Salesforce's · MindBlaze

A $1.6 billion ceiling is a strange number to find on an “unlimited” contract. That is the figure on the AELA the Department of Veterans Affairs signed with Salesforce in July 2026.

The word next to it, Agentic Enterprise License Agreement, is the one your Salesforce account team is probably using in your renewal conversation right now.

If you are being quoted an AELA, the decision is simple to state and hard to get right. Commit to a fixed fee for two or three years of unlimited agent usage, or stay on the meter and pay for what you consume.

My advice before you spend a week modelling it: the year-one math is not where this deal is won or lost. The renewal clause is. Negotiate that first, and the headline price second.

What an AELA is

An AELA is a fixed-price, fixed-term agreement. It gives you unlimited access to a defined set of Salesforce’s consumption-metered products, with no usage-based billing and no overage on top.

The products most often named in coverage of it are Agentforce, Data 360, MuleSoft and Slack. The term runs two to three years. The price is negotiated per customer and is not published anywhere.

So unlike the Core, Advanced and Max editions, there is no rate card to check your quote against. That matters more than it looks.

Every other way to buy Agentforce has a public anchor. Flex Credits list at $500 per 100,000, a standard action is 20 of them, a conversation is $2. Data 360 credits carry their own rate card with a multiplier per operation.

The AELA deletes all of that and replaces it with one number you cannot benchmark against a list price. You are negotiating in the dark unless you build the consumption model yourself.

The reframe Salesforce is selling, and Forrester is echoing, is that agents are economic actors that generate value rather than metered features that incur cost. If an agent closes a case, the argument goes, you should no more count its actions than count keystrokes.

A clean story. Whether it is your story depends entirely on how much you are going to run.

The problem the AELA is built to solve

The meter has a real cost beyond the invoice, and anyone who has run Agentforce at volume has felt it.

Consumption pricing turns every design decision into a budget decision. Add a grounding step and you add Data 360 query credits. Let a segment refresh hourly instead of daily and you multiply its cost. Give an agent one more action in its plan and you have changed the per-conversation rate.

The result is a tax on experimentation. Teams under-build agents to stay under a credit forecast, and finance spends the quarter reconciling a bill nobody can predict to two significant figures.

The AELA removes that friction in one stroke. Inside the covered scope, usage is free at the margin, so the Data 360 pipeline you were rationing and the extra agent you were holding back both stop showing up on a meter.

For an organisation that has already decided agents are core infrastructure and intends to scale them hard, that is the whole pitch, and it is a good one.

Adecco signed a multi-year agreement through 2027 for unlimited global access to Agentforce 360. Its stated goal is to have more than half its revenue powered by agentic AI by the end of 2026. You do not sign that deal to save money on a pilot. You sign it to stop thinking about the meter while you rebuild the business around agents.

Where unlimited beats the meter, and where it doesn’t

The crossover is about conviction and volume, not cleverness.

An AELA wins when three things are true at once. You have proven the use cases, so a flat fee is not paying to find out whether agents work. You intend to scale faster than you could comfortably forecast, so predictability is worth a premium. And the covered products are central to your roadmap, so you will use enough of the bundle to justify paying for all of it.

A staffing company rebuilding recruitment around agents, or a government department modernising a national contact centre, fits that shape.

The meter wins when any of those is shaky. In the proving phase, a flat multi-year commitment buys certainty you do not have yet, and Flex Credit optimization plus Data 360 credit discipline will carry you further for less.

If your usage is lumpy or seasonal, the meter bills the troughs at zero while the AELA bills them at full price. And if you will lean on one product and ignore the rest, you are buying MuleSoft and Slack to get Agentforce, which is rarely the cheapest route to Agentforce.

There is a quieter version of the same trap on the edition side. The new Core, Advanced and Max tiers already bundle a large pool of Flex Credits per user.

Before you conclude you have outgrown consumption, check whether the credits you already pay for cover the volume you are running. A lot of “we need unlimited” conversations end when someone opens Digital Wallet and sees the included pool sitting half-used.

The renewal clause is the whole deal

The part that should decide your answer is not the price.

Gartner’s Hannah Decker has warned that AELAs may be converted into defined-quantity contracts at the end of the term. Read that slowly.

The risk is not that year-one unlimited is a bad deal. The risk is the sequence after it. You spend two years scaling usage with the meter switched off. Consumption triples because nothing was stopping it. At renewal, the “unlimited” frame is replaced by a quantity based on everything you now consume, priced without the bargaining power you held the first time.

You optimised for throughput because throughput was free. The renewal prices that throughput back in.

Salesforce disputes the framing. The company told The Register it would not move away from capped agreements and that renewals would remain flexible, citing shifting AI compute costs as the reason it wants room to move.

Both things can be true. The first term is a good deal, and the renewal is where the vendor expects to make the economics work. That is not cynicism. It is how enterprise software has always been sold.

So the clauses that protect you govern the end of the agreement, not the beginning. Before you sign, get these in writing:

A renewal cap: a maximum price increase, in percent, if the deal converts to a defined-quantity contract at term end.
Exit terms spelled out: what you owe, what you keep, and what happens to agents and data if you walk.
Usage reporting you can audit during the term, not a number Salesforce reconstructs at renewal to set your next baseline.
A written definition of “unlimited”, naming which products are in scope and which usage types are carved out.

Decker’s own summary is the one to take into the room: understand the exit terms before you sign, track your usage, and negotiate the renewal caps now, while the deal is still unsigned and the bargaining power is yours.

Read the word “unlimited” like a lawyer

“Unlimited” in an AELA means unlimited within a defined scope. The scope is where the real terms live.

The VA contract shows how a headline can hide its own structure. The press release calls it a three-year, $1.6 billion agreement, but the footnote describes a one-year contract with two optional one-year renewals under that ceiling.

Those are not the same commitment. One is three years of certainty. The other is one year with the vendor and the customer each holding options. If a public, heavily scrutinised federal contract carries that much ambiguity in its term, assume your mid-market quote carries more.

Read the scope with the suspicion you would bring to any ceiling:

Which products are covered, and at which editions.
Which usage types are excluded from “unlimited”, because the carve-outs are where consumption charges sneak back in.
Whether net-new products Salesforce ships during your term fall inside the agreement or get sold to you again.
Whether the FedRAMP, HIPAA or residency posture you need is in scope or a separate line.

The headline number tells you almost nothing until you have answered those.

How to decide

Do the consumption model anyway, even though the AELA erases the meter. You cannot judge a flat fee without the usage curve it is replacing.

Build a two- to three-year forecast on the meter: agent volume, actions per conversation, Data 360 operations, segment refresh frequency, the lot. Our Agentforce ROI calculator gets you a defensible deflection and savings number to anchor it, and the Flex Credits post has the per-action math.

That curve is your walk-away price. If the AELA sits below it across the term and the renewal cap holds, the deal is real. If it sits above, you are paying a premium for predictability. A legitimate thing to buy, but it should be a conscious one.

Then weigh what the spreadsheet cannot answer. Whether you have proven the use cases, or are buying unlimited to avoid deciding what to build. Whether you will use enough of the bundle, or are paying for four products to get one. Whether you can live with the renewal, having modelled the defined-quantity version Gartner is warning about alongside the unlimited version Salesforce is selling.

The companies signing AELAs today are not doing it to save money this year. They are buying the freedom to stop rationing, because they have already decided agents are how the business runs.

If that is you, the AELA is the right instrument, and the renewal cap is the clause you fight for. If it is not yet you, the meter is still cheaper, and getting the data foundation and the use cases right first is what earns you the standing to sign a good one later.