Uploaded August 2026 | Updated September 2026, 3 weeks ago
Where is your enterprise AI budget actually going, and is that spend translating into measurable quality? As developer teams adopt coding agents like Claude Code, Cursor, and GitHub Copilot alongside custom production LLM applications, tracking inference spend across models, token types, and execution spans becomes critical to proving ROI.
In this episode of AI Builders, we detail how to use Arize AX to gain complete visibility into your AI bills, connect spend directly to eval quality metrics, and automate cost optimization workflows.
Chapters:
00:00 Where is your AI budget going?
00:39 Intro: Cost alongside quality
01:37 Coding agents vs. production agent spend
03:07 The two drivers of AI cost: model price and token usage
05:05 Why cheaper models do not always mean lower cost
06:32 Connecting cost to quality and ROI
08:27 What does a “better” AI result actually mean?
11:52 The cost-to-quality decision framework in Arize AX
14:11 Tracing cost down to individual model calls
15:40 Monitoring AI cost spikes in production
17:04 Using managed agents to investigate cost
19:17 Demo: Setting up the Arize AX Cost Agent
22:15 How managed agents work
24:07 Finding the biggest sources of wasted spend
27:07 Automatically proposing cost-saving code changes
29:15 Why you still need to validate cost reductions
31:07 Common ways to reduce AI costs
33:30 When your evals become too expensive
35:01 Testing cheaper models without sacrificing quality
37:45 Protecting the AI spend that creates value
39:18 Demo: Tracing Claude Code cost and efficiency
41:19 Using Signal to find coding agent inefficiencies
43:22 Turning trace insights into reusable agent skills
45:09 Five takeaways for reducing AI cost without hurting quality
46:17 Q&A: The cost and quality decision framework
47:15 Closing
Resources:
🔬 Phoenix (open source): phoenix.arize.com
🔗 Arize AX: arize.com
📖 OpenInference: github.com/Arize-ai/openinference
📖 Phoenix docs: docs.arize.com/phoenix
What techniques is your team using to keep LLM bills under control as you scale agents? Share your setup in the comments below!
What techniques is your team using to keep LLM bills under control as you scale agents? Share your setup in the comments below!
Subscribe and hit the notification bell to catch every episode of AI Builders! youtube.com/@arizeai?sub_confirmation=1
#AIEngineering #LLMCost #AIObservability
Where is your enterprise AI budget actually going, and is that spend translating into measurable quality? As developer teams adopt coding agents like Claude Code, Cursor, and GitHub Copilot alongside custom production LLM applications, tracking inference spend across models, token types, and execution spans becomes critical to proving ROI.
In this episode of AI Builders, we detail how to use Arize AX to gain complete visibility into your AI bills, connect spend directly to eval quality metrics, and automate cost optimization workflows.
Chapters:
00:00 Where is your AI budget going?
00:39 Intro: Cost alongside quality
01:37 Coding agents vs. production agent spend
03:07 The two drivers of AI cost: model price and token usage
05:05 Why cheaper models do not always mean lower cost
06:32 Connecting cost to quality and ROI
08:27 What does a “better” AI result actually mean?
11:52 The cost-to-quality decision framework in Arize AX
14:11 Tracing cost down to individual model calls
15:40 Monitoring AI cost spikes in production
17:04 Using managed agents to investigate cost
19:17 Demo: Setting up the Arize AX Cost Agent
22:15 How managed agents work
24:07 Finding the biggest sources of wasted spend
27:07 Automatically proposing cost-saving code changes
29:15 Why you still need to validate cost reductions
31:07 Common ways to reduce AI costs
33:30 When your evals become too expensive
35:01 Testing cheaper models without sacrificing quality
37:45 Protecting the AI spend that creates value
39:18 Demo: Tracing Claude Code cost and efficiency
41:19 Using Signal to find coding agent inefficiencies
43:22 Turning trace insights into reusable agent skills
45:09 Five takeaways for reducing AI cost without hurting quality
46:17 Q&A: The cost and quality decision framework
47:15 Closing
Resources:
🔬 Phoenix (open source): phoenix.arize.com
🔗 Arize AX: arize.com
📖 OpenInference: github.com/Arize-ai/openinference
📖 Phoenix docs: docs.arize.com/phoenix
What techniques is your team using to keep LLM bills under control as you scale agents? Share your setup in the comments below!
What techniques is your team using to keep LLM bills under control as you scale agents? Share your setup in the comments below!
Subscribe and hit the notification bell to catch every episode of AI Builders! youtube.com/@arizeai?sub_confirmation=1
#AIEngineering #LLMCost #AIObservability
