ping_wu

CEO @ x-googler / Co-founded: GCP Vertex AI; Google Contact Center AI;

[james hong @jhong]

Lots of VCs kindof sound like this: "Your basketball team doesn't have a moat, any healthy human can throw the ball into the basket thingy. "

People who have not run companies before don't understand how much the team matters more than some imagined structural moat.

[ping_wu @ping_wu]

https://t.co/kN4lAeiDR5

[ping_wu @ping_wu]

Today we're launching @cresta Deputy Agent: a long-running AI agent that automates the workflows behind customer experience and carries each one through to completion.

One customer request can span teams, systems, and weeks of work. Data is re-entered, approvals are chased, and progress updates depend on someone checking in. Without end-to-end ownership, every handoff becomes a potential source of delay, cost, and risk.

Deputy automates it from start to finish:

1️⃣ It runs as long as the work does. It tracks what's done, what's missing and what comes next, maintains progress through every pause, and resumes the moment the next input arrives.

2️⃣ It adapts. It uses real-time customer context, plus records, business rules and SOPs, to decide the next step.

3️⃣ People stay in control. The business decides what Deputy can do on its own and what needs approval. Every action is visible and auditable.

This is the next step in the evolution of the AI operating system for customer experience.

For the first time, businesses get continuity across the whole workflow, with one agent owning the full process.

[ping_wu @ping_wu]

https://t.co/dWsKuNlhxd

[ping_wu @ping_wu]

Today, we’re introducing @cresta Sentry-0.1, a personal AI agents detection model.

Cresta customers saw personal AI agent traffic spike more than 80x in September. With more consumers using these agents to make calls on their behalf, businesses need to rethink how much demand to expect and how to respond.

We built our Sentry model into Cresta’s AI operating system, alongside the intelligence, guidance, and automation already powering customer interactions. In early customer usage, these have looked like:

—Using indirect language like “the account” instead of “me” or “I”
—Stating limits like “I can’t click the link you shared”
—Combining multiple requests into one statement like “please confirm when it’ll be in stock, and let me know if shipping is free”

When Sentry recognizes a personal agent, Cresta AI Agents can follow your business’s policy for what personal agents can and can’t do. Furthermore, Cresta can guide your human representatives to receive relevant authentication and policy guidance. Plus, CX leaders can analyze tagged conversations in Cresta Insights and Cresta AI Analyst to understand what’s driving these calls and identify where policies or customer experiences need to change.

Super exciting time. A lot more to come from us in this space. Stay tuned!

[ping_wu @ping_wu]

I’m excited to share that @cresta is partnering with @comcast to transform the customer experience for Xfinity and Comcast Business customers.

Comcast uses Cresta to automate and augment the customer experience across voice and chat for sales, retention, and care. With Conversation Intelligence analyzing 100% of calls, each conversation makes the next one better.

For a Fortune 50 company like Comcast, that’s a powerful opportunity to transform customer experience at scale, bringing better service to millions of people.

Thank you to the Comcast team for the partnership! We’re excited for the work ahead.

[Tech Week @Techweek_]

WICKED hackathon for building voice agents with @cresta / @AnthropicAI

[ping_wu @ping_wu]

https://t.co/HMF8qInSex

[ping_wu @ping_wu]

We've been working with Anthropic to put years of CX engineering experience into the hands of more teams through Conductor, our agent-building agent.

Building production-ready AI agents takes judgment: which context matters, where a conversation can stay flexible, and which business rules must be enforced every time. We've built that judgment through real customer deployments, and Conductor helps teams apply it without starting from scratch.

The Claude Agent SDK handles the core development work: gathering context, using tools, and writing and running code. That frees our engineers to bring in business context from across the Cresta platform, including conversation intelligence, CX workflows, evaluations, and production feedback, throughout the agent development cycle.

In early use across Cresta, our customers and partners, Conductor has cut initial deployment time roughly in half. But the first build is only about 20% of the work. The rest is testing, optimization, and continuous improvement, and that's where we want each deployment's lessons to make the next agent better.

Link in 🧵 to the blogs.

