April 20, 2026

State of Agentforce for Sales

Chris Fezza, CEO

Agentforce has gone through more changes in eighteen months than most Salesforce products see in five years. Pricing models have changed three times. The product roadmap now spans prospecting, pipeline management, account research, quoting, and coaching.

‍

With so much coming out, it can be hard for RevOps leaders at mid-market companies to separate fantasy from reality. There's real capability here, but also real complexity in figuring out what's ready for production, what it actually costs, and where to start. And if Agentforce is even the best solution for the job!

‍

This piece covers the three things I think mid-market RevOps leaders should understand right now:

‍

What Agentforce sales capabilities are actually available,
How to use the Agentic Maturity Model to figure out where to start
and how pricing has evolved into something workable.

‍

What's actually available for sales teams now?

‍

Salesforce Agentforce now covers the full sales cycle for teams, from an SDR agent that handles outreach and meeting booking, to pipeline monitoring, account research, and quoting. A Prospecting Centre scores accounts by fit and intent, and the recent Bluebirds acquisition will add even richer signals. Reps also get a Sales Coach agent for ongoing skill development.

‍

Here’s a breakdown of all the features that are currently available or coming very soon:

‍

Feature	What it does	Availability	Best for	My take
SDR / Lead Nurturing	Sends personalized outreach, follows up, answers questions, and books meetings. Works for contacts and person accounts, not just leads.	Available Now	Inbound lead programs; top-of-funnel at scale	This is the most mature agent. I'd start here.
Prospecting Center	Scores accounts on Fit, Engagement, and Intent using CRM + external signals. Gives reps a prioritized target list.	Available Now	ICP-matched account prioritization	*Requires Data Cloud investment first.
Pipeline Management	Flags stalled deals, recommends next actions, and surfaces insights in Slack.	Available Now	Managers who want deal health visibility without extra meetings	Useful, but teams without Slack adoption will see limited value.
Account Research / Meeting Prep	Pulls account intelligence from Salesforce, third-party sources, and call history to generate summaries and talking points.	Available Now	AEs with high meeting volume	Underrated. The time savings here are immediate.
Bluebirds Integration	Combines LinkedIn signals, CRM data, and de-anonymized web traffic to build prioritized prospect lists with outreach sequences.	Coming Soon	Outbound-heavy teams	Watch this space. Promising but not yet native to Agentforce.
Sales Coach	Analyzes pitches and role-play sessions. Delivers feedback using CRM data.	Available but Early	Onboarding and skill development	Still maturing. Useful for structured programs, not yet plug-and-play.
Quoting Agent	Assists with quote generation and deal configuration inside Revenue Cloud.	Available Now*	Teams already on Revenue Cloud	*Only relevant if Revenue Cloud is in your stack.

‍

What this means for mid-market teams: The SDR / Lead Nurturing agent and Prospecting agent are the strongest starting points for sales teams. They address the top-of-funnel problem that affects every scaling company: reps spending too much time on manual outreach and research, and not enough time in conversations with qualified buyers.

‍

A practical first move is to pilot the Lead Nurturing agent on a defined segment (inbound leads from a specific campaign, or a set of ICP-matching accounts) and measure the results against your current manual cadences. This gives you real data on agent performance and consumption costs before expanding.

‍

Where should I start using Agentforce for my sales team?

‍

In 2025, Salesforce released the Agentic Maturity Model, a four-level framework that answers exactly that. It was designed for CIOs, but I find it most useful as an honest assessment tool for RevOps leaders.

‍

Here's how it maps to sales use cases:

‍

‍Level 1: Information Retrieval. Agents retrieve data and make recommendations. In a sales context, this looks like an agent pulling account history or product documentation to suggest next steps on a support ticket or sales call. The agent assists the human; it doesn't act independently. Most organizations deploying Agentforce today are here or approaching this level.

‍

Level 2: Simple Orchestration (Single Domain). Agents autonomously execute tasks within a single system. For sales, this is where the SDR agent lives: sending outreach emails, following up on cold leads, answering prospect questions, and booking meetings, all within Sales Cloud. The agent acts, but within defined guardrails and a single domain.

