CURRENT | Foundations

Field Guide

When Agents Can't Find What's Right in Front of Them

By Rina Takahashi— February 11, 2026

Feature image for article: When Agents Can't Find What's Right in Front of Them

A web scraper pulls product prices from a retail site every morning. One day it returns empty results. The prices are still there, displayed correctly on the page—customers see them, competitors see them. But the agent can't find them. Teams spend days checking pipelines and validating formats. Something broke, something most debugging approaches miss entirely.

Field Guide

When Agents Can't Find What's Right in Front of Them

By Rina Takahashi— February 11, 2026

A web scraper pulls product prices from a retail site every morning. One day it returns empty results. The prices are still there, displayed correctly on the page—customers see them, competitors see them. But the agent can't find them. Teams spend days checking pipelines and validating formats. Something broke, something most debugging approaches miss entirely.

Tools & Techniques

Tools in Context

Headless Browsers at Scale: Where Speed Determines Economics

The infrastructure lead switches 10,000 daily hotel price checks from headful to headless browsers. Monthly bill: $4,500 to $1,500. Tasks complete 2-15x faster. Same server handles three times the concurrent sessions. For teams running automation at volume, headless execution determines how many servers you need and what they cost. Then anti-bot systems evolve to catch the missing rendering pipeline, the absent plugins, the suspicious timing patterns. Engineering around detection becomes its own cost, one that sometimes exceeds the infrastructure savings.

Tools in Context

When Headful Browsers Justify Their Resource Overhead

The pricing monitoring system ran cleanly for three months—15,000 hotel properties checked daily, data flowing into automated pricing decisions. Then Booking.com updated their anti-bot system. Success rate dropped from 94% to 31% within 48 hours. Missing pricing data meant stale dashboards. Stale dashboards meant the system underpriced 400 properties based on weekend rates that were four days old. The revenue team calculated the loss. The cost of being wrong exceeded the cost of running more expensive infrastructure by a factor of eight.

Tools & Techniques

Tools in Context

Headless Browsers at Scale: Where Speed Determines Economics

The infrastructure lead switches 10,000 daily hotel price checks from headful to headless browsers. Monthly bill: $4,500 to $1,500. Tasks complete 2-15x faster. Same server handles three times the concurrent sessions. For teams running automation at volume, headless execution determines how many servers you need and what they cost. Then anti-bot systems evolve to catch the missing rendering pipeline, the absent plugins, the suspicious timing patterns. Engineering around detection becomes its own cost, one that sometimes exceeds the infrastructure savings.

Tools in Context

When Headful Browsers Justify Their Resource Overhead

The pricing monitoring system ran cleanly for three months—15,000 hotel properties checked daily, data flowing into automated pricing decisions. Then Booking.com updated their anti-bot system. Success rate dropped from 94% to 31% within 48 hours. Missing pricing data meant stale dashboards. Stale dashboards meant the system underpriced 400 properties based on weekend rates that were four days old. The revenue team calculated the loss. The cost of being wrong exceeded the cost of running more expensive infrastructure by a factor of eight.

Tools & Techniques

Tools in Context

Headless Browsers at Scale: Where Speed Determines Economics

The infrastructure lead switches 10,000 daily hotel price checks from headful to headless browsers. Monthly bill: $4,500 to $1,500. Tasks complete 2-15x faster. Same server handles three times the concurrent sessions. For teams running automation at volume, headless execution determines how many servers you need and what they cost. Then anti-bot systems evolve to catch the missing rendering pipeline, the absent plugins, the suspicious timing patterns. Engineering around detection becomes its own cost, one that sometimes exceeds the infrastructure savings.

Tools in Context

When Headful Browsers Justify Their Resource Overhead

The pricing monitoring system ran cleanly for three months—15,000 hotel properties checked daily, data flowing into automated pricing decisions. Then Booking.com updated their anti-bot system. Success rate dropped from 94% to 31% within 48 hours. Missing pricing data meant stale dashboards. Stale dashboards meant the system underpriced 400 properties based on weekend rates that were four days old. The revenue team calculated the loss. The cost of being wrong exceeded the cost of running more expensive infrastructure by a factor of eight.

