# Questions before your first AI workflow. > How to start with Genaima's AI autopilot, what to prepare and which decisions need your team. Specific connections and controls are confirmed for the workflow you choose. ## Data and security ### Where will our information be processed? That depends on the agreed design. Review application hosting, file storage, retrieval indexes, model endpoints, logs and support access together. The [security page](https://genaima.ai/security) explains the data-flow questions to settle before connecting business records. ### Is data used for model training? Confirm the selected provider's terms and account settings, together with the project's data-use agreement. Retention and training are separate questions, and requirements must be checked for the actual service rather than inferred from its brand. ### How do you keep different clients' records separate? Client boundaries need to be enforced in access, retrieval, storage and the review interface. The implementation should test an attempt to retrieve another client's material, not rely only on an instruction in the prompt. ## Approvals and control ### Does a workflow have to send or update anything? No. A pilot can prepare a report, answer or draft for a person to review. Sending messages and changing business records are additional capabilities with their own permissions and acceptance checks. ### How should approval work? The reviewer sees the exact output or proposed change, its evidence and the intended destination. The action path checks the reviewer's authority and handles rejected, expired or repeated requests. The control belongs in the application, not only in a model instruction. ### Can we inspect what happened? We agree the run record needed for review and troubleshooting, including source references, output versions, decisions and action results where applicable. Its access and retention are part of the scope; a universal live console is not assumed. ## Models and vendors ### Which models and tools do you use? The [tech stack](https://genaima.ai/tech-stack) covers models, agent tooling, automation, cloud, data and web frameworks. Selection depends on the task, data requirements and evaluation. A tool listed there is not necessarily used in every project. ### Can we use a model account we already have? We assess whether that account provides a suitable interface, permissions and terms for the workflow. A consumer subscription does not automatically provide an application API or the handling required for business records. ### Can models be changed later? Potentially, but it requires compatibility and quality checks. Prompts, tool use and outputs can behave differently between models. Portability should be considered in the design rather than promised as a configuration-only switch. ## Deployment ### Can the workflow run in our environment? Tell us the hosting and processing constraints first. We assess the full set of services, including model calls and logging, before confirming a supported arrangement. Private or on-site processing is not assumed to be feasible for every scope. ### Do we need to replace our ERP, CRM or POS? We begin with the current systems and their available features, exports and interfaces. A useful workflow may work alongside them, but the required records still need to be accessible and consistent. ### What happens after the pilot? Review the evidence against the acceptance criteria, then agree whether to continue, change or stop. Additional sources, users and actions need checks before a wider rollout, along with clear operating and support responsibilities. ## Working together ### Can we start with one task? Yes. A bounded task makes it easier to agree a useful output and compare it with the current process. The [workflow-selection playbook](https://genaima.ai/resources/workflow-selection-playbook) helps identify a candidate with available records and a named reviewer. ### How long will it take? Timing depends on source access, data preparation, integrations, required controls and review availability. We agree the sequence after inspecting a representative example rather than promising a fixed rollout from a website enquiry. ## Integrations ### Are all the homepage tools already connected? No. The logos show the technology stack we work with. Access to your accounts and the particular records a workflow needs must be verified during scoping. ### Can we include WhatsApp or spreadsheets? Spreadsheets can be assessed as structured inputs. Messages require an authorised way to provide the relevant material and a reliable link to the account or task. Automatic access to personal or group conversations is not assumed. ## Getting started ### What should we send first? Describe the recurring task, its current steps and the output you want. A redacted example and the name of the person who reviews the work help us assess a first scope. Use the [contact page](https://genaima.ai/contact) to begin. ### Are the demos real customer results? The [demos](https://genaima.ai/#demos) use illustrative or synthetic data to show interfaces and workflows. The [selected work](https://genaima.ai/work) pages describe published project scopes separately. Demo figures should not be treated as verified customer outcomes. ### How will we know the workflow helped? Measure preparation time, review time, corrections and completion against a baseline. The [pilot measurement template](https://genaima.ai/resources/pilot-measurement-template) helps keep the comparison consistent and separate task improvements from wider business outcomes. - [Launch your autopilot](https://genaima.ai/contact) — Tell us which workflow you want to improve and any requirements this page has not covered. ## Related - [Security & governance](https://genaima.ai/security) - [How a workflow runs](https://genaima.ai/how-it-works) - [Implementation partner](https://genaima.ai/partner) - [Questions to ask AI agent vendors](https://genaima.ai/insights/questions-to-ask-ai-agent-vendors) --- - Canonical page: https://genaima.ai/faq - More: [Home](https://genaima.ai/) · [About Genaima](https://genaima.ai/about) · [Contact Genaima](https://genaima.ai/contact) · [How a workflow runs](https://genaima.ai/how-it-works) · [The AI operating layer](https://genaima.ai/ai-operating-layer) · [Security & governance](https://genaima.ai/security) · [FAQ](https://genaima.ai/faq) · [Implementation partner](https://genaima.ai/partner) · [Solutions](https://genaima.ai/solutions) · [Industries](https://genaima.ai/industries) · [Compare](https://genaima.ai/compare) · [Insights](https://genaima.ai/insights) · [Glossary](https://genaima.ai/glossary) · [Tech stack & tools](https://genaima.ai/tech-stack) · [Resources](https://genaima.ai/resources) · [Pricing & scope](https://genaima.ai/pricing) · [Selected work](https://genaima.ai/work) · [llms.txt](https://genaima.ai/llms.txt) - Contact: hello@genaima.ai