‍

Level 3: Complex Orchestration (Multi-Domain). Agents work across multiple systems and data sources. Think of an agent that handles a deal from prospecting through quoting to order creation, pulling data from Sales Cloud, Revenue Cloud, Snowflake, and your ERP. This level requires harmonized data across systems and a robust governance framework. A small number of enterprises are running Level 3 agents in production today.

‍

Level 4: Multi-Agent Orchestration. Multiple agents collaborate across disparate tech stacks. Salesforce is transparent that no one has reached this level yet. It's the forward-looking vision for the platform.

‍

What this means for mid-market RevOps teams: Even though this is targeted at CIOs, I think the maturity model is a useful assessment tool for RevOps leaders too. Most mid-market companies should be targeting Level 1–2 for their first Agentforce investments.

‍

The prerequisite work (clean data, well-defined processes, clear governance) matters as much as the AI configuration itself.

‍

Before I'd recommend any client stand up an Agentforce agent, I want to see three things in place: clean and well-structured CRM data, documented lead qualification and handoff logic, and a team member who owns agent monitoring and iteration. The technology can be stood up in weeks. Getting it to produce trustworthy results takes longer.

‍

Pricing options (and sometimes confusion?!)

‍

When Agentforce launched in September 2024, pricing was straightforward on paper: $2 per conversation. In practice, it caused immediate confusion. A single-turn question and a complex multi-step interaction cost the same. For teams trying to forecast AI spend, the model was a black box.

‍

By May 2025, Salesforce pivoted to Flex Credits, a consumption-based model at $0.10 per action (20 Flex Credits per action, sold in blocks of 100,000 for $500). Each discrete task an agent performs (updating a record, summarizing a case, answering a product inquiry) draws from a shared credit pool. The Digital Wallet gives you real-time visibility into consumption across orgs and agents.

‍

Then came the per-user options. Agentforce add-ons at $125/user/month give licensed users unmetered employee-facing agent usage within Salesforce. For organizations that want the full bundle, including base license, unmetered AI, Data Cloud access, and a million Flex Credits annually, Agentforce 1 Editions start at $550/user/month.

‍

Salesforce now offers three payment structures:

‍

Model	How it works	Best for
Flex Credits Pay-as-you-go	$0.10/action (20 Flex Credits per action). Sold in blocks of 100,000 for $500. Digital Wallet gives real-time consumption visibility.	Teams piloting 1–2 use cases who want to measure actual usage before committing.
Per-User Add-on	$125/user/month. Unmetered employee-facing agent usage within Salesforce for licensed users.	Teams with high, predictable agent usage across a defined group of reps.
Agentforce 1 Edition	$550/user/month. Full bundle: base license, unmetered AI, Data Cloud access, 1M Flex Credits/year.	Orgs ready to go all-in with Data Cloud already in their stack.

‍

My recommendation: The original $2/conversation model priced out experimentation. The current structure is more workable. Start with Flex Credits on a pay-as-you-go basis to pilot one or two use cases, measure actual consumption, and then decide whether per-user licensing makes sense at your volume. Don't overbuy upfront. The pricing is designed to let you scale into it.

‍

What actually separates the teams that get ROI with Agentforce?

‍

We work with mid-market sales teams on Salesforce every day, and here's what I see: the companies that will get the most out of Agentforce are the ones that treat it as a RevOps initiative, not a technology project.

‍

That means having your ICP criteria documented before you configure the agent. It means knowing your lead qualification logic and handoff points between SDR and AE, not just generally, but specifically enough to put it in a prompt. It means owning the data quality problem before it becomes the agent's problem.

‍

Agentforce won't fix a broken process; it'll run a broken process faster.

‍

Agentforce is real, it's production-ready for specific use cases, and the economics now work for mid-market companies. The question is whether your org is ready to support it.

‍

If you're evaluating Agentforce for your sales team, we're happy to walk you through the readiness assessment. Get in touch for a free consultation.