Systems | Development | Analytics | API | Testing

Ep 82 | Behavior Change: How Startups Are Making AI Stick

Behavior change is the biggest hurdle in AI adoption. AI only creates value when people make it part of their everyday work. In this episode of The AI Forecast, Paul Muller sits down with Varun Puri, CEO and co-founder of Yoodli, to discuss why successful AI adoption starts with changing how people work. Drawing on his experience at Google, Google X, and as the founder of an AI startup, Varun shares practical lessons on embedding AI into everyday workflows and building habits that stick.

A Google Data Cloud Leader's Formula for Token-Efficient AI

For most enterprises, autonomous AI agents still feel like a risk waiting to happen. Andi Gutmans, Vice President and General Manager for Data Cloud at Google, joins Cindi Howson to explain what it takes to build the data foundation that trustworthy agentic AI depends on. He breaks down how organizations can finally activate the 90 percent of enterprise data that's unstructured, why tokenmaxxing is the wrong way to measure AI value, and how open standards like Apache Iceberg are helping leaders tear down fragmented, multi-cloud data silos and unify data across clouds.

Securing Non-Deterministic Systems: Comprehensive AI Security Testing for Enterprises

Enterprise QA teams are discovering that deploying machine learning models breaks their existing validation pipelines. Legacy testing environments rely on a simple truth: fixed inputs must produce predictable outputs. Because intelligent architectures operate on probabilistic distributions, deterministic testing alone can no longer guarantee reliability. When conducting a code review or architectural risk assessment, treating an active model as a standard black-box API leaves critical flaws unaddressed.

Tosca's AI is coming to on-premises deployments this fall

This fall, Tosca’s suite of AI capabilities will be available for on-premises deployments — giving teams the tools to meet the pressure of the AI era head-on. Customers will gain access to leading AI features, including Agentic Test Automation, AI-powered test creation, MCP connectivity, and AI Workspace, no matter their deployment type or existing infrastructure. AI’s unprecedented development means testing practices that were standard a year ago could already be outdated.

Token-Maxxing and Inference Ops: The New FinOps Frontier

A Head of Product at a major sportswear retailer has a brilliant idea: let’s build an app that sales staff on the shop floor can have on their tablets, and ask their questions there and then, where they serve customers. They set about building. In order for the app to answer questions, it needs to have the information from the 2026 Spring/Summer Catalogue, a mammoth manual, let’s say 300k tokens.

Why Token-Maxxing Is the Wrong Way to Measure AI Success

Silicon Valley has been measuring AI success by token consumption. The more tokens, the more AI transformation. Right? Wrong. Andi Gutmans, Vice President and General Manager for Data Cloud at Google, joins Cindi Howson on the podcast to share that the best context is the context that drives the outcomes you need with the least amount of tokens and processing. Efficiency, not volume, is where the real value is.

How to Accelerate Vulnerability Remediation with AI

Perforce QAC and Klocwork's new AI-assisted code remediation capabilities combine deep static analysis with AI-guided fix recommendations, helping developers resolve issues faster while maintaining compliance, security, and code quality. In this webinar, you'll see a live demo of how teams can accelerate remediation, reduce rework, and enable flexible AI-powered workflows directly within their development environment.

AI Transport v0.6.0: mid-run steering

Steering lets a client redirect an agent while the agent is still working, so a follow-up message reshapes the answer in flight instead of cancelling the Run and starting over, or waiting for it to finish. In the AI Transport SDK, a Run encapsulates the agent's output for a single turn (including multiple iterations around an agentic loop). The last release, v0.5.0, made a single agent turn survive a crash by splitting a Run into re-attemptable Steps.