Partner Academy

Deliver an FDE project

A forward deployed engagement turns what a client asked for into a workspace their team runs on. Scope it so it can be accepted, build it so the permissions hold, hand it over so it survives you leaving.

01 Define the project

Scope it so it can be accepted

Most delivery arguments are scoping arguments that surfaced late. The interview, the boundary and the acceptance standard are one piece of work, and all of it happens before anyone opens the workspace.

01

Interview the people who do the work

Not only the buyer. The person who runs the process every week knows the exceptions, and the exceptions decide whether the App gets used after you leave.

02

Write the process as it runs today

Who does what, in what order, how often, and where it stalls. This is the document the build gets measured against, so it is worth two pages rather than a paragraph.

03

Draw the boundary out loud

What is in this project and what is the next one. Naming the second phase explicitly is what stops it from living on as an expectation.

04

List the deliverables

Rooms, agents, connections, apps and workflows, written down as a list. A client accepts a list. Nobody accepts an impression.

05

Agree the acceptance runs

The specific flows that have to work, on real data, on the client's own accounts. Write them before you build them, not after.

06

Name who owns it afterwards

The person at the client who will edit guidelines, add members and read a run history once you are gone. If nobody is named, the project has no ending.

A project is finished when someone at the client can change it without calling you.

Kylon Partner Program
02 Build and verify

Permissions and connections first, screens second

A Kylon delivery is a working workspace rather than a codebase: rooms, agents, connections, apps, and the workflows between them. What breaks after handover is almost always access, so access is what gets built and tested first.

01

Permissions before features

Members, roles, room visibility and what each agent may reach. Give the least access that still finishes the work, and say what happens when someone leaves.

02

Connections per agent

Mail, CRM, CMS, dashboards and the rest attach to a person or to an agent, with a reason, rather than to the workspace as a whole.

03

Guidelines carry the working rules

A room's standing instructions are where most of a client's rules end up living, because agents read them before replying. Write them with the client, not for them.

04

Apps on the data agents already hold

Build the business app inside the workspace, on the same context and records the agents work with, instead of beside it.

05

Workflows with a run history

A trigger, runs you can point at, and an approval step wherever a person has to sign off. A workflow nobody can inspect is a workflow nobody will trust.

06

Ship it the same way twice

Rooms, agents, permissions and skills go out through the CLI, so a build you designed once is repeatable on the next client instead of rebuilt by hand.

03 Hand over and maintain

Finished means they can change it

Acceptance is a session, not a mail. You run the agreed flows on the client's own data, with the people who will own it, and you leave them able to change what you built.

01

Run the acceptance flows together

The runs you wrote into the scope, on real data, in front of the owner. Anything that fails here is cheaper to fix than the same thing found in month two.

02

Train the owner, not the room

One named person who can add a member, edit guidelines, read a run history and tell you what changed. Everyone else can learn from them.

03

Hand over the map

What sits where, which connection belongs to whom, and what to do when a run fails. A page of this is worth more than a day of training.

04

Agree what maintenance means

What you watch, what they watch, and how a change request reaches you. Ambiguity here is what turns a finished project into an open-ended one.

05

Book the first change

Something small, a few weeks out. It proves the handover worked and it is where the second project usually comes from.

06

Keep what you learned

The exceptions and the shortcuts from this build are what make the next one faster to quote and faster to deliver.

04 What to build from

You are not starting from an empty workspace

Scenario and industry builds, the product reference, and working templates you can open and adapt rather than design from nothing.

01

What partners build

Client builds by scenario and by industry, and the three working templates partners start most projects from.

See what to build
02

Training and support

The three moves every build repeats, the workspace walkthrough, and the contact you reach while a client build is live.

See training
03

Product documentation

Agents, skills, workflows, apps and the CLI, kept current with the product. Keep it open while you build.

Open docs
04

Industry CRM Gallery

Eight industry CRM starters on live data, useful as the first screen of a scoping conversation.

Open the gallery
05

Workspace Quick Start Guide

Rooms, agents and members step by step, with a cheat sheet for the moves you repeat on every build.

Open the quick start
06

Partner toolkit

The program story and the apps you can demo live, for the meetings before the project starts.

Open the toolkit
What you leave with

One client project through acceptance, with a named owner on the client's side.

More for partners

Bring a project to us before you quote it

Tell us the client and the work, and we will go through the scope and the acceptance standard with you while it is still cheap to change.