Companies make a mistake when they discuss AI strategy as though agentic AI were a single capability to deploy uniformly across an organization, according to Marc Escobosa, VP, Salesforce Futures. AI excels at some types of work and is unreliable at others — sometimes dangerously so. The agentic AI that suffices for one process might be a poor fit in another. Fortunately, the agentic unevenness is predictable. Two questions explain where an agent is likely to excel or struggle in an enterprise: ➡️ Has the LLM been trained on enough relevant data to be able to work its magic? ➡️ Can a human verify the answer to ensure it’s correct? Answering those two questions reveals four broad categories of tasks, each calling for a different role for humans: 1️⃣ High verification and abundant context: Let the machine do it — but check its work 2️⃣ High verification and missing context: AI could do this if only it knew what we know 3️⃣ Low verification and abundant context: AI can give a pretty good answer — and that’s the problem 4️⃣ Low verification and missing context: Nobody knows, somebody still has to decide For leaders, the takeaway is not to ask simply, “Where should we deploy AI?” but “What kind of work is this, and what should humans do here?” https://sforce.co/3Wt8YvD
Salesforce News & Insights
Technology, Information and Internet
San Francisco, CA 128,450 followers
News and insights from Salesforce. For media inquiries contact pr@salesforce.com.
About us
News and behind-the-scenes insights from Salesforce. For media inquiries contact pr@salesforce.com.
- Website
-
http://www.salesforce.com/news
External link for Salesforce News & Insights
- Industry
- Technology, Information and Internet
- Company size
- 10,001+ employees
- Headquarters
- San Francisco, CA
- Founded
- 1999
Updates
-
Half of all Slack messages now come from apps and agents, which have moved into the channels where work already happens, taking action alongside the teams they support. Consider Vercel, a leader in agentic infrastructure, which runs its entire company in Slack. Its inbound sales agent handles lead qualification, research, and outreach drafting, so a single rep can do the work that would otherwise take a team of 10. Meanwhile, at Anthropic, 65% of all code changes now ship through Slack. Anthropic is one of the nine of the 10 top frontier AI companies that run their operations on the platform. In this Salesforce Scoop, get the full picture of Slack by the numbers.
-
Ask Slackbot: I’m prepping for a customer discovery call. Do I accept all of their opinions or challenge their assumptions? Find out what advice Slackbot has for this Forward Deployed Engineer. Ask Slackbot is a monthly column that answers reader questions about work in the age of agentic AI. Answered by the Slackbot of someone in a different Salesforce role each month, it’s grounded in that employee’s real Slack conversations, customer implementations, research, and relevant documents. It can search messages, read files, synthesize patterns, and connect insights — all while respecting permissions and data privacy within the Salesforce Trust Layer. It offers some broad life and work lessons, too. Got a question particular to your job function? Email AskSlackbotColumn@salesforce.com. https://sforce.co/48wEsUm
Ask (a Forward Deployed Engineer’s) Slackbot
-
Picture a race car flying down a country highway, only to find itself stuck behind a tractor. The car’s engine is spectacular. But it can’t go any faster than the tractor. That’s what many companies encounter, often without realizing it, when they build with AI. They drop an agent into one step of a workflow, and everything before and after it still moves at the speed of the organization around it. This is why you hear that 95% of companies aren’t getting value from AI. The value gets trapped at every human-scale step the agent has to wait on. The fastest way to become an agentic organization is to create a secure and governed open environment where everyone has the opportunity to build. At Salesforce, 74% of employees have used or interacted with an agentic coding tool like Claude Code or Codex at least once, and usage among nontechnical employees is up to 66%, more than triple the usage in Q1. Most of them can’t write code, but they’re willing to experiment with telling a machine what they need. But if the right methodology and infrastructure are not in place, everyone building everything everywhere can be an insecure nightmare. In a new byline, Joe Inzerillo, President, Enterprise and AI Technology at Salesforce, shares how the company is addressing this - by developing Rapid Innovation Squads (RIS). https://lnkd.in/ePd4VUmB
