SHAREPLANE AGENT CONTEXT
Schema: shareplane-context/2.0
Schema URL: https://next.shareplane.malott.ai/schemas/agent-package.schema.json
Projection: Generated public-safe plain text. Not canonical Markdown.
Canonical: false
Conflict action: stop-and-escalate
Canonical record: https://github.com/pinklon/pinklon-shareplane-next/tree/33d227f6000da2491f19209edde916d8846f287f/content/artifacts/do-not-outsource-the-brain/artifact.json
Canonical record SHA-256: e4d0fdb32c839b79db1881edad9bf40c65c192cc967586bda2642818b1626af4
Source content SHA-256: 63133cc82d3e6155f11bf68072e2d94eda6120bf727f80ec9a95d34221b43f70
Generation receipt: https://next.shareplane.malott.ai/build-receipt.json
Content role: artifact-content
Content trust: untrusted-data
Instructions allowed: false
Operational authority: none

IDENTITY
Artifact ID: artifact:do-not-outsource-the-brain
Slug: do-not-outsource-the-brain
Canonical URL: https://next.shareplane.malott.ai/artifacts/do-not-outsource-the-brain/
Title: Do Not Outsource the Brain
Abstract: Article-first SharePlane op-ed arguing that AI makes high-context internal judgment more valuable, so enterprises should outsource capacity rather than the cognitive center of the business.
Author: Tony Malott
Author URL: https://malott.ai
Published: 2026-07-07
Updated: 2026-07-07
Format: teaching-artifact-worked-example
Privacy: public-safe-aggregated
Topics:
- ai-governance
- strategy
- governed-ai
- governance
- static-publishing
- teaching-artifacts
Audience:
- engineering-leadership
- enterprise-architecture
- ai-governance
- technical-review
- artifact-authors
- governance-review

PROVENANCE
Posture: public-source-supported-owner-thesis
Private sources used: true
Private sources published: false
Public-safe boundary:
Uses the locked self-contained HTML article source and its reader-facing public source index. No raw private source packets, transcripts, screenshots, employer materials, internal documents, local paths, controlled information, backend service, runtime AI, analytics, package tooling, or dependency additions are published.

CLAIMS
None declared.

PUBLIC SOURCES
Source: source:do-not-outsource-the-brain:pwc
Title: PwC, 2026 Global AI Jobs Barometer
Type: public-source
Role: evidence-support
Description: Supports the labor-market argument: AI is creating a two-track labor market, professionalised roles are growing faster, AI-exposed jobs are changing faster, and junior roles increasingly require senior skills.
Locator: https://www.pwc.com/gx/en/services/ai/ai-jobs-barometer.html

Source: source:do-not-outsource-the-brain:bcg
Title: BCG, AI at Work 2026: Why Strategy Matters More Than Tools
Type: public-source
Role: evidence-support
Description: Supports the “tools are not the moat” argument. AI use is widespread, but organizations struggle to convert saved time into business value without strategy, guidance, and workflow redesign.
Locator: https://www.bcg.com/publications/2026/ai-at-work-why-strategy-matters-more-than-tools

Source: source:do-not-outsource-the-brain:deloitte-state
Title: Deloitte, State of AI in the Enterprise 2026
Type: public-source
Role: evidence-support
Description: Supports the governance, scale, workflow redesign, human oversight, and worker-skill argument. Deloitte explicitly ties scaling to governance, data, regulation, workforce readiness, and redesigned work.
Locator: https://www.deloitte.com/de/de/issues/generative-ai/state-of-ai-in-enterprise.html

Source: source:do-not-outsource-the-brain:deloitte-trends
Title: Deloitte, 2026 Global Human Capital Trends
Type: public-source
Role: evidence-support
Description: Supports the orchestration thesis: competitive advantage shifts from static scale to the ability to sense, assemble, and reallocate people, skills, data, and AI in real time.
Locator: https://www.deloitte.com/nl/en/services/consulting/research/human-capital-trends.html

Source: source:do-not-outsource-the-brain:deloitte-outsource
Title: Deloitte, Outsourcing for Strategic Advantage
Type: public-source
Role: evidence-support
Description: Supports the balanced sourcing argument: internal capability and external providers both matter, but they must be orchestrated intentionally around outcomes.
Locator: https://www.deloitte.com/global/en/services/consulting/perspectives/the-power-multidimensional-workforce-outsourcing-strategic-advantage.html

