Salesforce has just made its most capable software the default rather than the upgrade, and that is a genuinely good week for anyone running revenue on the platform. On 3 September 2026 the company introduced three replacement editions across Agentforce Sales, Agentforce Service and Agentforce Industries: Core at $195 per user per month, Advanced at $395, and Max at $550. Every one of them arrives with Slack, Tableau Next, a Premier Success Plan, enterprise-grade security and a standing allowance of Flex Credits already inside the box.
The instinct with any repackaging is to reach for the calculator, and the commercial reading matters. The more immediately useful exercise is a shorter one. Work out what your organisation can now do that it could not do the week before, then work out how quickly you can put that capability in front of the people who would use it. For a large number of Salesforce customers, the software they will be running by the end of this quarter is materially larger than the software they were running in August, and the seat price did not move to make room for it.
What Salesforce Actually Shipped on 3 September
The three editions replace a structure that had grown difficult to explain. Enterprise, Unlimited and the Agentforce 1 Edition each carried a different mix of AI capability, and most of the interesting functionality sat at the top. Core now carries what Salesforce describes as 70 per cent more value than legacy Enterprise. Advanced carries 50 per cent more than legacy Unlimited. Max is priced identically to the Agentforce 1 Edition it succeeds, with roughly 60 per cent more in it, and existing Agentforce 1 customers can move across at no additional cost.
The Flex Credit allowances are where the practical difference shows up. Core includes 500,000 credits, Advanced one million, and Max 2.75 million. Credits are the unit Salesforce uses to meter agent work, so an allowance of that size is not a token gesture: it is enough capacity to run real agent workloads without a separate purchase order every time a team wants to try something. Pricing on existing legacy contracts is unchanged, so nobody is being pushed anywhere on renewal day, and pricing for the Industry editions follows later this autumn.
The bundling itself is the headline. Slack, which many organisations were buying separately, is now inside the licence. Tableau Next, which moved to role-based pricing rather than pure consumption, comes with the edition. So does the Premier Success Plan, which used to be a percentage uplift on the contract. Individually these are line items. Together they describe a platform that assumes collaboration, analytics and agent execution are part of a CRM rather than adjacent products you assemble around one.
What Is Salesforce Agentforce, and What Now Sits Inside It?
Salesforce Agentforce is Salesforce’s platform for building and running autonomous AI agents that act on CRM data rather than simply describing it. An Agentforce agent can research an account, draft and send outreach, update pipeline records, resolve a service case or hand a conversation to a colleague, all governed by the same permissions, sharing rules and audit trail as any other action in the org. As of the September 2026 editions, Agentforce capability is included in every tier of Agentforce Sales, Agentforce Service and Agentforce Industries rather than reserved for a premium edition, and each tier carries a Flex Credit allowance that funds agent execution. Sales Cloud and Service Cloud are now branded Agentforce Sales and Agentforce Service to reflect that agents are core to the product rather than an add-on to it.
Ninety-Five Per Cent of Users Were Never on the Premium Tier
Here is the number that makes this announcement more interesting than a pricing tweak. Roughly 5 per cent of Sales Cloud and Service Cloud users sat on the higher-end editions where the most advanced AI capability lived. Everything Salesforce has built over the past two years, the agent runtime, the reasoning layer, the newer analytics, was reaching a small slice of its own installed base.
Moving that capability down into the standard editions changes the size of the population that can use it, and it changes it immediately rather than at the pace of an upgrade cycle. The teams that benefit most are not the flagship accounts that already had everything. They are the mid-sized organisations that ran a competent, unglamorous Sales Cloud instance and made a reasonable decision two years ago that the premium tier was not worth the delta. That decision has just been made for them, in their favour.
This is a good position to be in. The capability arrives without a business case, without a procurement cycle and without the internal argument about whether AI is worth a budget line this year. What it needs instead is someone willing to spend a fortnight finding out what is now switched on.
Start With an Entitlement Audit, Not a Renegotiation
The single highest-return action available to most Salesforce customers this month is an entitlement audit: a plain list of what the organisation now holds by default, set against what it currently pays for separately. Nearly every mid-market CRM estate we see carries a small constellation of point tools bought to fill gaps in an older edition. A separate collaboration tool. A reporting layer somebody licensed for one department. A conversation intelligence product that duplicates something now sitting in the platform.
