The trigger
A Meta or LinkedIn lead form, your website form, a Shopify event, a scheduled follow-up. Seven sources ship today and they all feed the same pipeline.
You don't have a lead problem — you have a follow-up problem. Choose your industry and floGPT sets up the trigger, the outbound voice agent, the qualification questions and the field capture as one pipeline. No prompt to write.
The call script, the qualification questions and the capture fields come with the pack. You're picking which one, not writing one.
This picks the trigger. You can add the other six sources later — they all feed the same pipeline.
Thirteen packs, each with its own script and qualification questions. Choose the closest match — you can change any of it afterwards.
Not a board with cards you drag between columns. A call pipeline: the five things that have to happen between a form submission and a booked job, each one a part you can open and edit.
A Meta or LinkedIn lead form, your website form, a Shopify event, a scheduled follow-up. Seven sources ship today and they all feed the same pipeline.
The agent dials out automatically — usually within a minute of the submission, while the lead still remembers filling in the form.
Your industry pack's questions, asked in conversation: scope, timeline, budget range, decision-maker. Not a survey read aloud.
Every answer typed and validated — phone, email, date, number, enum — re-asked on the call if it doesn't validate, then written to the call log and a spreadsheet row.
Booked into your calendar, warm-transferred to a human closer, or ticketed with the transcript attached. Never just dropped.
You'd have to edit one. Thirteen industry packs ship with the greeting, the qualification questions, the typed capture fields and the post-call actions already written for that vertical. Picking yours is the setup step; rewriting it is optional.
It's the same voice agent that answers your inbound calls, in 16 voice languages including Hindi and nine Indic locales. It asks, it listens, it re-asks when an answer doesn't validate, and it refuses to diagnose, quote or advise — it escalates to a human instead.
Then this is the shift that happens before your team does. The agent takes the first call, so the ones your closers pick up have already stated scope and timeline — and the ones that were never going to buy never reached a calendar.
Quiet hours, a do-not-call block list, retry caps and per-day calling windows are controls you configure. They're off by default and they apply to trigger-fired calls, so switch them on and review them before you go live. This is not a compliance determination on your behalf.
Two steps need you, and we'd rather say so here than surprise you after signup.
Starting from a website form instead? That trigger works immediately — post your submissions to the webhook URL and the first call goes out on the next dial cycle.
The pipeline is assembled from products that each have their own page.
Book a free 20-minute consultation. Tell us your industry and we'll run your pack against a test number while you listen.
Tell us your industry — we'll run your pack live on the call
Three things, in one step: a Call Pipeline seeded from that industry's pack (the greeting, the qualification questions, and the typed fields the agent captures), a voice agent configured to run it, and the outbound trigger that fires it when a new lead arrives. You still add a phone number, and for Meta or LinkedIn you connect that ad account — everything else is already wired.
Usually within a minute of the form being submitted. The lead is ingested immediately, then the dialer picks up queued contacts on a per-minute cycle and dials within the concurrency limit you set, so bursts queue rather than all firing at once.
Yes. The industry pack is a starting point, not a lock-in. Everything it created — the prompt, the qualification questions, the capture fields and their validation rules, the post-call actions — is editable in the call pipeline editor, and the changes apply to the next call.
Seven trigger sources ship today: Meta lead ads, LinkedIn Lead Gen forms, a generic webhook (your own website form or any external system), Shopify events, time-relative follow-ups, internal events from the workflow engine, and data-collection row watches. You pick one during setup; the rest can be added later and they all feed the same pipeline.
Each field is typed and validated — phone, email, date, number, enum — and asked again on the call if the answer doesn't validate. Afterwards a second pass re-reads the transcript to fill anything missed, keeping a value only when it can cite the lines where the question was asked and answered. The result lands on the call log and, if you enable it, as one row per call in a spreadsheet-style Data Collection with optional Google or Microsoft Sheets sync.
Quiet hours, a do-not-call block list, retry caps and per-day calling windows are controls you configure per organization and per trigger. They're off by default and they apply to trigger-fired calls, so switch them on and review them before you go live — floGPT does not make a compliance determination for you.
Yes, for outbound. The pipeline, agent and trigger are created immediately, but a call needs a number to dial from — link one you already have on the platform or request one, and the trigger starts dialing as soon as it's attached.
The call still ends somewhere. Depending on how the conversation went, the agent books into your calendar, warm-transfers to a human closer, or opens a ticket with the transcript and captured answers attached so someone can follow up. No lead ends the call unrecorded.