The questions we get asked.

That it runs in pre-production, every unit test green. Done means proven on the real deliverable: the crew has run the integration, regression and acceptance tests on your Salesforce, and the round ended with no deviation.

That round is signed by our expert and validated by you. Only then does the deliverable go to production, with your approval.

A Product Owner approach, not a developer approach: you define the need, an IA-Pros Salesforce expert is its Product Owner, and we give that Product Owner a team of agents.

A production line, not an assistant. The factory delivers a module that is ready, unit tests passed, documentation up to date. The crew confronts it with your real Salesforce. Every deviation goes back to the factory, as many rounds as it takes, and every round is signed by our expert and validated by you. Whoever proves the code never writes it.

With tests that run. The code the factory produces is a standard Salesforce project, readable, with its unit tests. The crew, other agents on other models, runs on your environment the tests that prove it does what the need requires and that existing behaviour is preserved.

A deliverable is always validated by people other than those who wrote it: tests that run, our expert who signs, and you who validate.

The factory runs on several providers. A change of model happens at IA-Pros, in our clean room, and the crew keeps proving the deliverable on your environment, as it does every round.

Your deliverable, for its part, is a standard Salesforce project: code, tests and documentation that any Salesforce developer can read, and that run on your Salesforce with the vendor’s tools alone.

You, with your consultants or your advisory firm, and with us if you wish. Design stays with you and your advisers, at its own pace.

Our Salesforce specialists start from your specifications, three lines or a hundred pages, whatever their stage of maturity: they challenge them with you and estimate them with the factory, which builds a mock-up from the pre-sales stage. What our agents speed up is everything that follows: estimate, mock-up, code, tests, documentation.

The scope sets the duration, and the first review gives you an estimate, with costed options.

What you can expect: a project live three to six times faster than with a conventional team, because our production line is already running and you no longer wait for a team to free up. On your side, the workload goes down too: you approve, you do not manage a team.

The principle stays that of a fixed price: a change of scope is a quote. It is priced quickly, because the factory estimates and mocks up continuously, and it is signed separately.

What changes with an agentic production line is the flexibility inside the scope. Every round confronts the deliverable with your need, every deviation goes back to the factory, and the adjustments you ask for when validating a round enter the loop, as many rounds as it takes. A conventional team counts that flexibility in days; we count it in rounds.

That is true of most orgs that have served well. The first review starts there: an assessment of what you have and where improvements are possible.

Then, depending on what you decide: an audit of your org with a priced plan to set it straight, your legacy automations migrated to Flow with a written record, or a project that builds on the existing setup as it is. The crew surveys your existing setup before every round, and the regression tests prove that a change preserves behaviour.

No. The AI is not in the target platform, it is in the method: our agents already build, test and document on your SFRA, as it is, with no migration.

A move to Storefront Next is still a good decision when it rests on facts measured on your storefront: performance that has hit its ceiling, features on your roadmap that only Next provides, hosting on Managed Runtime. We measure it with you and we price it; if the measurements do not justify it, we tell you. When it is decided, it happens page by page, with no big bang.

A defect in the deliverable after go-live: a deviation between what runs in production and the scope you validated. We fix it at no charge, through the same loop, during the warranty period. Its length is set in the contract.

A change in the need, for its part, is priced separately.

That is the plan. You take over a project you already know: a standard Salesforce project, written with the vendor’s official tools, which any Salesforce developer can read; regression tests that run on your environment and prove that a change preserves behaviour; documentation produced every round, from the first one.

You validate every round signed by our expert, so the knowledge stays with you, and our agents stay at IA-Pros. Your team takes over whenever you decide, with the standard tools alone. Between hiring a team and a year-round outsourcing contract, we are the third way: in-house-grade results, the code in your hands, no hiring.

Only the crew, our delivery point, works on your sandboxes, through Salesforce’s official APIs: an API client you create in your Account Manager, with the permissions you set, and which you can revoke at any time. The factory, for its part, works at IA-Pros, in our clean room, on our own environments. What enters your org is the deliverable: a standard Salesforce project.

Deployment of the Salesforce work to production requires your approval and uses your repositories and your pipeline.

The factory receives what it needs to produce: your specifications or statement of need, your design system, your existing code. It works at IA-Pros, on our own environments, and your customer data is not handed to it.

The crew works on your sandboxes, through the official APIs, with the permissions you set, preferably on test data sets; if a sandbox holds a copy of your production, that is your decision and you grant the access. In every case, the models read your data through APIs, for the task and for the task only, and none of it is used to train a model.

What already exists in our method, and what you receive with every deliverable: documentation produced every round; the tests that prove the code, delivered with it; the count of deviations and their fixes, kept by the crew every round; our expert’s signature and your validation, round by round; and, for an Agentforce agent, the log of each of its actions and, when the journey requires it, its scenario written in Agent Script.

Your data stays your data and is never used to train any model. How your use is classified under the regulation is your call, with your advisers; we hand you the records.