Freelance project · AI products
A voice agent that answers, qualifies and books, around the clock
AI Calls is a voice assistant that handles incoming and outgoing calls: it qualifies leads, answers FAQs and books appointments, 24/7. I designed and built the web app teams use to run it, from campaign dashboards and call history to lead lists, workflows and integrations.
01The problem
A voice agent can take calls all day, but the team behind it still needs to know who it spoke to, what happened and what comes next. The product had to make a constant stream of automated calls as traceable as a sales rep’s notes, and let non-technical teams change how the agent behaves.
Who it’s for
Sales teams
Qualified leads and booked meetings without dialling every number themselves.
Support teams
FAQs answered at any hour, with a clear handover when a person is needed.
Operations leads
Campaign metrics and call history they can check at a glance.
Workspace admins
Workflows, integrations and tags they can set up without code.
02The main flow
01
Import a lead list
02
Launch a campaign
03
Agent makes calls
04
Review call outcomes
05
Meeting gets booked
03Structure
Dashboard
Calls
Leads
Automation
Settings
04Key decisions
Decision 01
A call log sorted by time shows what happened today, but not where a lead stands after several calls.
Considered
A flat, time-ordered log of every call.
Chosen
Every call is tied to a lead: the lead detail shows each call, its outcome and the next task, and the call history links straight to it.
Why it works
Teams think in leads, not calls. One record per person means a booked meeting or a missed callback is never lost in a long log.
Decision 02
What the agent does after a call branches: book a slot, send a follow-up, create a task or hand over to a person.
Considered
A long settings form with a toggle for each possible follow-up.
Chosen
I designed a visual workflow builder where each step is a block and each call outcome is a branch, with integrations as steps in the same flow.
Why it works
Branching logic is easier to read as a picture than as nested toggles, and teams can see the whole path before they switch it on.
Decision 03
Once the agent runs 24/7, nobody can review every call, so the team needs to know where to look.
Considered
Opening the app on a feed of the latest calls.
Chosen
The dashboard opens on active campaigns and their key metrics, and lets teams drill into the calls behind any number.
Why it works
Managers decide at campaign level, such as pausing a campaign or changing its workflow, and only drop to single calls when a number looks wrong.
05What’s inside
Tracks the key metrics of every active call campaign in one view.
Lists every incoming and outgoing call with its lead and outcome.
Brings scraped prospect lists into one place, ready for a campaign.
Sets what happens after each call, from booking to follow-up tasks.
Custom tags and controls let each team fit the agent to its pipeline.
06Screens
07How it was built
I designed it and wrote the code, so the design survived contact with reality.
Interface
Services
AI and voice
Data
08Outcome
Delivered
What I learned
With voice AI, the product isn’t the call. It’s everything a team needs to trust that the call went well.
A visual workflow made automation feel safe to change for people who would never edit a config file.
Next step
Pick a time that suits you. You'll leave the call with a clear next step, whether we work together or not.
