The use case: how Intercom deploys Honeycomb
Intercom uses Honeycomb distributed tracing and SLOs to observe and operate its Fin AI customer service agent. They track end-to-end time to first token and cost per interaction in the same traces, with low-level telemetry for investigating degradations.
More reported results
Figures published in Honeycomb’s case study. They are not independently measured by usedby.
This optimization alone helped us to shave two seconds off the median time to first token.
In the customer’s words
We were so focused on Fin’s latency that we lost touch with our own business by incurring extra operational cost. Luckily, thanks to the great tracing we already had, we were able to identify and fix the issue. To validate our fixes, we used the same query that Finance was using to quantify the extra LLM cost. We were happy to observe the LLM waste % going down, confirming our fix.Other AI tools used by Intercom
Other public relationships recorded on usedby.
Explore Intercom’s AI stackOther companies using Honeycomb
Customers across industries, with public evidence.
AmperityConfirmed
Booking.comConfirmed
Cruise.co.ukConfirmed
DepotConfirmed
DuolingoConfirmed
FenderConfirmedGo straight to the source
Based on Honeycomb’s published case study. We check the public evidence; the results are reported by the vendor.
honeycomb.io- Source published
- Feb 10, 2026
- Last checked
- Nov 13, 2026
Open source capture What the story claims, and what we checked
What we compared with the page.
- The figure: 60%Checked
The figure is on the page; its label is our wording.
- The publication dateChecked
Read from the page’s own metadata, never guessed.
- Intercom uses HoneycombChecked
Confirmed line.
- The result itselfNot checked
We quote it; we did not measure it.
So, we added Fin’s cost per interaction to the same distributed traces we already had.






