Straight answers.
The Agentforce and Claude questions you're being asked, answered without a sales call.
Should we use Agentforce or Claude?
The dividing line is people or records. When a person wants help with their own work, such as preparing for a call or reviewing their deals, that suits Claude. Salesforce in Claude works as the signed-in user and, by default, proposes changes for them to approve. When a job has to happen on every record without anyone asking, such as routing each new lead or checking for duplicates every night, it needs an agent inside Salesforce, which is Agentforce. Some teams need both: Claude for their people and an agent for their records. We decide it with you job by job. If Salesforce already does it, we'll say so.
Do we need Data Cloud?
It depends on the job. The Readiness Assessment checks whether you already have Data Cloud, what your edition includes, and whether any use case on your shortlist depends on it. If one does, the report says so before you buy anything. For what Salesforce includes and charges, see Salesforce's Agentforce pricing page.
What do we already own?
We find out before we quote anything. The Readiness Assessment reviews your edition, Agentforce entitlements and Flex Credits, Data Cloud, Slack, your Sales and Service add-ons, and the AI tools your teams already use or pay for. If something you already own does the job, we tell you in writing.
Does this work on Enterprise, Unlimited, Core, Advanced or Max?
What each edition includes is set by Salesforce and can change, so we don't restate it here. For the current line-up, see Salesforce's Agentforce pricing page. We check your edition and entitlements in the Readiness Assessment and make every recommendation against what your org actually has. Salesforce in Claude has its own requirements while it is in beta, and we confirm them against Salesforce's current terms before we scope a rollout.
What do Flex Credits cost, and can we cap them?
Flex Credits are priced and billed by Salesforce, so we don't quote a rate. See Salesforce's Agentforce pricing page. Whether and how you can cap usage depends on Salesforce's current terms and your contract with them, so we check it for your org during the assessment. Our prices never include Salesforce licences, Agentforce entitlements or Flex Credits. Every recommendation shows its cost three ways: what you would pay Salesforce, what you would pay anyone else (such as Claude seats) and what you would pay us.
Where does our data go, and what does Quebec's Law 25 ask of us?
Our Agentforce agents run inside your Salesforce org. Your metadata, prompts and approval log stay in your org, and model calls go through Salesforce's Einstein Trust Layer in the region your org settings allow. There is no Arkeforce server. Salesforce in Claude works differently: Claude runs in Anthropic's cloud and reaches your Salesforce as the signed-in user, so we review what each user can see before anyone signs in. During an assessment, metadata and aggregate figures pass through our tooling. We sample only what's needed and keep no client data after the readout. Your AI and data-use terms are confirmed before any tool connects to your org. Working this way supports your Law 25 obligations. Confirm the specifics with your counsel.
Can you build an agent that isn't in the catalogue?
Yes. When no catalogue agent, Salesforce feature or Claude skill covers a job, we can build a custom agent for it, on Agentforce, on Claude or on another model. These are often jobs that span several systems. We scope a custom agent after the Readiness Assessment and agree a fixed price before we start.
Who does the work?
Our engineers. The team that assesses is the team that builds: the engineers on your first call review your org, write the fit report and build in your sandbox. We work faster with AI assistance, and every change is reviewed and tested by our engineers before it reaches production.