Forward Deployed Agent
our forward deployed agent turns
invisible work into AI teammates
01
observe
[fda] watches, quantifies
02
diagnose
assess insights; rank by hours back and business impact
03
plan
[fda] specialist agents queue the implementation plan
04
execute
press go, deploy AI 19.5x faster
Free · Unlimited machines · Metadata only, never your content
v0.2.1 · 3.5 MB Windows 10+ · 10.7 MB macOS 13+
We put fda on 4 of our own machines.
Nobody changed how they worked. Three weeks later, here is what it found.
The same things, over and over
Add it up
168 hrs / month
One job's worth of busywork, wasted. Every week.
How the forward deployed agent works
01
Install the agent
Deployed in an afternoon, under 1% CPU. Nobody changes how they work.
02
Work normally, two weeks
It learns the real job — the tools, the loops, the switching nobody logs.
03
Get your report
Where the hours actually go, and what a teammate could take over. Free, yours to keep.
04 — We build it with you
Side by side with your team
We build it with your team, not for them — sitting alongside the people whose work it learned. It ships after the two-week scoping period, trained on your workflows, not a generic model.
What it takes over
Trained on 530 hours of your team's actual work
fda learns the work. Your AI teammate does it.
— 02 / Learn · Map · Take over
Remove monotony from your day-to-day.
01 — Learn
Install once. Forget it exists.
Deploy across the company in an afternoon. The agent runs under 1% CPU, changes no workflows, and no one has to remember to start a timer. It observes; your team works.
Capture stream · local buffer
02 — Map
fda finds what no one would report.
Repeated sequences, re-entered data, tools that never talk to each other — patterns spread across weeks and thousands of individual actions, invisible from inside the work.
03 — Take over
Every piece of monotony becomes work your AI teammate can own.
What it takes over, what that gives back, how long it takes to build — ranked by value. The deliverable is a hiring plan for software, not a dashboard.
— 03 / The finding
Our internal case study.
4 people, 3 weeks, 530 hours. Nothing redacted, nothing illustrative. Click through all 5 pages.
We measured ourselves first.
It was worse than we thought.
We put our forward deployed agent on 4 of our own machines — recruiting, client success and the founder's desk — and recorded 3 weeks of real work. Every application, every browser tab, every copy and paste. No surveys.
168 hrs / mo
= one full-time employee of repetitive work, spread across 4 people. You already employ the 5th person. You just can't see them.
~2,300
copy-pastes into AI tools across the team — context shuttled by hand between Slack, Gmail, the ATS and the models
~25,500
app switches measured — one every 75 seconds of work
2,400+
AI sessions already running — every one of them fed by hand
attain internal case study — july 2026
This is the same report you get after your two weeks.
Want this report for your own team?
530 hours analysed. Yours in two weeks, free.
— 04 / Scale
More desks = better AI teammate.
Every machine you add sharpens the picture. What looks like one person's quirk turns out to be how the whole company works.
One desk
A habit
How one person works. Easy to dismiss as personal preference.
One team
A pattern
The same workaround on four desks is not a quirk. It is a process nobody designed.
The company
A system
fda learns how the business actually runs.
Free on every machine, so don't be shy.
Unlimited seats · Deploys in an afternoon
2 weeks of work.
1 new teammate.
Free · Unlimited machines · Metadata only, never your content
Verify the build
macOS · v0.2.1 · 10.7 MB · SHA-256 1316C3F4 … 2D46E410
Windows · v0.2.1 · 3.5 MB · SHA-256 D3FD0B49 … 6EFD9A5F