Source: source:do-not-outsource-the-brain:mckinsey
Title: McKinsey, The State of AI in 2025
Type: public-source
Role: evidence-support
Description: Supports the adoption-versus-impact argument: 88% report regular AI use in at least one business function, but only about one-third are scaling, and high performers redesign workflows and apply management practices around validation, KPIs, talent, operating model, technology, data, and adoption.
Locator: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai

Source: source:do-not-outsource-the-brain:gartner
Title: Gartner, 2025 Agentic AI cancellation forecast
Type: public-source
Role: evidence-support
Description: Supports the AI theater and agent-washing argument: Gartner predicts more than 40% of agentic AI projects will be canceled by the end of 2027 and warns that vendors are rebranding assistants, RPA, and chatbots as agentic AI without substantial capability.
Locator: https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-end-of-2027

Source: source:do-not-outsource-the-brain:wef
Title: World Economic Forum, Future of Jobs Report 2025
Type: public-source
Role: evidence-support
Description: Supports the skills-gap and workforce-transformation argument: AI and information processing are expected to transform business for 86% of employers, and skill gaps are the largest barrier to transformation.
Locator: https://www.weforum.org/publications/the-future-of-jobs-report-2025/digest/

Source: source:do-not-outsource-the-brain:mit
Title: MIT NANDA, The GenAI Divide: State of AI in Business 2025
Type: public-source
Role: evidence-support
Description: Use as caveated directional evidence only. It supports the pilot-to-production chasm and workflow-integration argument, but its own report notes that some figures are directionally accurate based on interviews rather than official company reporting.
Locator: https://mlq.ai/media/quarterly_decks/v0.1_State_of_AI_in_Business_2025_Report.pdf

RELATIONSHIPS
Relationship: migratedFrom
Target: repository:pinklon/shareplane
Posture: declared
Evidence: site/pages/do-not-outsource-the-brain/index.html

Relationship: relatedTo
Target: artifact:demo-debt
Label: Demo Debt
Display posture: Operationalization companion
Description: Do Not Outsource the Brain explains why high-context internal judgment must remain owned; Demo Debt shows what happens when organizations mistake impressive output for supportable capability.
Posture: declared-migrated-target
Evidence: site/registry.json relatedArtifacts in source order

Relationship: supports
Target: legacy-pattern:shareplane-premium-self-contained-editorial-thesis-pattern-v01
Display posture: Legacy pattern support
Posture: declared-legacy-pattern
Evidence: site/registry.json supports in source order

ARTIFACT CONTENT
[HEADING 1] Do Not Outsource the Brain

[PARAGRAPH] In the AI era, contingent labor still has a place. But the enterprise moat belongs to the people who understand the business, the systems, the failure modes, and what should be built next.

[PARAGRAPH] By Tony Malott

[PARAGRAPH] Original thesis from Tony Malott’s own thinking, generalized professional experience, public sources, and independent analysis.

[LIST ITEM] Core thesis

[LIST ITEM] 1. The old sourcing model

[LIST ITEM] 2. AI moves value upstream

[LIST ITEM] 3. Tool access is not differentiation

[LIST ITEM] 4. The labor market is rewarding judgment

[LIST ITEM] 5. Outsourcing still has a place

[LIST ITEM] 6. The real risk is cognitive outsourcing

[LIST ITEM] 7. Governance & maintainability

[LIST ITEM] 8. The domain-systems orchestrator

[LIST ITEM] 9. Internal capability is not bureaucracy

[LIST ITEM] 10. The AI failure pattern is visible

[LIST ITEM] 11. The solution is not "AI everywhere"

[LIST ITEM] 12. The new moat

[LIST ITEM] Claim / Evidence Ledger

[LIST ITEM] Source Index

[HEADING 2] Core thesis

[PARAGRAPH] AI does not make enterprise labor interchangeable. It makes high-context internal capability more valuable.

[PARAGRAPH] For years, large organizations have used contingent labor, outsourcing, and managed service providers to gain flexibility. That model still has a place. Elastic work, bounded execution, commodity services, specialized surge capacity, and repeatable operational tasks can often be sourced externally without damaging the enterprise. The mistake is assuming that same logic applies to the cognitive center of the business.

[PARAGRAPH] AI changes where value accumulates.

[PARAGRAPH] The scarce layer is no longer just syntax, ticket throughput, script writing, documentation production, or raw task execution. AI keeps pushing those things down the cost curve. The scarce layer is now judgment: knowing what should be built, why it matters, how the business actually works, which constraints matter, what failure looks like, who must own the outcome, and whether the solution will still be maintainable after the first impressive demo.

