Forward Deployed Engineering: FDE Pods get AI working inside your business
A pod is a small team that sits with your people, picks one job AI can take on, and builds it into your systems.
What is a forward deployed engineer?
A forward deployed engineer is an engineer who works inside your business to get AI doing real work. They sit with the people who do a job every day, learn how it is really done, and build AI into the tools those people already use, in your own cloud account, one job at a time. When it runs, they hand it over and step back, and everything built stays yours: the code, the guide to running it and the people who know how.
The name comes from the military, where forward deployed forces are kept close to where they will be needed instead of being sent from home. Software companies took the phrase for engineers placed with their customers. Here it means an FDE Pod works where the job is done: in your tools and your meetings, with the people who do it.
- They sit with your team
- They build it into your systems
- They hand it over and step back
The pod joins your meetings and tools, and watches how the job is really done before designing anything. The AI goes into the tools your people already use, in your cloud account, built with your people. Your team runs it. The code is yours, and the work ends on a date written down.
Why companies bring one in
Most businesses already have AI tools. Few have one doing a real job every day. That takes connecting it to your systems, your data and the people who do the work.
An AI pilot stuck in the demo
It worked in a demo. Nobody connected it to your systems, so the job is still manual.
A job done by hand, all week
Order questions, returns, invoices and listings are small jobs that together fill most of a team’s week.
Answers spread over three systems
The answer sits in orders, the courier’s site and the helpdesk. A person opens all three each time.
A tool that does not fit
The vendor’s AI works one fixed way. Your process works another, so nobody uses it.
An old system that has to stay
The order system is old and stays. The AI has to read it and write to it safely.
An AI that must be checked first
A reply, a refund or a payment needs a person’s yes. The check has to be built in.
Bring one in for a real job that someone on your team already owns.
We build AI into that job, inside your systems, with the people who do it.
What an FDE Pod does
What it does for your business
- Time back
Routine work off your team’s week
The AI reads, looks things up and drafts. Your people decide.
- Proof first
Proof on your own cases first
The AI is scored on your past cases before the full build starts.
- Less waste
A small build that gets used
One job, with an owner and an agreed score, in tools people already use.
- Builds up
Each next job starts further ahead
It reuses the connections, data and checks the first job built.
A first job at an online retailer
One sample job with made-up figures: an online retailer’s order questions. Your own figures come out of week one.
- A question arrives
- The AI drafts the reply
- A support agent reads it and sends it
Before
- 2,000order questions a week
- 6 minuteseach: find the order, check the courier, write the reply
- 200 hoursof the team’s week
- 4 hoursuntil a customer hears back, on average
After the first job
- 2,000order questions a week, each read and drafted by the AI
- 1 minuteeach: a support agent reads the draft and sends it
- 33 hoursof the team’s week
- 10 minutesuntil a customer hears back
A person still sends every reply. The AI cost is about three cents a reply, on your own cloud bill.
Other first jobs in retail
- Returns and refunds
- Product listings
- Supplier invoices
- Reviews and complaints
How a first job runs, week by week
We sit with your people, pick one job, agree how it is scored, and gather real past cases.
What you hold: A signed job sheet, and real past cases to test on
Your sign-off: Go ahead, or stop here
A prototype in your cloud account, tried on the past cases and scored against the score you agreed.
What you hold: The scores, a working prototype, the cost of each task
Your sign-off: The approach we build, or stop here
Connected to the systems you use, behind approvals and alerts, with your people building beside ours.
What you hold: The workflow live, tested against the score you agreed
Your sign-off: Acceptance against those past cases
Written steps for when things go wrong, training, and two weeks watching it together before we step back.
What you hold: The code, the written steps and people who run it
Your sign-off: You close the engagement
How to start, and how it grows
- Start here
Week one on its own
We learn the job and write the job sheet. You keep it and decide.
- Then
One job, through to handover
The four steps above, ending with your team running it.
- As it grows
Job after job
Each next job gets its own job sheet, signed before work starts.
Compared with the options you already know
A consultant
What you keepA plan to build
A software product
What you keepA licence, while you pay
An engineer you direct
What you keepThe work they did for you
A managed service
What you keepA platform looked after
An FDE Pod
What you keepOne job running, yours
Compare in detail
| Compared on | A consultant | A software product | An engineer you direct Cloud Staffing | A managed service Managed Services | An FDE Pod |
|---|---|---|---|---|---|
| Who does the work | They study it and write it up | Your people, inside the vendor’s tool | Our engineer, as part of your team | Our team, from outside yours | The FDE Pod with your people, inside your team |
| Who leads day to day | Your sponsor | You, within what the tool allows | You | We do, against a service level | We do, against the job sheet you signed |
| What you hold at the end | A plan to build | A licence, for as long as you pay | The work your engineer did, as you directed it | A platform looked after while the service lasts | One job running in your systems, its code and guides, your people trained |
| Best when | You need a decision, not a build | A standard job the tool already does | You know what to build and need hands | You want a platform looked after, month after month | A job with an owner that AI should take, and nobody to build it |
| How it ends | With the report | When you cancel | When you say so | When you end the service | On the date on the job sheet, with everything handed over |
An FDE Pod
- Who does the work
- The FDE Pod with your people, inside your team
- Who leads day to day
- We do, against the job sheet you signed
- What you hold at the end
- One job running in your systems, its code and guides, your people trained
- Best when
- A job with an owner that AI should take, and nobody to build it
- How it ends
- On the date on the job sheet, with everything handed over
A consultant
- Who does the work
- They study it and write it up
- Who leads day to day
- Your sponsor
- What you hold at the end
- A plan to build
- Best when
- You need a decision, not a build
- How it ends
- With the report
A software product
- Who does the work
- Your people, inside the vendor’s tool
- Who leads day to day
- You, within what the tool allows
- What you hold at the end
- A licence, for as long as you pay
- Best when
- A standard job the tool already does
- How it ends
- When you cancel
An engineer you direct Cloud Staffing
- Who does the work
- Our engineer, as part of your team
- Who leads day to day
- You
- What you hold at the end
- The work your engineer did, as you directed it
- Best when
- You know what to build and need hands
- How it ends
- When you say so
A managed service Managed Services
- Who does the work
- Our team, from outside yours
- Who leads day to day
- We do, against a service level
- What you hold at the end
- A platform looked after while the service lasts
- Best when
- You want a platform looked after, month after month
- How it ends
- When you end the service
What stays with you
You own what we build, and we step back once your team runs it.
The code is yours
The workflow, its code and its tests stay with you.
Your data stays in your cloud
We work through access you grant, and we keep no copies.
Your cloud, the model that fits
AWS, Microsoft Azure or Google Cloud, and the AI model your own cases favour.
A person approves what the AI does
Nothing is sent or changed in your systems without a person’s yes.
A written end date
It goes on the job sheet once the job is chosen.
Questions
Is this staffing or consulting under another name?
No. A consultant leaves you a plan, and with Cloud Staffing our engineer joins your team while you lead. Here an FDE Pod leads one job to production with your people, tests it against the score on the job sheet, and leaves your team running it.
Where does the FDE Pod work from?
Inside your team and your tools, from wherever your team works. How often you meet, and where, goes on the job sheet.
Who is in an FDE Pod, and what does it take from our people?
An engagement lead, engineers and an architect who checks the design, all from Cloud 23. From you, the owner of the job for an hour or two a week, your security contact at the start, and the people who will run it for part of their week while it is built, because they build it with us.
What happens when the engagement ends?
Your people run it, with the written steps they helped write and the code in your repository. The end date is on the job sheet, and our access ends with it.