In Dialogue With Complexity

An Interview with the Acceleration Curve That Judges Your Humanity

In Dialogue With Complexity

An Interview with the Acceleration Curve That Judges Your Humanity

Pattern Recognition

Platform Vendors Launch Before Production Proves Out

Three major vendors launched agent management platforms in February 2026. OpenAI released Frontier on February 5. Salesforce expanded Agentforce throughout January and February. AWS grew Bedrock AgentCore. Gartner called these platforms "the most valuable real estate in AI."

Only 11% of organizations have agents in production, according to Deloitte. Gartner forecasts over 40% of agentic AI projects will be canceled by end of 2027. Vendors are racing to own infrastructure before production use cases prove out at scale, betting that controlling the platform matters more than solving deployment challenges first.

Pattern Recognition

Platform Vendors Launch Before Production Proves Out

Three major vendors launched agent management platforms in February 2026. OpenAI released Frontier on February 5. Salesforce expanded Agentforce throughout January and February. AWS grew Bedrock AgentCore. Gartner called these platforms "the most valuable real estate in AI."

Only 11% of organizations have agents in production, according to Deloitte. Gartner forecasts over 40% of agentic AI projects will be canceled by end of 2027. Vendors are racing to own infrastructure before production use cases prove out at scale, betting that controlling the platform matters more than solving deployment challenges first.

IN SUMMARY

Early adopters:

OpenAI's Frontier launched with HP, Intuit, Oracle, State Farm, Thermo Fisher, and Uber as initial enterprise customers testing the platform.

Pilot failures:

MIT study analyzing 300+ enterprise initiatives found 95% of generative AI pilots fail to deliver measurable ROI in production environments.

Process gaps:

Celonis survey of 1,649 business leaders found 76% admit current processes hold them back despite wanting to become "agentic enterprises" within three years.

Vendor hesitation:

Enterprises avoid committing to single-vendor platforms as AI strategies evolve rapidly, preferring hybrid approaches with multiple LLM options available.

Integration obstacles:

46% of respondents cite integration with existing systems as their primary barrier to deploying agents at scale across their organizations.

Questions Worth Asking

Most teams pick tools based on demos that showcase perfect scenarios. Then production arrives with its corrupted files, peak-hour spikes, and users who do things no one anticipated.

We've seen this pattern repeatedly. The system that looked solid in the vendor presentation becomes the thing demanding constant attention once real load hits. The gap between demo and deployment is where expensive mistakes live.

These questions surface what matters when systems scale. They distinguish tools that fade into infrastructure from those that become operational burdens. Ask them before you're locked in.

Questions Worth Asking

Most teams pick tools based on demos that showcase perfect scenarios. Then production arrives with its corrupted files, peak-hour spikes, and users who do things no one anticipated.

We've seen this pattern repeatedly. The system that looked solid in the vendor presentation becomes the thing demanding constant attention once real load hits. The gap between demo and deployment is where expensive mistakes live.

These questions surface what matters when systems scale. They distinguish tools that fade into infrastructure from those that become operational burdens. Ask them before you're locked in.

Production Reality

Does It Handle Corrupted Files?

Demos use clean inputs. Production delivers corrupted PDFs, malformed JSON, and queries that make no sense. The tool that chokes on the first weird file isn't a tool. It's a liability waiting to page you at 2 AM.

Parallel Execution

Can It Run 10,000 Tests Simultaneously?

Sequential testing is a luxury production can't afford. You need parallel execution across distributed resources, not a queue that takes hours. If the tool can't shard tests and run them concurrently, you're building a bottleneck into your pipeline.

Environmental Parity

Does QA Actually Match Production?

Code that works in staging and fails in production is expensive. Not just frustrating. Your QA environment needs the same hardware, network topology, and configuration drift that production has. Otherwise you're testing a fantasy version of your system.

Integration Depth

Are Integrations Native or Duct-Taped?

Bolted-on integrations break. Native ones fade into infrastructure. Check whether the tool connects to your actual systems—GitHub, Datadog, PagerDuty—without custom glue code. If you're writing adapters, you're maintaining two systems instead of one.

Rollback Capability

How Fast Can You Undo This?

Rollback sounds simple until you try it with database migrations and dependent services. The tool should make reverting trivial, not a multi-step procedure requiring three teams. If rollback takes longer than the deploy, something's wrong with your architecture.

Vendor Viability

Will This Company Survive Your Contract?

Tool selection is a three-year bet minimum. Check how long they've existed, whether they're actively shipping updates, and if their business model makes sense. Migrating off a dead vendor is worse than choosing wrong in the first place.