[PARAGRAPH] That is the part enterprises cannot casually outsource.

[PARAGRAPH] Drucker lens: the decisive asset is not labor hours. It is knowledge converted into repeatable performance.

[PARAGRAPH] Torvalds and Knuth lens: the value is not the code-shaped artifact. The value is architecture, maintainability, failure discipline, and knowing when a clever thing is actually garbage with better lighting.

[HEADING 2] 1. The old sourcing model was built around labor flexibility

[PARAGRAPH] Contingent labor and outsourcing make sense when the work is elastic, bounded, and easy to specify. If demand spikes, a company can add capacity. If conditions change, it can reduce capacity. If a task is standardized, external scale can be rational.

[PARAGRAPH] That is not the problem.

[PARAGRAPH] The problem starts when companies treat internal capability as an avoidable cost rather than a strategic asset. That thinking was already risky before AI. With AI, it becomes more dangerous because the people who understand the work can now use AI to multiply their output.

[PARAGRAPH] The question is no longer, “How many people do we need to perform this task?”

[PARAGRAPH] The better question is, “Which people understand the system deeply enough to supervise AI, redesign the workflow, validate the output, and own the result?”

[PARAGRAPH] That is a very different labor equation.

[BLOCKQUOTE] Do not outsource the brain. Outsource capacity, not cognition.

[HEADING 2] 2. AI moves value upstream

[PARAGRAPH] AI lowers the cost of producing drafts, scripts, summaries, prototypes, analysis scaffolding, documentation, test cases, and workflow fragments. That does not make expertise less important. It makes expertise more leveraged.

[PARAGRAPH] The person with domain knowledge, system context, business fluency, and operational judgment can now do the work of many narrow executors, not because AI is magic, but because AI turns thought into artifacts faster.

[PARAGRAPH] The person without that context can also produce artifacts faster. That is the problem. AI makes shallow output cheaper too. A weak operator with a strong model can generate impressive-looking nonsense at enterprise speed. Humanity, naturally, has turned this into a procurement category.

[PARAGRAPH] The differentiator is not access to the model. Everyone gets access.

[PARAGRAPH] The differentiator is knowing:

[LIST ITEM] What problem is worth solving

[LIST ITEM] Which process is actually broken

[LIST ITEM] Which requirements are real versus inherited theater

[LIST ITEM] Which constraints are regulatory, operational, security, cultural, or political

[LIST ITEM] Which outputs are plausible but wrong

[LIST ITEM] Which automations are useful

[LIST ITEM] Which automations are dangerous

[LIST ITEM] Which AI use cases should not exist at all

[LIST ITEM] Which solutions can be supported after the vendor leaves

[PARAGRAPH] That is the moat.

[BLOCKQUOTE] AI makes shallow output cheaper. It makes deep context more valuable.

[HEADING 2] 3. Tool access is not differentiation

[PARAGRAPH] BCG’s 2026 AI at Work research shows how widespread AI use has become: 74% of frontline employees report using AI every day or a few times per week, and 42% of regular frontline AI users report saving eight hours per week. But BCG’s core warning is the important part: most organizations have not figured out how to convert saved time into value, and strategic clarity matters more than access to tools.

[PARAGRAPH] BCG also reports that 66% of regular frontline AI users receive limited or no guidance on what to do with time saved, and more than half are not reinvesting that time into more strategic work. That is the enterprise AI trap in one ugly sentence: the tool works locally, but the organization fails systemically.

[PARAGRAPH] The lesson is blunt: buying tools is not transformation. Transformation requires workflow redesign, management clarity, incentives, capability building, and a decision system that tells people where AI belongs and where it does not.

[HEADING 2] 4. The labor market is already rewarding judgment

[PARAGRAPH] PwC’s 2026 Global AI Jobs Barometer analyzed more than one billion job ads across six continents and found that AI is creating a two-track labor market. Skills needed for the most AI-exposed jobs are changing more than twice as fast as skills in the least AI-exposed jobs, and jobs “professionalised” by AI are growing twice as fast as democratized jobs, with 42% faster wage growth since 2021. PwC also found that the most AI-exposed junior roles are seven times more likely to demand traditionally senior skills like leadership.

[PARAGRAPH] That supports the thesis directly. AI is not simply flattening skill. In many roles, it is raising the bar. Routine work is being compressed, and higher-order skills are being pulled forward earlier in the career ladder.