This work is worth doing whatever the outcome, which is what makes it easy to justify. An accurate picture of what your CRM estate contains improves integration planning, sharpens your security review and shortens every future vendor conversation. If it also lets you retire two subscriptions, that is a bonus rather than the premise.
CRM technical debt is the accumulated cost of decisions that made sense at the time but now slow the platform down: redundant fields, overlapping automations, unused customisations, and integrations built to bridge gaps the vendor has since closed. It behaves like financial debt in that it compounds quietly, and the interest is paid in release cycles that take longer than they should. The most effective way to reduce it is not a rewrite but a regular inventory: identify what is genuinely used, retire what duplicates native functionality, and document what remains. A vendor repackaging like this one is an unusually good trigger for that inventory, because the list of things now covered natively has just changed.
How Do You Measure the Return on Capability You Did Not Buy?
Measuring CRM ROI is usually framed around a purchase, which makes bundled capability awkward to assess. The framing that works better is to measure the throughput of the work itself. Pick two or three processes where an agent could plausibly take load, establish what they cost in hours and cycle time today, then measure the same figures ninety days after the agent goes live.
To measure CRM return on investment credibly, set a baseline before any change is made, then track three categories over one or two quarters. First, cycle time on the processes you are automating: how long a quote takes to produce, how long a service case stays open, how long it takes a seller to prepare for a meeting. Second, capacity released: hours per week returned to the team and what those hours were redirected to. Third, revenue-linked outcomes with a clear line of sight to the change, such as conversion at the stage the agent supports, rather than total pipeline. Compare against the pre-change baseline rather than against a vendor benchmark, and keep the measurement window long enough to survive a quarter-end.
The Flex Credit allowance makes this considerably easier than it was. Credits give you a real, meterable unit of agent consumption included in the licence, so you can run a genuine pilot at known cost and compare consumption against outcome. That is a much better starting position than a proof of concept negotiated separately and paid for out of an innovation budget.
How to Improve Adoption When the Tools Arrive Before the Demand
There is a pleasant problem in bundling, which is that capability now arrives ahead of demand for it. Nobody asked for Tableau Next in June. It is in the edition in September. The organisations that get value from this will be the ones that treat it as a small internal launch rather than a footnote in a release note.
Improving CRM adoption rates comes down to sequencing rather than persuasion. Start with one team and one process that is visibly painful, rather than a platform-wide rollout. Configure the capability against how that team actually works, using their language and their stages, not the demo configuration. Put a named owner on it who is a practitioner rather than an administrator. Measure use weekly for the first month and publish the numbers, because visible measurement changes behaviour faster than training does. Then let the first team’s result recruit the second. Adoption follows evidence produced by peers far more reliably than it follows a mandate, and a bundled capability is ideal for this because the cost of a failed first attempt is close to zero.
The practical sequence for the next six weeks looks something like this. Audit what you now hold. Pick one process per business function where an agent or a dashboard would remove real friction. Configure it with the team that will use it. Measure it against the baseline you took before you started. That is four weeks of focused work, and it will tell you more about what this platform can do for your organisation than any amount of evaluation will.
The Sirocco Perspective
We work across Salesforce, HubSpot and Microsoft Dynamics 365, and we like seeing capability of this kind ship as standard. It raises the baseline for everyone. When advanced functionality only reaches the top 5 per cent of an installed base, the market learns very little about what works, because the sample is made up entirely of organisations with the budget and the internal capacity to make almost anything work. Move it into the standard edition and the evidence gets much more useful to the rest of us.
Our advice to clients this month is consistent regardless of platform. Do the entitlement audit before you do anything else, because it is cheap, it is quick, and it changes what the rest of the conversation is about. Then pick one process, not ten, and prove the value where people can see it. We have watched this pattern work across all three major platforms, and the constraint is almost never the technology. It is whether somebody owns the first use case closely enough to make it succeed.
The wider point is that platform boundaries are moving in the customer’s direction. Salesforce has put collaboration, analytics and agent execution inside the licence. Other vendors will respond, because that is how this market has always worked. For buyers, that means more capability per contract and more freedom to choose a platform on how well it fits the business rather than on which add-ons you can afford. That is a good direction of travel, and the organisations that build the habit of auditing, piloting and measuring will be the ones that compound the advantage.
If you would like a second opinion on what your current edition entitles you to and where the quickest wins sit, schedule a consultation with our team. We are happy to walk through it with you.
Get in Touch
If you are working out what your new Salesforce edition actually entitles you to, or which process should be the first one an agent takes on, we would be glad to help you map it.