[ping_wu @ping_wu]

https://t.co/F6eavQ6OBY

[ping_wu @ping_wu]

On the heels of Muse, Instinct, and other personal agents, lots of CX leaders have been asking us the same question...

What happens to the contact center when the caller is an AI agent and not a person?

I've written up my full thoughts below. https://t.co/7AD0TdTf4T

[ping_wu @ping_wu]

https://t.co/TO95c8Avgv

[Lin Qiao @lqiao]

Frontier labs should give top universities real access to frontier models, and structure AI safety to be an open scientific problem.

Can we detect deception, dangerous capabilities, or loss of control before they matter?

Open research compounds. It scales fast when top AI safety researchers learn and build on each other.

[ping_wu @ping_wu]

https://t.co/bbiwVCAHHL

[ping_wu @ping_wu]

Nobody wants to read a paragraph to book a seat.

Excited to launch Custom UI Components for @cresta's AI Agent, expanding what our agents can do beyond language!

Digital channels have a capability voice does not: they can render an interface.. Often, a visual component is the fastest way for a customer to express intent: selecting a time from a calendar, comparing plans side by side, choosing a seat from a cabin map. They shouldn't have to translate prose into action!

Cresta’s Custom UI Components are not hard-coded additions bolted onto the chat experience. Each one is *generated* in Cresta Conductor (our agent-building-agent) at design time, alongside the agent's prompts and tools. So the UI component and the agent’s behavior evolve together.

Each interaction also has a plain-text representation. When a customer selects seat 14C, that action becomes part of the conversation record, preserving the state and context when the interaction moves to a human or another system.

The next generation of AI agents will not be measured only by how well they answer. They will be measured by whether the experience actually resolves what the customer came to do.

Link in thread to learn more.

[Deedy @deedydas]

I heard the greatest story of promotion incentive hacking at FAANG today.

A very well known Engg VP knew that:
a) you can only really be promoted in ~2 yrs. the few cycles after a promotion are pointless
b) you are promoted on your impact, complexity, scope, leadership and independence

In order to make sure his entire team gets promoted in due time, he would just ensure the ~25% of his team 2yrs into their level and "up for promo" this cycle take credit for the work of the entire team, with their consent, gets promoted, and repeat round robin.

Over the years, everyone on his team would get promoted always on time leading to others wanting to join his team. His headcount would keep increasing, and he'd get promoted too! Yet effectively the team was producing about ~50% of their entire capacity because after a promo, no one really had to do any work. Almost everyone on their team clears $1M+/yr.

Beautiful and disgusting example of gaming broken incentive structures at large institutions.

[Lin Qiao @lqiao]

Cybersecurity isn’t a fortress problem, it’s an immunity problem. Think vaccines.

Eliminating pathogen is not practically possible. Vaccines don’t eliminate pathogens. They teach the immune system to recognize, adapt, and respond faster as threats evolve.

Attackers will keep mutating. AI will accelerate that. Only if we can simulate what attacks look like, we can defend. Open models are the best tool to drive both offense simulation and defense.

We as a community should build an adaptive immune systems for software.

[ping_wu @ping_wu]

https://t.co/325UineuhM

[ping_wu @ping_wu]

Today we're launching three new products that push the frontier of conversation intelligence. We're raising our own bar again with this next generation of @cresta Insights.

- Cresta AI Analyst has evolved into a deep research agent with reasoning capabilities. It plans and executes research, builds charts and cohort comparisons, drafts briefs, and recommends next steps, all from a plain English question.

- Topic Discovery can now support a durable, customizable AI-powered taxonomy that becomes a trusted source of truth for why customers are reaching out, without the maintenance burden of traditional approaches.

- Real-Time Trends (RTT) brings the concept of X’s surging topics to omnichannel customer conversations. It continuously detects emerging issues, anomalies, and trending topics. Some of our airline customers now detect outages from call center conversations before their operation teams do.

Together, these capabilities redefine the frontier of this entire category. We’re one step closer to our vision: one proactive customer intelligence system that doesn't wait for someone to ask the right question.

Link in thread to the full blog.

[ping_wu @ping_wu]

https://t.co/W6TEjCTZOD