[PARAGRAPH] This matters because companies that over-index on external execution capacity may underinvest in the exact internal people who can convert AI into durable advantage.

[PARAGRAPH] The future belongs to people who can combine:

[LIST ITEM] Domain knowledge

[LIST ITEM] Systems thinking

[LIST ITEM] Business translation

[LIST ITEM] Technical fluency

[LIST ITEM] Judgment

[LIST ITEM] AI supervision

[LIST ITEM] Process redesign

[LIST ITEM] Operational accountability

[PARAGRAPH] That is not generic labor. That is institutional capability.

[BLOCKQUOTE] The companies that rent their AI thinking will rent their future.

[HEADING 2] 5. Outsourcing still has a place, but not at the brain layer

[PARAGRAPH] This is not an anti-outsourcing argument. That would be lazy, and therefore perfectly suited for a conference panel.

[PARAGRAPH] Deloitte’s 2026 outsourcing research says many organizations are moving toward multidimensional workforce models that combine global capability centers, AI, and outcome-based managed service providers. Deloitte also reports that 70% of surveyed organizations brought previously outsourced work back in-house during the last five years to strengthen internal capabilities, improve service quality, and reduce vendor markups. At the same time, 67% adopted outcome-based outsourcing models, showing that external providers remain important when tied to measurable results.

[PARAGRAPH] That is the right balance.

[PARAGRAPH] Outsource capacity. Outsource commodity execution. Outsource bounded specialization. Outsource surge work. Outsource work where the requirements are stable and the learning does not need to compound inside the enterprise.

[PARAGRAPH] Do not outsource the brain.

[PARAGRAPH] The brain is the internal capability that understands business context, system dependencies, failure history, regulatory boundaries, operational handoffs, support economics, user behavior, and long-term maintainability.

[PARAGRAPH] When that capability lives outside the company, the company may still receive deliverables. What it loses is memory.

[PARAGRAPH] And in the AI era, memory is leverage.

[HEADING 2] 6. The real risk is cognitive outsourcing

[PARAGRAPH] The failure mode is not that a vendor builds a bad AI solution. That happens. The market will survive. Probably by selling the cleanup project.

[PARAGRAPH] The deeper failure mode is cognitive outsourcing: the company pays external providers to build AI systems it does not understand, cannot evaluate, cannot maintain, and cannot evolve.

[PARAGRAPH] That creates demo debt.

[PARAGRAPH] Demo debt is the gap between what looks impressive in a controlled narrative and what survives inside real operations.

[PARAGRAPH] Demo debt appears when:

[LIST ITEM] The proof of concept worked only because experts hand-fed it clean context

[LIST ITEM] The system cannot handle messy exceptions

[LIST ITEM] The AI workflow breaks against permissions, identity, data quality, latency, or handoff boundaries

[LIST ITEM] Users do not trust the output

[LIST ITEM] The assistant adds review burden instead of removing work

[LIST ITEM] The automation is less efficient than deterministic scripting

[LIST ITEM] No one knows how to tune, repair, validate, or retire the thing after launch

[LIST ITEM] The vendor leaves and the internal team inherits an artifact, not a capability

[PARAGRAPH] Gartner’s 2025 agentic AI forecast supports this concern. Gartner predicts that over 40% of agentic AI projects will be canceled by the end of 2027 because of escalating costs, unclear business value, or inadequate risk controls. Gartner also warns about “agent washing,” where vendors relabel assistants, RPA, and chatbots as agentic AI without substantial agentic capability.

[PARAGRAPH] The point is not that agentic AI is fake. The point is that weak use cases, weak governance, and weak ownership are expensive.

[BLOCKQUOTE] Demo debt is what happens when an AI system survives the presentation but dies in operations.

[HEADING 2] 7. AI increases the importance of governance and maintainability

[PARAGRAPH] Deloitte’s 2026 State of AI in the Enterprise report says 34% of surveyed organizations are using AI to deeply transform products, services, processes, or business models, 30% are redesigning key processes around AI, and 37% are using AI at a more surface level with little or no change to existing processes.

[PARAGRAPH] The same Deloitte report says governance is the difference between scaling successfully and stalling out. It emphasizes senior leadership involvement, human oversight, auditability of automated decisions, retained records of system behavior, independent validation, and integration with existing risk structures.

[PARAGRAPH] That is the boring part. Naturally, it is also the part that decides whether anything works.

