Sashank Gondala

Cofounder/Chief TokenMaxxing Officer @furtheraicom || Past: @Apple, @GeorgiaTech, @iitbombay

[Sashank Gondala @sgondala2]

We raised $25M in Series A funding, led by @a16z ( @joeschmidtiv, @astrange)

@furtheraicom is building the first AI workspace for Insurance 🧵

[Sashank Gondala @sgondala2]

We've built out internal versions of this (Sherlock and Poirot) few months ago, and they've become a huge force multiplier across the company.

Engineers use them for triage, debugging, and code generation.

What's been more surprising is the adoption across non-technical teams. People use them to understand product behavior, answer questions about usage, pull billing metrics, and explore how different features are being used.

Turns out tagging an agent in Slack is a surprisingly natural interface for getting work done.

[Claude @claudeai]

Introducing Claude Tag, a new way for teams to work with Claude.

In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.

[FurtherAI @furtheraicom]

Today we're launching Eval Studio — a test bench for AI workflows in insurance.

Every few months a new model lands. Every time one does, the same question hits every team running AI in production: does everything still work?

The honest answer, until now, was: we'll find out.

[Sashank Gondala @sgondala2]

@MaruPelkar Insane

[Aman Gour // FurtherAI @amangour30]

What's holding back large-scale AI deployments?

The short answer: Architecture.

At @ITI_Insurtech, we dug into this topic with Arron Lamp CIO of @TMHCCInt's Public Risk Group  - one of the hardest environments in insurance to deploy AI at scale:

↳ 15 lines of business
↳ Large, complex, unusual risks
↳ Policies that are public, so every decision gets scrutinized

If you can make AI work here, you can make it work practically anywhere.

[Joe Schmidt IV @joeschmidtiv]

https://t.co/i7teOUTWgZ

[Richard Sutton @RichardSSutton]

The bitter lesson in 26 words:

Don’t be distracted by human knowledge, as AI has been historically.
Instead focus on methods for creating knowledge that scale with computation, like search and learning.

[Sashank Gondala @sgondala2]

There's no playbook for agent-first software yet. We're writing it in real time.

If these problems excite you, would love to talk!

Full writeup here - https://t.co/BuGI3ISihb

[Sashank Gondala @sgondala2]

Insurance is one of the world's most complex data environments, and what we're building is fundamentally an AI problem.

Here's some of the hard technical challenges we are solving at @furtheraicom 🧵

[FurtherAI @furtheraicom]

Excited to share that @furtheraicom has been named to the 2026 @CBinsights AI 100 — joining the 100 most promising private AI companies in the world, and the only AI for insurance company on the list. 🎉

Insurance is one of the most operationally complex industries in the world. Every submission, policy, and claim is shaped by expertise, judgment, and context — which is why meaningful AI transformation requires real partnership, not just product.

At @furtheraicom, we work alongside insurance companies on that transformation, helping them reshape how the business runs, how decisions get made, and how they compete in an AI-native era. Being recognized alongside the leaders shaping the future of AI tells us we're on the right track.

A heartfelt thank you to @CBinsights, our trusted partners, customers, and team. This milestone belongs to all of us.

The video, of course, was made with AI.

Further.

[Sashank Gondala @sgondala2]

Synthetic Data for Insurance

There's no ImageNet for insurance documents. No large-scale labeled datasets of SOVs, loss runs, or bordereaux.

We're creating synthetic data to improve our agents.

The challenge is generating documents that capture real-world messiness. Too clean and it doesn't help. Too messy and it's not useful either.

You need the right distribution of chaos.

[Sashank Gondala @sgondala2]

@JYFHuang I guess it just knows that it can’t ask for more information

So on boundary conditions it defaults to figuring it out over asking

[Sashank Gondala @sgondala2]

niche claude code observation

if you tell it that you're gonna sleep and finish the work by the time you wake up, it seems to run longer and finish things better

[Sashank Gondala @sgondala2]

@MaruPelkar Lol 🤣

[Sashank Gondala @sgondala2]

@MaruPelkar @x402scan @base This looks quite cool!

[Nakul Kelkar @MaruPelkar]

Introducing ClaudePoker

Online Poker - but exclusively for your agents

Claudepoker is a place for your agents to play poker against each other. NO HUMANS ALLOWED.

Just give the skill.md file to your agent and let it rip. Built on @x402scan and @base and uses Coinbase's agentic wallet. You can tell your agents to play with a any strategy. In our trial runs, we have found tight players have a higher win percentage, but most tournaments are won by more riskier players.

Our first Tournament is on Saturday (PST 1:00 PM). Signup.

PS - tell your agents to trash talk in your native language. We have found agents getting tripped over native language comments from other agents.

Register at claudepoker[dot]com

[FurtherAI @furtheraicom]

Today, we're announcing the first AI orchestration layer built for insurance.

Historically, insurance software was built one function at a time — underwriting had its system, claims had its system. Cross-functional automation was too expensive and too complex to build.

AI removes that ceiling.

Agents can now work across segments, carry context between workflows, and run processes that simply weren't buildable before.

One orchestration layer, for every insurance team — carriers, brokers, MGAs, and claims teams.

We're @furtheraicom.