Unclassified

17 — Differentiation

Status: CANONICAL · Commercial OS
Owner: Product Marketing
Last verified: 2026-07-21

Purpose

The defensible differences — what only IntegrateWise can claim, why competitors structurally can't follow, and how each difference converts to a sales argument. This is the moat register; 09 Battle Cards applies it per competitor.

Target audience: Sales, marketing, founders, investors.

1. The Differentiation Stack (five moats, evolved from the investor canon)

Moat 1 — Shared-State Collaboration (the category moat)

What: Human and Twin work on the same workbench over the same live state. Not chat beside work — collaboration inside work. The Twin's interventions (NOTICE → PROPOSE) land in the semantic slot of the work itself.

Why others can't follow: Copilots are architecturally sidecars to a host app; assistants are architecturally conversation threads. Neither has a shared work-state model to collaborate on. Rebuilding around one means abandoning their core UX.

Sales argument: "Ask where their AI's output lives. If the answer is 'in the chat,' you're looking at a sidecar, not a colleague."

Moat 2 — TruthLayer: Approval-Gated Memory (the trust moat)

What: AI proposes memory; humans approve it; only then does it become organizational knowledge. Every other AI auto-writes memory — it guesses, and its guesses compound.

Why others can't follow: Auto-memory is their engagement engine; gating it degrades their core metric. For us, gating is the product philosophy.

Sales argument: "ChatGPT remembers what it wants. IntegrateWise remembers what you approve."

Moat 3 — The Adaptive Spine: Truth Outside the Apps (the data moat)

What: Canonical operational state independent of any SaaS vendor, with per-field provenance and confidence. Tools stay; truth is projected to workbenches by the Projection Engine.

Why others can't follow: Every incumbent's truth is its own silo — Salesforce's truth is Salesforce. A vendor whose business is the silo cannot sell truth-above-the-silo without cannibalizing itself.

Sales argument: "Your CRM's truth is the CRM. Your organization's truth should belong to your organization."

Moat 4 — Governance as Architecture (the enterprise moat)

What: Proposal → approval → execution → audit is the physics of the system. Fail-closed boundaries; identity-bound, tenant-scoped, policy-checked exchanges; the platform cannot be talked out of governance by a clever prompt.

Why others can't follow: Agent vendors race toward autonomy as their differentiator; retrofitting mandatory human gates reverses their pitch. Regulation is moving toward us, not them.

Sales argument: "Their roadmap removes the human. Ours arms the human."

Moat 5 — Compounding Continuity (the time moat)

What: Every approved fact, decision, and handover enriches the Spine. Value compounds with tenure; switching cost is accumulated institutional intelligence — earned, not imposed.

Why others can't follow: Session-based tools reset by design. Continuity requires the other four moats first.

Sales argument: the continuity math — "Month 1: one week of memory replaces three status meetings… Month 12: institutional memory no longer leaves with the people."

2. Differentiation Claims — Approved Wording

Claim	Status	Guard-rail
"Only workspace where humans and AI collaborate on shared, governed state"	Approved	Keep "governed" — ungoverned shared canvases exist
"The only AI memory that requires your approval"	Approved	Verified against market scan 08; re-verify quarterly
"AI never owns organizational truth"	Approved (doctrine)	Architecture claim — demo the write path
"Swap AI models without losing anything"	Approved	Provider-agnostic architecture; demo config switch
"Nothing executes without approval"	Approved	Always pair with audit-trail view
"70+ connectors"	Approved	Update count quarterly from live registry

Forbidden differentiation: price ("cheaper than Gainsight" invites race-to-bottom), model quality ("better AI" — we rent models like everyone), connector count as primary story (reads iPaaS).

3. The Differentiation Test Grid

Use in deal reviews. If we can't check ≥3, we're in the wrong deal or told the story wrong.

Prospect can restate why a Twin ≠ a copilot (shared bench vs. sidecar).
Prospect saw the approval gate and audit trail live.
Prospect understands truth lives in their Spine, exportable, provider-agnostic.
Prospect's security/IT acknowledged governance-by-architecture.
Prospect articulated what compounds by month 12.

KPIs: win rate in deals where ≥3 boxes checked vs. <3 (expect >2x gap); competitive-loss reasons tagged to a moat (any moat losing >30% of its matchups triggers repositioning review).