What happens when AI agents start negotiating with each other, and who's responsible when they do? Silvio Savarese and Sabastian Niles break down the New AI Trust Architecture: 5 requirements for safe, accountable agent-to-agent communication. sforce.co/4gwgT3p #AgenticAI #AItrust #AIgovernance

The New AI Trust Architecture: 5 Requirements for Agent-to-Agent Communication

5 requirements for trustworthy AI agent-to-agent communication, from Salesforce's Chief Scientist and Chief Legal Officer. Standards, identity, and accountability, explained.

salesforce.com

(1/5) Model cards are nutrition labels for AI. Now they include environmental impact. 🌱 @Salesforce.com is adding standardized energy + carbon metrics: sforce.co/4umu8qm

Measuring AI’s Environmental Impact: How We’re Operationalizing Transparency Through Model Cards

Today, Salesforce is expanding its AI model cards with standardized environmental impact metrics. This update helps customers better understand the energy

sforce.co

(1/8) Can Language Models Remember What They Learn? LLMs learn from feedback. But most post-training is amnesiac: rollout → reward → update → forget. What if you keep the signal? Procedural Memory Distillation (PMD): learning from experience, not just feedback. 🧵

6th Multimodal Algorithmic Reasoning Workshop at #CVPR2026 is Thursday (6/4), 8:55 AM–12:30 PM MDT, Room 601, Colorado Convention Center. 👥 Keynote by Juan Carlos Niebles @jcniebles.bsky.social, organized by Honglu Zhou @hongluzhou.bsky.social 📆 sforce.co/4ueOT7j

Can Language Models Remember What They Learn? Introducing Procedural Memory Distillation (PMD): sforce.co/4dAjQOu PMD turns model attempts into reusable training memory, conditions a self-teacher on it, and distills the guidance into the student's weights.

Can Language Models Remember What They Learn?

Post-training methods (RLVR, On-policy distillation) are Episode-local Language models are getting better at learning from feedback during post-training. In reinforcement learning with verifiable rewa...

sforce.co

📣 1/5 Counterparty Modeling is Not Strategy: The Limits of LLM Negotiators sforce.co/3RTTU7Q New research tests whether LLM agents can turn knowledge of a partner's preferences into better deals. The finding? LLMs can know what the other side wants without bargaining strategically on it.

1/5 RLVR trains LLMs with pass/fail rewards — but every near-miss rollout is wasted. What if models could actually *learn* from their mistakes? New paper: "Learning from Language Feedback via Variational Policy Distillation" Read: sforce.co/4uv2f0k 🧵👇

When AI Becomes Invisible: The Rise of Ambient Intelligence Silvio Savarese on the shift from AI as a tool you go to, to a presence already there: always-on, aware, adaptive, and anticipatory. sforce.co/3Pwo68s #FutureOfAI #EnterpriseAI #AmbientIntelligence

When AI Becomes Invisible: The Rise of Ambient Intelligence

A perspective on the future of enterprise AI that is ambient: context-aware, always on, and invisible to users.

sforce.co

Nvidia has adopted our xLAM function calling dataset in NeMo Gym, their new library for building RL training environments for LLM agents. Check out our function calling dataset: sforce.co/4uZkL0P And see how it works in Nemo Gym: bit.ly/4tuu0pg #AgenticAI #ReinforcementLearning

SFR-VibeTrain: The Agent That Trains Agents A messaging-native agentic control plane for model and agent training. Describe the task, upload data (or don't), and it reasons, launches, analyzes, ablates, and deploys. sforce.co/4dwzYzo #FutureOfAI #EnterpriseAI #AgenticAI

SFR-VibeTrain: The Agent That Trains Agents

What if launching an RL training run felt less like operating a GPU cluster and more like talking to a sharp research engineer in Slack? Training AI models is still strangely artisanal, involving…

sforce.co

A great moment for the lab! 🎉 "Don't Stop Early" earned its spot on the Industry Track for advancing enterprise deep research at scale. Controlled information flow and evidence-aware termination prevent premature stopping and uneven coverage. bit.ly/49fk2zQ #ACL2026

Salesforce AI Research@sfresearch.bsky.social · 4mo ago

(1/7) We have 6 papers accepted to ACL 2026, advancing work across web agent evaluation, LLM reasoning verification, uncertainty quantification, long-context efficiency, and multilingual judge systems. ACL 2026 takes place July 2-7 in San Diego, California. #ACL2026 #FutureOfAI #EnterpriseAI

🎉 3 papers from @salesforce.com AI Research accepted to #ICML2026, spanning Mixture-of-Experts efficiency and multi-agent reasoning + verification. Learn more ↓

Our Infrastructure & DevOps team was featured by Google Cloud. Lavanya Karanam and Avinash Gudagi on how @salesforce.com is solving one of the hardest problems in large-scale AI training: keeping powerful GPUs fully fed. https://bit.ly/4n6HDc3 #FutureOfAI #EnterpriseAI

3 AI trends shaping #EnterpriseAI: simulation environments, ambient intelligence, and agent-to-agent communication. Itai Asseo on our focus on Enterprise General Intelligence (#EGI) in @gregorojstersek.bsky.social's #TDX26 recap. https://bit.ly/4mT4irW #FutureOfAI

3 Key AI Trends and How Salesforce Engineers use AI

Recap from the Salesforce TDX 2026: AI Agents, 3 key AI trends and how engineers use AI.

bit.ly

(1/5) 📊 Don't Stop Early: Scalable Enterprise Deep Research with Controlled Information Flow and Evidence-Aware Termination bit.ly/49fk2zQ

