Today, we are introducing Traversal Workers. No tag required.
The industry's first AI agents built to proactively investigate wherever they're needed, from alert triage and deployment monitoring to complex production incidents, without being scheduled or summoned. It feels like having your favorite SRE teammate seamlessly problem solving with you.
Because AI in production shouldn't just know how to investigate. It should know when to act.
Learn more: https://t.co/ruwvgrHUtT
What better place for @traversal_ai 's first out-of-home campaign than the central nervous system of NYC: the subway system. When it’s down, the whole city comes to a halt.
Production plays the same role for our customers. When it goes down, their business stops.
We’re proud New Yorkers, and proud that teams rely on Traversal when their systems are on the line.
Most AI pilots die in the InfoSec and procurement stage.
This is especially true of AI SRE products, which need access to sensitive company log data.
We explain why, and how to avoid it.
When vendors demand root access, intrusive agents, and broad cloud API permissions before proving they can deliver value, it typically stalls momentum and kills a pilot before it ever gets started. We even heard from a customer that another AI SRE vendor had asked to install a sidecar in every pod — just for the pilot!
Here’s the thing: an AI SRE's job is primarily to interpret telemetry, not create it. Asking for invasive access early is often a signal that a vendor is either taking technical shortcuts or intentionally vying for vendor lock-in.
There's a smarter path.
At Traversal, some of our largest deployments run entirely on read-only API access to existing observability platforms like Datadog, ServiceNow, or Splunk. When accuracy demands more, we move to pipeline integration (Kafka, Cribl, OpenTelemetry) to deepen analysis without adding operational overhead. Agents and sidecars? They’re rarely necessary, if ever.
The other place pilots get stuck: LLM approval. Many enterprises maintain pre-approved model lists, and vendors who can't adapt to your existing AI infrastructure force you into months of additional InfoSec reviews.
The gold standard isn't a single integration pattern — it's flexibility. Start with the lightest integration (usually API access), prove value quickly, and expand only as needed. Work with vendors who adapt to your environment, not the other way around.
In her latest piece for @axio, Director @sais_idev @nagpaltweets writes on how global nutrition has improved, but targeted investments are still needed. https://t.co/NVqE1zj57L
Excited to discuss my research on solidarity levies, and hear fellow panelists discuss other innovative financing mechanisms for development. If you're in DC, swing by SAIS during lunchtime this Wednesday!
Lesson 2 - Learn more effective ways of raising resources for water from the health and transport sectors. Check out our full report to learn more! @sais_idev@urbaninstitute
Excited to have partnered with the Urban Institute to research innovative financing solutions for safe drinking water provision. Check out a summary of our findings here!
The @WHO estimates that more than 2 billion people lack access to safely managed drinking water services. From @Matt_Eldridge & @CHauenstein: https://t.co/n2rEBU2mbg