[PARAGRAPH] AI systems are not fire-and-forget deliverables. They require:

[LIST ITEM] Evaluation

[LIST ITEM] Monitoring

[LIST ITEM] Human validation

[LIST ITEM] Failure classification

[LIST ITEM] Data governance

[LIST ITEM] Access governance

[LIST ITEM] Prompt and context governance

[LIST ITEM] Model-routing decisions

[LIST ITEM] Exception handling

[LIST ITEM] Operational support

[LIST ITEM] Retirement criteria

[LIST ITEM] Cost controls

[LIST ITEM] Ownership boundaries

[PARAGRAPH] If the enterprise cannot do those things internally, it does not own the capability. It rents the appearance of one.

[HEADING 2] 8. The most valuable role is the domain-systems orchestrator

[PARAGRAPH] The new high-value operator is not simply a developer, architect, analyst, service owner, or business partner. The emerging role is a domain-systems orchestrator.

[PARAGRAPH] That person can translate messy business need into executable work. They know when to use AI, when to use deterministic automation, when to simplify the process, and when to stop because the requested solution is nonsense with budget approval.

[PARAGRAPH] They can supervise agents without pretending autonomy is magic. They can inspect outputs, detect plausible garbage, classify failures, and keep humans in the loop where judgment matters. They understand that AI is not the operating model. AI is one component inside the operating model.

[PARAGRAPH] Deloitte’s 2026 Human Capital Trends report frames this clearly: organizations should stop layering AI onto legacy roles and processes and should instead design human-AI interactions deliberately at strategy, governance, workflow, role, and team levels. It also says decision rights, escalation paths, accountability, leadership, psychological safety, and culture are central to human-machine work design.

[PARAGRAPH] That is exactly the operating layer companies need to build internally.

[HEADING 2] 9. Internal capability does not mean internal bureaucracy

[PARAGRAPH] Internal capability is not headcount nostalgia. It does not mean every company should build everything itself or hire armies of permanent staff.

[PARAGRAPH] It means the enterprise needs a small, serious internal core that owns the cognitive control plane:

[LIST ITEM] Which problems matter

[LIST ITEM] Which workflows deserve automation

[LIST ITEM] Which AI patterns are approved

[LIST ITEM] Which sources are authoritative

[LIST ITEM] Which claims AI systems may make

[LIST ITEM] Which outputs require human review

[LIST ITEM] Which vendors are useful

[LIST ITEM] Which vendor outputs are unacceptable

[LIST ITEM] Which systems are maintainable

[LIST ITEM] Which solutions should be killed

[PARAGRAPH] That internal core can still use contractors and vendors. In fact, it should. But vendors should extend internal capability, not replace it.

[PARAGRAPH] Deloitte’s outsourcing research makes this distinction useful: next-generation providers can complement in-house efforts by bringing specialist knowledge, advanced technology, and scale, but the best model is a multidimensional sourcing strategy, not blind dependence on one labor channel.

[PARAGRAPH] The operating principle is simple:

[PARAGRAPH] Use external capacity to scale execution. Keep internal capability to own judgment.

[BLOCKQUOTE] A vendor can deliver an artifact. The enterprise still has to own the judgment.

[HEADING 2] 10. The enterprise AI failure pattern is already visible

[PARAGRAPH] McKinsey’s 2025 State of AI survey found that 88% of respondents report regular AI use in at least one business function, but only about one-third say their organizations have begun scaling AI programs across the enterprise. McKinsey also found that high performers are far more likely to redesign workflows, define when model outputs need human validation, embed AI into business processes, and track KPIs for AI solutions.

[PARAGRAPH] That means the difference between AI theater and AI value is not whether the organization has access to AI.

[PARAGRAPH] It is whether the organization changes the work.

[PARAGRAPH] The World Economic Forum’s Future of Jobs Report 2025 reinforces the workforce side. It says AI and information processing are expected to transform business for 86% of employers by 2030, and that AI and big data, networks and cybersecurity, and technological literacy are among the fastest-growing skills.

[PARAGRAPH] WEF also reports that skill gaps are the biggest barrier to business transformation, cited by 63% of employers, while 85% plan to prioritize upskilling their workforce.

[PARAGRAPH] So the strategic question becomes uncomfortable:

[PARAGRAPH] Why would an enterprise underinvest in internal AI-era capability at the exact moment skill gaps are the largest transformation barrier?

[PARAGRAPH] Because short-term cost optics are easy. Capability math is harder. Naturally, management dashboards prefer the easy thing.

