Activate · Vendor landscape
Agentic workflows are reshaping knowledge work. Alongside, a new architecture is forming to capture organisational knowledge that was previously invisible. The opportunity is to deliberately compose this across the customer corridor to produce organisational and customer value. Knowledge Strata is the lens. Three forces shape what comes next.
Convergence: vendors growing into adjacent knowledge layers. Corridor: agents crossing vendor boundaries inside the customer corridor. Structural + Commercial: thirty years of vendor-aligned teams and contracts quietly shaping future AI architectures.
The same four-layer reading applies to every category. The pattern is consistent within a category — Salesforce and Dynamics share the same shape — and different across categories. This is the cornerstone: vendor strategy read through Knowledge Strata.
| Category | L1 Transferable | L2 Domain | L3a Documented | L3b Tacit |
|---|---|---|---|---|
Engagement Salesforce · Dynamics 365 · HubSpot · Oracle CX · Pega | AI in product | Customer domain | Pipeline + service + marketing records | Sales + service patterns |
Record SAP · Oracle · Dynamics F&O · Workday · TechnologyOne · Infor | AI in product | Industry processes | Transactions, config, master data | Approval routing, exceptions |
End-user Compute Microsoft 365 · Google Workspace · Slack · Atlassian · Zoom | AI in product | Generic work context | Documents, email, meetings, chats | Collaboration patterns |
Workflow Ops ServiceNow · Atlassian · Pega · Power Platform · Nintex | AI in product | Process domain | Runbooks, tickets, escalations | Operational tacit, agent traces |
Hyperscaler AWS · Azure · GCP · OCI · IBM | Compute + model hosting | — | — | — |
Data Platform Snowflake · Databricks | — | Semantic + cross-app | Cross-app L3a archive | — |
Frontier AI Anthropic · OpenAI · Google DeepMind · Meta · Mistral | Model = L1 | Via fine-tune + retrieval | Skills, projects, custom GPTs | Memory, decision patterns |
The asymmetry is structural. Engagement and Record own L3a and are extending into L3b. End-user Compute owns the conversational L3a. Workflow Ops owns operational tacit (L3b). Frontier AI owns the model (L1) and is positioning to own decision-pattern L3b. No category captures all four layers. Which is why composing across them is the architecture decision.
Pillar 2 names four rungs of activation: access, embedding, workflow, trust. Where a vendor plays on that ladder is as important as which layer they capture.
The compression insight. Embedded enterprise vendors — Salesforce Agentforce, Microsoft 365 Copilot, SAP Joule — collapse the first three rungs of the activation ladder into a single deployment decision.
Compressed by the embedded vendor — one deployment decision
Still open — your strategic surface
When an organisation deploys Agentforce:
What’s left of the rungs is rung four: trust. Which is why verification architecture — human-in-the-loop checks, evaluation harnesses, provenance tracking, defensibility scaffolding — is the next strategic surface, not another vendor category. It sits downstream of the embedded deployment, not in competition with it.
What’s genuinely new. Cross-system integration isn’t new. Organisations have been moving data across the customer corridor for decades. What is new is intelligence acting across the integrated stack: agents that read records, reason against domain context, apply organisational rules, and decide as one continuous flow.
The combination of system of record and intelligence is the new shape. The Corridor section names how it composes.
Read the grid with this in mind. Embedded vendors compress activation inside their corridor. Two surfaces emerge next: Trust inside, and intelligence acting across what’s already integrated.
Vendors growing into adjacent layers
Each vendor is extending beyond where it started. Salesforce moving past the customer record, SAP past the transaction, frontier AI into how decisions get made. The boundaries between what they capture are dissolving.
Agents crossing the customer corridor
Most enterprises are organised along the customer corridor: separate teams for engagement, record, workflow, and end-user compute. Intelligence can now reason across all of them. The work between them, previously nobody’s territory, becomes strategic.
Vendor-aligned teams + contracts
Where your existing investment sits matters. Heritage vendors package AI into specific commercial SKUs (RISE with SAP, Copilot per-user, ServiceNow Pro Plus). Cloud-native providers offer it pay-as-you-go. Thirty years of vendor-aligned teams and contracts can quietly shape your AI architecture. Seeing them lets you choose deliberately.
The grid tells you what each vendor captures. Convergence, Corridor, and Structural + Commercial tell you how the landscape is moving and what is open to you. Together they enable you to knowingly make decisions and compose a deliberate stack.
Pillar 3 — Value Your Knowledge — picks up from here. With AI freeing capacity, where do you direct it? Four strategic moves are open: absorb adjacent work, expand the same service to more demand, productise a codified method, or specialise into your hardest tacit work. The architecture you compose enables the strategy.