📊 Full opportunity report: The Agent Trap: Why 90% of AI “Launches” Are Infrastructure Liars on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
In 2026, 90% of AI ‘agent’ launches are actually features built on vendor infrastructure, not independent platforms. This mislabeling creates vendor lock-in and security risks, with only 10% representing genuine infrastructure plays.
Last week, a vendor announced an AI agent marketed as a platform capable of transforming enterprise workflows. However, analysis shows that the product is merely a feature on vendor infrastructure, illustrating a broader industry pattern where 90% of so-called AI agent launches are not true platforms but simple feature upgrades.
In May 2026, a vendor released an AI agent described as a platform, but further scrutiny revealed it lacked core characteristics of a true agent, such as runtime independence, state persistence, and governance controls. Instead, it was a SaaS feature tightly integrated into the vendor’s cloud infrastructure, with no portability or independent control.
This pattern is widespread: most enterprise ‘agent’ launches rely on vendor-controlled environments, with limited ability for customers to swap models, persist state outside vendor systems, or govern security and compliance effectively. Learn more about OpenAI’s new agent SDK. Only about 10% of launches in 2026 qualify as genuine infrastructure platforms, capable of running autonomously and with full control.
Industry experts note that this mislabeling is driven by marketing and pricing strategies, not technical capabilities. The distinction between features and platforms is becoming a key procurement skill for enterprises aiming to avoid vendor lock-in and security vulnerabilities. Understanding what makes a true AI platform is essential.
The agent trap.
Why 90% of AI “launches” are infrastructure liars.
A vendor announces an “AI agent.” The product is a chat box that summarises meeting notes — wired to a SaaS via OAuth, no runtime, no audit trail, no portable state. List price: $30 per seat per month. This is the agent trap. The label has been stripped from its meaning. What enterprises are buying — under the word agent — is overwhelmingly a feature on top of someone else’s infrastructure.
Most “agents” are features wearing infrastructure as a costume.
In 2026, the word agent has been stripped from its meaning. Vendors monetize the label. Buyers inherit the dependency. The asymmetry has a number — and the number does the work this story needs.
Mastering Enterprise Platform Engineering: A practical guide to platform engineering and generative AI for high-performance software delivery
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
A request that fails three or more is a feature.
Run the request against five questions before signing any “AI agent” PO. The 90% fail at least three. The 10% pass all five. Price the line item accordingly — because the vendor won’t.
Does it run when no human is logged in?
A real agent runs on a schedule, on a trigger, or as a daemon. If it only works when a user opens a tab, it’s a feature.
Can you swap the model without losing the work?
Real agents treat the model as substitutable. The runbook, tools, memory, and workflow survive a model change. Features are welded to one model.
Where does the state live?
Real agents persist state to a customer-controlled store with a schema you can query. Features persist to “your conversation history” inside the vendor’s database.
What does the audit trail look like to your SOC?
Real agents emit events into a SIEM or webhook stream the security team subscribes to. Features emit nothing — or vendor-side logs you can’t ingest.
What do you keep when the contract ends?
Real agents leave you with skills, prompts, runbooks, memory, integrations as exportable artifacts. Features leave you with the labor you sank into the vendor’s UI — and nothing else.
TOPDON Carpal OBD2 Scanner Bluetooth – Scanner for Car, AI-Assisted All System Diagnostic Tool for iOS & Android, Vehicle Health Check, 6 Maintenance Services, Smog Check, Repair Guide
[Easy OBD2 Bluetooth Scanner, Complete Car Insights] Enjoy peace of mind with our comprehensive vehicle health check. From…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Salesforce isn’t selling agents. It’s removing the seat.
The dominant 2026 enterprise pattern is “headless 360” — the same Customer 360 / Employee 360 data model the suite sold for two decades, except agents now read and write directly. SDR · CSM · support agent are increasingly configurations of an agent runtime, not job descriptions for human seats.
The 9% genuinely AI-driven layoffs cluster exactly where headless is shipping.
Tier-1 support, junior software engineering, structured-data work — paying customers of a UI. If agents become the operators, the seat license attached to the human disappears. The vendor still gets paid; they just get paid per agent action instead of per human login.
Before · Per-seat humans
After · Headless 360
Managing AI Risk: A Practical Approach to Responsibly Managing AI with ISO 42001
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
A feature cannot be routed.
When you buy a feature agent from a SaaS vendor, you commit to whatever model the vendor chose, at whatever margin the vendor charges. Real infrastructure exposes the model layer. If the vendor can’t tell you what model is running underneath, that is the answer.
QUERY
independent AI runtime environment
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The leverage moves to whoever owns the motherboard — not the chip.
Claude is increasingly the engine inside other people’s products. Legal-tech vendors, customer-success platforms, contract-review startups. This is the Intel Inside playbook. The implication for buyers is not “therefore buy Anthropic.” It is the reverse.
Built on a single closed model.
Brand sits on top of someone else’s chip. Looks like a platform. Priced like one.
