FAQ
Questions
and answers.
Answers stay open. Start with a section, or with the question people ask first.
Q → A
A signal comes in. A decision is made. Systems are updated. A person steps in if needed.
Everyone asks this
What’s the difference between an AI agent and a chatbot?
A chatbot answers. An agent finishes the work in your systems.
01 · What this is
What this is
What does Bloomit Labs actually build?
We consult marketing, sales, and customer success teams, then deploy agents into the marketing platform, the CRM, and the success tools those teams already use.
Who is this for?
Companies that want agents in marketing, sales, or customer success — inside the systems those teams already use.
What work do the agents actually take on?
Scoring and handoff in marketing. Routing and follow-up in sales. Renewal, onboarding, and expansion in customer success. One process at a time, saved in the tool that team already uses.
Will this sit on top of the systems we already use?
Yes. Marketing automation, the CRM, and customer success tools. Your tools stay.
Do we buy a platform and set it up ourselves?
No. We don’t hand you a builder or a generic agent to click into place. We learn how the work already moves in your systems, then design and ship the one that runs that process—exceptions, approvals, and the points where a person still needs to step in.
02 · Fit & ROI
Fit & ROI
How do we know which process to automate first?
Start with a repeating job in marketing, sales, or customer success that has a clear outcome — a campaign, a follow-up, or a renewal.
What ROI should we expect?
You get hours back and a live process. We start with one workflow. We’ll say if it isn’t worth building.
We’ve tried automation before and it stalled. Why would this be different?
Most tools fail because they ignore how work actually moves—exceptions, workarounds, and real systems. We map the messy process first, build failure routes and human review into the design, and ship a first version against real cases early.
Should we hire you or build in-house?
Hire us when you have a clear process, lack dedicated AI capacity, and want a working system in weeks. Build in-house when AI is core product work and you already have the team. A common path: we build and validate the first version; your team maintains it once it’s documented.
03 · Trust
Trust
Where does our data go?
We design around your constraints: integrations into systems you control, credentials in your environment, and clear rules about what leaves your perimeter. We don’t use your business data to train a public model. Details are written into the scope before we build.
Who owns the code, prompts, and workflows after delivery?
You do. Automations and agents live in your accounts where possible. You keep the credentials, documentation, and artifacts.
How do you keep an agent from doing the wrong thing?
Guardrails are built into the system: action limits, policy filters, confidence thresholds, and a person in the loop on important steps. Agents act with the access you already grant. Important tool calls and decisions can be logged so you can see what ran and why.
Will people still be involved?
Yes—where judgment matters. Routine steps complete alone; ambiguous or high-stakes steps escalate to a person with a clear queue and context. Your team stays in control of exceptions and ongoing refinement.
04 · Working together
Working together
What happens on the first call?
We pick one real process and walk it: who touches it, where it stalls, and which systems hold the current records. You leave with a first-cut map—what’s worth automating, what we’d connect, and a clear path to a first version. The map is yours whether we work together or not.
How much time does our team actually need to spend?
You’re the experts on how the work really moves. We need a few working sessions to map it and check the edge cases — not a standing committee or a second job. We design and ship; you stay available for decisions, reviews, and the people who own the systems.
How long until something is live?
One process, a handful of systems: often in production against real cases in a matter of weeks. We map first, then put a first version live before a full cutover. Messier exceptions and more systems take longer. We do not stretch the timeline to fill a quarter.
How do you price this?
By the process, the systems, and how far you want to go — not a mystery product. We quote after we’ve seen the work. A focused pilot is the usual first price, so you can see hours come back before you commit to a larger system.
What do we get when you’re done?
A production workflow or agent in your environment, integrations that stick, monitoring and failure alerts, and a handoff: runbook, walkthrough, and clarity on how to change it. Optional ongoing improve-and-scale support if you want it.
What happens when something breaks in production?
Automations include error routes: retries where safe, alerts where not, and logs to diagnose. We design for failure up front. Support after launch can be a handoff to your team or an ongoing cadence with us—decided in scope, not after a surprise.