for every single interaction. The system prompt is never exposed. The context selection — what knowledge and which actions are relevant to this specific user at this specific moment — is the product. This is not a prompt wrapper.

"We'll just use Anthropic Managed Agents." Managed Agents run your brain. Foldspace is the face your users see and the data flywheel they generate. The two are complementary — Optibus runs their own orchestration and uses Foldspace for the interface layer above it. Most AI-Native teams end up needing both.

"MCP makes this obsolete." Foldspace ships MCP-compatible natively. Your MCP servers, your tool ecosystem, any ChatGPT App integrations you've already shipped — they all connect. We are not competing with MCP. We are the interface layer that sits above it.

KEY FEATURES FOR THIS BUYER

• Three-layer architecture — clean separation between your IP (Layer 1) and the interface (Layer 2)
• Trust Lab — golden sets, shadow mode, canary rollouts, full-loop evals
• A2A + MCP — connects to your existing agent loops and tool ecosystem
• Optimizer — routing, cost reduction, latency tuning — automated
• Enterprise security — ISO 27001, SOC 2 Type II, GDPR, JWT/HMAC, full audit trails

PROOF POINT

Optibus — Built their own Layer 1. Bought Foldspace for Layer 2. "We had the orchestration loop. We didn't want to build the interface on top of it. Foldspace shipped it in days." — [Mickey to confirm quote + attribution]

CTA

[Talk to our engineering team →] (links to /for/engineering)

FOR DEVELOPERS

EYEBROW

For Developers

HEADLINE

Ship a production AI layer this sprint. Own the code.

BODY

Two plugins. One for building inside your codebase. One for building without it.

The Developer Plugin (in-codebase) works inside Claude Code and Cursor — you build the agentic experience against your SDK and APIs, in the environment you already use. The Remote Plugin (no-codebase) works through Chrome MCP with the Foldspace Chrome extension — the extension injects the SDK, the SDK pulls actions from a remote repo. A PM or growth engineer can stand up live agentic experiences and demos with no backend changes.

CODE BLOCK

// Developer Plugin — define an action in your codebase
export const createReportAction = defineAction({
  name: 'create_report',
  description: 'Generate a report for a given date range and account',
  parameters: z.object({
    dateRange: z.object({ from: z.string(), to: z.string() }),
    accountId: z.string(),
    format: z.enum(['pdf', 'csv', 'dashboard'])
  }),
  execute: async ({ dateRange, accountId, format }) => {
    return await reports.generate({ dateRange, accountId, format });
  }
});

Caption: "Your action. Your API. Your permissions. Foldspace handles the intent layer above it."

KEY FEATURES FOR THIS BUYER

• Developer Plugin — build in Claude Code or Cursor, against your own SDK
• Remote Plugin — no-codebase deployment via Chrome MCP
• Agent Studio — hand off action authoring to PMs once the schema is defined
• Full TypeScript SDK with type inference
• MCP-native — your existing MCP servers connect out of the box

CTA

[Read the docs →] · [View on GitHub →]

ZONE 8 — SOCIAL PROOF

Full-width section. Outcome-anchored — not just logos.

SECTION HEADLINE

Results from teams who shipped.

CUSTOMER GRID (3 cards minimum)

MixMax "We replaced a 15-minute onboarding flow with a single prompt." → Day-1 retention: 57% → 90% → Time to first outcome: under 60 seconds

AskCody [Mickey to supply quote] → Daily AI-driven actions: +344%

Qonto [Mickey to supply quote] → Feature activation: 8% → 16%

Optibus [Mickey to supply quote] → Built own orchestration layer, deployed Foldspace interface in days

LOGO WALL

MixMax · Optibus · AskCody · Vidmob · Qonto · SugarCRM · Anvil · EverCommerce [Mickey to confirm which logos are cleared for public display]

ZONE 9 — ENTERPRISE TRUST

Gear-shift zone — tone becomes more formal. For the economic buyer and security reviewer.

SECTION HEADLINE

Enterprise-grade from day one. Not an add-on.

COMPLIANCE GRID

[Design note: large badge/chip grid — visually prominent]

ENTERPRISE FEATURE LIST

• SSO — SAML 2.0 and OIDC
• Dedicated support — named CSM and engineering contact for enterprise accounts
• SLA — 99.9% uptime commitment
• Custom data retention — configurable per your compliance requirements
• On-prem / VPC deployment — available on request

CTA

[Talk to our enterprise team →]

ZONE 10 — FINAL CTA

Clean, simple close. Two paths — one for each primary buyer.

HEADLINE

Ready to build AI-Native?

SUB-HEADLINE

See Foldspace running inside a product like yours — or go straight to the docs.

CTA PAIR

• [Book a demo] — primary button (VP Product, CTO)
• [Start with the docs] — secondary link (IC Dev, AI Engineer)

FOOTNOTE TRUST SIGNALS

ISO 27001 · SOC 2 Type II · GDPR · No re-platform required · Ships in hours

METADATA & SEO

Page title: Foldspace — The Agentic Interface Layer for AI-Native Software Meta description: Foldspace is the agentic interface layer for AI-Native software. Users say what they want; the Product Agent does it; every interaction makes your product smarter. Deploy in hours — not quarters.

Primary keywords (own these):

• agentic interface layer
• AI-Native software platform
• Agent-Led Growth
• conversational analytics for SaaS
• agentic onboarding

Anchor links for paid campaigns:

• /product#for-product-growth → ICP Section A
• /product#for-engineering → ICP Section B
• /product#for-developers → ICP Section C
• /product#trust-lab → Pillar 3
• /product#analytics → Pillar 2

REVIEW CHECKLIST

Before this page ships, Mickey to confirm:

• MixMax quote and attribution cleared for public use
• AskCody quote sourced and cleared
• Qonto quote sourced and cleared
• Optibus quote sourced and cleared
• Logo wall — which of the 8 logos are cleared for public display
• Conversational Analytics and Trust Lab — confirm "Early Access" is the right label and CTA
• "20–35% token cost reduction" stat — customer-attested or internal benchmark? Needs sourcing before it ships publicly
• "Under 8 hours engineering time" stat — same question
• GitHub link for Developer section — confirm public repo exists or remove
• On-prem / VPC deployment — confirm this is offered or remove
• SSO — confirm SAML/OIDC is available or adjust
• ALG definition — confirm positioning vs. Lifesight / Insight Partners (see positioning/icp-a-algplg.md §5b)
• Avoid-list check — run final copy through icp-a-algplg.md §4 and icp-b-ainative.md §4