Congratulations to our researches on winning an Outstanding Paper Award at #ICLR2026 for LLMs Get Lost in Multi-Turn Conversation! 🏆 Authors: Philippe Laban, Hiroaki Hayashi, Yingbo Zhou, Jennifer Neville https://bit.ly/3ZoC6T1

Attending #ICLR2026? Meet us at @salesforce.com Booth #203 to hear the latest on how we're evolving enterprise general intelligence. For more, follow our journey at ICLR: https://bit.ly/4vLy9qohttps://bit.ly/4vLy9qo #FutureOfAI #EnterpriseAI

📍 Live from Rio: 21 papers at #ICLR2026, starting today. Our work spans agent architectures, reasoning, evaluation, deep research reliability, and scalable RL, tackling what matters most for #EnterpriseAI. https://sforce.co/4vRgqOs #FutureOfAI

Salesforce AI Research at ICLR 2026

Salesforce AI Research will present 21 accepted papers at ICLR 2026, the Fourteenth International Conference on Learning Representations. The conference runs April 23–27 at the Riocentro Convention an...

sforce.co

Can AI coding assistants maintain effectiveness as codebases scale 100×? LoCoBench-Agent evaluates agents across 10K to 1M token contexts, spanning 8,000 scenarios in 10 programming languages and four difficulty tiers. https://sforce.co/4txCDj4 👥: Jielin Qiu & Huan Wang #EnterpriseAI #AgenticAI

Beyond 100K Tokens: Evaluating AI Agents in Long-Context Software Engineering

As codebases grow to millions of lines of code, can AI agents still understand, reason, and code effectively? LoCoBench-Agent delivers the answer: a comprehensive benchmark for evaluating AI coding as...

sforce.co

The Illusion of Certainty: Decoupling Capability and Calibration in On-Policy Distillation: bit.ly/48iccVY On-policy distillation boosts accuracy but causes severe overconfidence. CaOPD uses a student-grounded empirical target for Pareto-optimal calibration. Code: bit.ly/4cUtCKO

At #TDX26, Itai Asseo @iiitaiii.bsky.social on Enterprise General Intelligence: refining generic LLMs into reliable enterprise agents through AI Foundry and Agentforce Labs, including eVerse, the Learning Engine, and Agent Startup. https://sforce.co/3OrRYCb #FutureOfAI #EnterpriseAI

TDX Live Blog: All the Highlights You Missed

On-the-ground reporting from the must-attend developer event for the Agentic Enterprise

sforce.co

(1/7) We have 6 papers accepted to ACL 2026, advancing work across web agent evaluation, LLM reasoning verification, uncertainty quantification, long-context efficiency, and multilingual judge systems. ACL 2026 takes place July 2-7 in San Diego, California. #ACL2026 #FutureOfAI #EnterpriseAI

That's a wrap on #TDX26! The @salesforce.com AI Research team was on the ground in San Francisco, connecting with the community and sharing the latest from our labs. From research demos to conversations about what's next in #EnterpriseAI, it was great to be part of the energy!

Reliability over raw power. 🔬 @aimagazine.bsky.social features Silvio Savarese on why enterprise AI value comes from integrated systems—not bigger models—and how AI Foundry is building that foundation. https://bit.ly/41z0G4E #SystemLevelAI #AIFoundry

Salesforce AI: Reliability Trumps Raw Model Capability

As AI matures, enterprise success hinges on integrated systems that deliver consistent performance across the most complex professional business workflows

bit.ly

AI Foundry: Turning foundational research into enterprise AI products faster. Silvio Savarese and Itai Asseo in TechFinitive: bit.ly/4ccYmFy #FutureOfAI #EnterpriseAI #AgenticAI

Salesforce launches AI Foundry as it doubles down on three "big bets" to accelerate enterprise AI

Details on Salesforce AI Foundry, launched today to help AI researchers, customers and partners to collaborate in a safe environment

bit.ly

The model wars are over. Enterprise AI success now lives at the system level. Silvio Savarese and Itai Asseo in @technologymag.bsky.social: bit.ly/4ttSBu4 #FutureOfAI #EnterpriseAI #AgenticAI

Salesforce AI Foundry: System Reliability Beats Model Power

The era of the model wars is over, with enterprise AI success occurring at the system level, demanding reliability and full integration over model power

bit.ly

Ambient intelligence is moving from research into live sales and service workflows. Silvio Savarese discusses with @cxtoday.com: bit.ly/47DUhJ1 #FutureOfAI #EnterpriseAI #AgenticAI

Salesforce Brings Ambient Intelligence to Sales Calls

Salesforce showcases ambient intelligence for real-time sales workflows and highlights Agentforce upgrades focused on scale, governance, and enterprise outcomes.

bit.ly

Three agentic AI trends shaping the enterprise through 2027. Silvio Savarese and Itai Asseo discuss AI Foundry with @cio.com: bit.ly/4sg6iMf #FutureOfAI #EnterpriseAI #AgenticAI

Salesforce AI Research identifies trends shaping agentic AI

Simulation environments, agent-to-agent ecosystems, and ambient intelligence will be at the heart of the Salesforce product roadmap through AI Foundry initiative.

bit.ly

One demo. Reliable replay. No cloud calls. GPA turns a single recorded workflow into deterministic desktop automation, entirely on-device. 🔎 Explore GPA: bit.ly/48r7Onp 📖 Read the blog: sforce.co/4sYdhu8 #EnterpriseAI #GUIAutomation

Why #EnterpriseAI demands a shift from models to systems. Itai Asseo discusses AI Foundry with @diginomica.com: bit.ly/4scZwGT #FutureOfAI #AgenticAI

The big bets are on as Salesforce pitches the need for enterprise transition from model to system level AI

Itai Asseo, VP of Salesforce AI Research, explains some new enterprise realities on the way.

bit.ly