-
Any company can get a smarter model. What none of them can rent is their own context: their customers, relationships, permissions, and rules. Today Salesforce introduced Agentforce Coworker, an AI teammate that puts the full context and capabilities of Salesforce to work for every employee. Built on AIforce, the live interface for Salesforce, Coworker is a trusted teammate from day one. "Agentforce Coworker has been a game changer for Fulton Bank. We're already seeing substantial daily usage and real examples of how Coworker is helping teammates find information faster, prepare more effectively, and move work forward,." said Brandon Bernard, Senior Vice President/Sales Technology Strategist, Fulton Bank. https://lnkd.in/gRP4qic5
-
More than a decade ago an engineering team built Slack because they needed a way to coordinate code, log alerts, and talk to each other all in one place. Now, as AI agents move from pilots into live business production, they’ve entered the channels where people already collaborate: Through open standards like MCP, agents can pull context from Slack across platforms like GitHub, Salesforce, and Jira to take direct action in real time. Slack now processes 127 million of these automated actions, called MCP tool calls, every week, a figure that is doubling quarter over quarter. Six years after Slack joined Salesforce, Slack’s origin in humans and software working side by side has become the execution layer for humans and agents working together. "Some tasks feel like stopping at a traffic light every few minutes of a journey. The journey shouldn’t take that long, but you get stuck at every red light. Slack helps us remove those obstacles that stop the flow of work," said Colin Strother, Executive Vice President, Rochester Electronics. https://lnkd.in/es2BgawQ
-
Salesforce News & Insights reposted this
As more and more businesses rely on AI, workers are experiencing a bottleneck: interface isolation. According to a recent Salesforce survey of 250 U.S. executives: 🔹 68% of respondents see AI insights go unshared 🔹 66% say people unknowingly use AI to solve the same problem as their colleagues 🔹 59% observe work stall because it's trapped in one person’s AI tool What’s more, because only 15% of executives say their AI is embedded seamlessly across multiple systems and interfaces, employees waste a whopping 22 business days a year toggling between AI and other windows. Dive into the results of this new survey: https://lnkd.in/e7_8JWQP
-
Salesforce News & Insights reposted this
Agentic AI is already excellent at some tasks and getting better at more. So should we just extend the line and assume it'll be good at everything? I don't think so. Some limits will fall. Rubrics and preference data are turning ”subjective“ work into something models can learn to check. Other limits won't, at least not soon: work where there's no way to verify the answer before the decision window closes and you have to act. Knowing which is which should shape your AI strategy. https://lnkd.in/gVvzvRPw #AI #AgenticAI #EnterpriseAI
-
Everyone’s got an AI assistant. But those assistants all live in silos. According to a new Salesforce survey, 59% of executives see work stall because it’s trapped in one person’s AI tool; 66% watch colleagues unknowingly solve the same problem independently; and 68% of executives see AI insights go unshared. Plus, employees lose an average of 3.6 hours a week toggling between AI and other apps. That’s roughly 22 business days a year — about a month of work — gone. In this Salesforce Scoop, learn about how executives perceive AI usage at their companies, and why they want AI that works across the whole team.
-
AI is still jagged. Enterprise strategy should be, too. Companies make a mistake when they discuss AI strategy as though agentic AI were a single capability to deploy uniformly across an organization. It isn’t. The agentic AI that suffices for one process might be a poor fit in another. Sometimes, AI agents can perform so reliably that a person’s main job is to delegate the tasks to AI and audit the results. Elsewhere, the AI might have little to no relevant training. For leaders, the takeaway is not to ask simply, “Where should we deploy AI?” but “What kind of work is this, and what should humans do here?” Read more from Marc Escobosa, VP, Salesforce Futures about why even within an organization, there is no single AI strategy. https://lnkd.in/es8Br4fE