[HEADING 2] 11. The solution is not “AI everywhere”

[PARAGRAPH] The solution is disciplined selectivity.

[PARAGRAPH] Use AI where uncertainty, language, synthesis, judgment support, exploratory analysis, or adaptive interaction is the work.

[PARAGRAPH] Use deterministic software where the process is known, repeatable, and rule-bound.

[PARAGRAPH] Use automation where consistency matters more than improvisation.

[PARAGRAPH] Use humans where accountability, ethics, exception handling, system judgment, or business meaning matter.

[PARAGRAPH] Use vendors where bounded expertise, scale, or acceleration is needed.

[PARAGRAPH] Use internal capability where context compounds.

[PARAGRAPH] That last phrase is the article’s practical decision rule:

[PARAGRAPH] If context compounds, keep the capability inside.

[BLOCKQUOTE] If context compounds, keep the capability inside.

[HEADING 2] 12. The new moat

[PARAGRAPH] The new enterprise moat is not the model. The model will change.

[PARAGRAPH] It is not the prompt. The prompt will decay.

[PARAGRAPH] It is not the vendor implementation. The vendor may leave before the support burden becomes obvious.

[PARAGRAPH] It is not the assistant interface. Most assistant interfaces currently feel like making a human fill out a form so a chatbot can slowly do something a script could have done better. A triumph of progress, apparently.

[PARAGRAPH] The moat is the internal capability to decide:

[LIST ITEM] What should be built

[LIST ITEM] What should not be built

[LIST ITEM] What should be automated deterministically

[LIST ITEM] What should use AI

[LIST ITEM] What should remain human

[LIST ITEM] What failure looks like

[LIST ITEM] What evidence is sufficient

[LIST ITEM] What support model is viable

[LIST ITEM] What economics justify continuation

[LIST ITEM] What must be retired

[PARAGRAPH] That is where enterprise value is moving.

[PARAGRAPH] AI does not eliminate the need for internal expertise. It punishes organizations that hollow it out.

[HEADING 2] Claim / Evidence Ledger

[TABLE ROW] Claim | Evidence Posture | Best Source Support

[TABLE ROW] AI access is no longer differentiation. | Strongly supported. AI use is broad, but value depends on strategy and workflow redesign. | BCG 2026, McKinsey 2025.

[TABLE ROW] AI raises the value of judgment, leadership, and expertise. | Strongly supported. AI-exposed jobs are changing faster and often demand more senior skills. | PwC 2026, WEF 2025.

[TABLE ROW] Outsourcing remains valid, but must be used at the right layer. | Strongly supported. Deloitte shows both insourcing for capability and outcome-based outsourcing for external leverage. | Deloitte Outsourcing 2026.

[TABLE ROW] Agentic AI has a real hype and failure problem. | Strongly supported. Gartner explicitly forecasts cancellations and warns about agent washing. | Gartner 2025.

[TABLE ROW] Governance, evaluation, and human oversight determine whether AI scales. | Strongly supported. Deloitte and McKinsey both link value to governance, workflow redesign, validation, and operating practices. | Deloitte 2026, McKinsey 2025.

[TABLE ROW] The internal moat is domain knowledge plus systems judgment. | Inferred thesis, supported by labor, AI adoption, governance, and sourcing evidence. | PwC, BCG, Deloitte, McKinsey, WEF.

[TABLE ROW] Many custom enterprise AI tools struggle to reach production because they do not fit real workflows. | Directional support only. Useful but should be caveated. | MIT NANDA 2025, directional due to stated limitations.

PUBLIC SURFACES
Human page: https://next.shareplane.malott.ai/artifacts/do-not-outsource-the-brain/
Metadata JSON: https://next.shareplane.malott.ai/artifacts/do-not-outsource-the-brain/artifact.json
Receipt: https://next.shareplane.malott.ai/artifacts/do-not-outsource-the-brain/receipt.json
Context: https://next.shareplane.malott.ai/artifacts/do-not-outsource-the-brain/context.txt
Agent-package manifest: https://next.shareplane.malott.ai/artifacts/do-not-outsource-the-brain/agent-package.json
Agent-package ZIP: https://next.shareplane.malott.ai/artifacts/do-not-outsource-the-brain/agent-package.zip
Collection catalog: https://next.shareplane.malott.ai/catalog.json
Graph: https://next.shareplane.malott.ai/graph.json
Agent index: https://next.shareplane.malott.ai/llms.txt
