Engine 02
AI Enablement
For the operator who owns how the work actually gets done.
Teams of agents deployed into the segments that carry the most repetitive load, built on Dust and wired into the tools that segment already works in. This engine does not print money. It makes the day run faster, with fewer errors and less of it spent on the parts a machine does better.
From a process to an agent team
- 01
Start from a process, not a tool
We pick a process your team already runs in software, because that is the only kind an agent can take over. Most often it sits in sales or marketing. Ops is fair game too.
- 02
Build the team, not the prompt
An AI team is several agents working together toward one goal, rather than one clever prompt trying to do everything. We build them on Dust and deploy them into the tools that segment already lives in.
- 03
Wire the automation layer
n8n handles the repetitive plumbing between systems, so the agents spend their time on judgment rather than on moving data from one place to another.
- 04
Give the team an assistant
Claude sits alongside as the assistant people talk to directly, with access across the tools it needs at once instead of one integration at a time.
- 05
Extend as the next process surfaces
The first build is never the last one. Once a segment sees how much an agent can absorb, the next candidate is usually obvious.
Outcome
Hours back, every week
The stack
What each piece is for
Claude
The assistant your team talks to
Certified Expert. One place to ask, with access across the tools it needs at the same time.
Dust
Where the agent teams live
Certified Partner. Multiple agents deployed across the company, whether or not they are related to each other.
n8n
The automation layer
The repetitive plumbing between systems, so the agents are not doing data entry.
What we run
On the account, every month
Agent teams scoped per business segment, starting with sales and marketing
Agents built on Dust and deployed into the tools that segment already uses
The repetitive plumbing automated in n8n
Claude set up as the assistant your team works with directly
Processes documented in your workspace, so your team can run them alongside us
Your data stays in your own cloud. Agents run against your systems, and nothing is copied out to ours
Models and prompts kept current as the underlying technology moves
What you bring
A process worth fixing, and the person who runs it today
Access to the tools that process already lives in
The business context the Company Brain and the agents read from
Not the right fit if
The process is not digital yet. An agent can only take over work that already happens in software.
You are looking to cut headcount rather than to make the work run better.
Nobody will own the agents once they exist.
Results
What AI Enablement has produced
- 12
- AI workflows for automated account research
- 2,000+
- Accounts mapped and enriched across four markets
- 4
- European markets, each with its own playbook
iFeel
iFeel
iFeel
AI Enablement, in more detail
Not in revenue. This one improves processes rather than producing pipeline: the same work happening faster, with fewer errors, and less of your team's day going into the parts a machine does better. If a process is already digital, an agent can usually take a piece of it.
Mostly sales and marketing, which is where the repetitive load usually sits. We are happy to work in ops as well.
A team of agents working together toward one goal, each handling the part it is suited to, rather than a single prompt asked to do everything. That is the difference between something that survives contact with a real process and something that impresses in a demo.
Yes, though usually inside a wider engagement rather than as a standalone build. iFeel runs twelve AI workflows we built for automated account research across four European markets. Orb runs agents for account scoring and MQL qualification. VeUP runs a tiered reply system where lower-priority responses are answered autonomously and the top tier is drafted for a human to approve. BlueMargin is the first engagement where the agents are the product rather than a component. We run our own too: Radar keeps the team current, and a reporting agent assembles our numbers.
Nowhere. The agents run against your own systems, in your own cloud. Nothing is copied into a OneThousandPieces environment to make them work.
Accounts running this engine
Write-ups go up as each client signs them off.
Runs well with
The four engines read from the same Company Brain, so each one makes the next cheaper to stand up. Plenty of clients run just this one.
Ready to build AI Enablement?
Book a call and we'll map the engine against how your revenue motion runs today.