- Cabinet vendor sells the platform pricing
- Chip vendor (Anthropic / OpenAI) sets margin
- If the chip vendor moves up the stack, cabinet gets squeezed
- Customer keeps nothing portable when leaving
Runtime that uses models.
Routing, governance, audit, skills layer. The chip is replaceable. The motherboard captures value.
- Multiple models, swappable per-request
- Customer-controlled governance plane
- Skills + integrations are exportable artifacts
- Survives the chip vendor moving up the stack
Skills are the portable infrastructure.
A skill written for Claude Code can be loaded into Codex, into Cursor, into any agent runtime that understands the format. The skill is the IP the customer wrote. The model is the chip. A buyer with 40 skills against an internal runtime can swap the model layer in an afternoon.
declarative · versioned · portable
If the vendor cannot or will not tell you what model is running underneath, that is the answer. You’re not buying an agent platform. You’re buying a wrapper.
Five questions any executive can ask in any vendor pitch.
- Does it run when no human is logged in?
- Can I swap the model without breaking the workflow?
- Where does the state live, and can I query it directly?
- Does it emit events my SOC can ingest?
- When the contract ends, what do I keep?
Four assignments. By role.
Run the five-point filter against every agent line item.
Reclassify each as feature or infrastructure. Re-price accordingly. The exercise will recover budget — usually significant budget.
Inventory the OAuth scopes granted to feature agents.
After Vercel, the agent supply chain is your perimeter. Tokens granted to chat-box agents holding Workspace, GitHub, and CRM scopes are the largest unmanaged risk in the stack.
Per-seat agent SaaS is the most expensive way to buy LLM compute.
Per-action and per-token routing typically costs 60–85% less for the same throughput. Demand the comparison. Vendors that refuse to provide it have answered the question.
Add “AI infrastructure vs feature” to the quarterly risk review.
If management cannot draw the line, the line has not been drawn — and someone else is drawing it for you, on a price tag.
Implications of Mislabeling AI Agents in 2026
This trend matters because labeling features as platforms obscures the true level of control, security, and portability enterprises have over their AI tools. Relying on vendor-controlled infrastructure increases dependency, risks data security breaches, and complicates compliance efforts. It also hampers the ability to adapt or migrate AI workflows, locking organizations into vendor ecosystems and potentially inflating costs.
Understanding this distinction is critical for enterprise decision-makers, as it affects long-term strategy, security posture, and operational resilience in AI deployments.
The Evolution of ‘Agent’ Definitions and Industry Practices
Prior to 2024, an ‘agent’ was a well-defined process: a continuously running, governable entity that maintained state, took actions, and was externally controlled. However, in 2026, the term has been co-opted by vendors to describe simple chat interfaces or feature add-ons that lack core agent capabilities.
This shift is driven by marketing strategies aimed at capturing enterprise budgets and creating lock-in. Vendors bundle features into ‘agent’ labels to command higher prices, even when these products lack the fundamental qualities of true autonomous agents. The industry has thus moved towards a ‘headless 360’ model, where enterprise data models are accessed directly through agent-like interfaces, blurring the line between features and platforms.
“The label has been chosen for what it does to the price tag, not for what it describes.”
— Thorsten Meyer
“Most enterprise ‘agent’ launches rely on vendor-controlled environments, with limited portability or control, making them more features than true platforms.”
— Industry expert
Extent of Industry-Wide Mislabeling and Future Trends
While estimates suggest that 90% of AI ‘agent’ launches are features, the precise percentage may vary, and the pace of genuine platform development remains uncertain. It is also unclear how quickly enterprises will adapt procurement practices to distinguish true platforms from features.
What Enterprise Buyers Should Do Moving Forward
Enterprises should implement rigorous filtering criteria, such as testing for runtime independence, model swapability, state control, security logging, and portability, before investing in AI tools labeled as agents. For more details, see our guide on evaluating AI platforms. Future developments may include clearer industry standards and more transparent vendor disclosures, but organizations must proactively scrutinize offerings to avoid vendor lock-in and security risks.
Key Questions
What is the main difference between a feature and a true AI agent?
A true AI agent operates independently, maintains persistent state, can be governed externally, and can be swapped or replaced without losing work. Features lack these capabilities and are tied to vendor infrastructure.
Why are vendors labeling features as agents?
Vendors do this to command higher prices, create perceived value, and lock customers into their ecosystems, often masking the limited capabilities of the actual product.
What risks do enterprises face by buying feature-based ‘agents’?
They risk vendor lock-in, security vulnerabilities, inability to migrate or control workflows, and increased costs over time due to dependency on vendor infrastructure.
How can organizations identify genuine AI platforms?
By applying a five-point filter: checking for runtime independence, model swapability, state control, security logging, and portability of workflows and data.
What is the industry trend regarding AI ‘agent’ definitions?
The industry is moving towards a headless 360 model, where enterprise data is accessed directly via agent-like interfaces, often blurring the lines between features and platforms.
Source: ThorstenMeyerAI.com
