Chief of Staff Agent
Inbox, priorities, and the updates that go out in your voice.
You, Your Team, and a Little Legion of AI Agents
We help founders and operators scale their deep expertise, so they can be more creative, more innovative, and put deeper intelligence to work to outrun the competition.
Hire a Little Legion, 1000x Your Team
We build you a custom AI agent team around the work you're already repeating every week: the inbound that needs a decision, the report that's rebuilt every month, the drafting that has to sound like you. Every business has these. Your agents run on your own accounts and your own files. You keep the calls only you can make.
Founder / Operator
You
Head of Recruiting Agent
Every candidate intro, first pass, and shortlist, with reasoning.
User Research Lead Agent
Customer calls, surveys, and interviews distilled into insights.
Market Analyst Agent
Competitor decks, market reports, and trend signals tracked weekly.
Financial Analyst Agent
Monthly spend, margin, variance, and the numbers behind every update.
How We Build Your Legion
Four phases. Each phase is a working session with your team, run on frameworks we've built over fifteen years. You keep what every session produces. We run the phases with you, or for you.
01
Discovery
People
Process
Platform
Assess
Approve
Intake
Triage
Log
Draft
Send
handoff
rework
We map your operating model across people, process and platform, then trace every handoff and every loop where work comes back. Most of what we find is machinery. Wiring, not judgment. What we're hunting is narrower: the one or two steps where your real judgment lives.
- Discovery Worksession
- Operating-Model & Function Map
02
Prioritization
JudgmentJudgment
Rule fragilityFragility
Volume & valueVolume
Unstructured dataData
Risk & safetyRisk
Agent‑fit
Automate
High volume with low judgment isn't agent work.
We score every candidate workflow on five dimensions: judgment, rule fragility, volume and value, unstructured data and risk. Then we read the profile rather than the total, because a workflow can score well and still be the wrong shape for an agent. Shape wins. Work that doesn't fit goes to conventional automation.
- Agent Suitability Filter
- Opportunity Register
- Go / No-Go Gate
03
Build & Refine
Orchestrator
Specialist pool
Agent
Agent
Agent
Shared memory
Producer
Reviewer
refine
Gate
pass
Ship
Human
rework
We wire the agents into a workflow: an orchestrator over a pool of specialists, a shared memory, and a producer and reviewer refining the output in a loop until it holds up. A deterministic gate then ships the work or escalates it to a person. The gate runs on rules, not opinion. Rework re-enters the loop. Nothing goes live until it's cleared the Faithfulness-First Eval.
- Agent Spec
- Boundary Test
- Flow Patterns
- Faithfulness-First Eval
04
Deploy & Operationalize
The Playbook
How each agent runs
Frameworks
Function Map
Suitability
Agent Spec
Flow Patterns
Eval Harness
Rollout
Maturity ladder
Crawl
Walk
Run
Rollout checklist
Pick the first function
Run it beside the team
Hand the playbook over
You get the playbook, the frameworks and a Crawl-Walk-Run rollout, and we train your team to run the agents, refine them as the work changes and build the next agent once we're gone. Yours to keep. You end up with a map of what your agents own and what your people own, function by function. The Automation Roadmap tells you what to build after that.
- The Playbook
- Crawl-Walk-Run
- Training & Enablement
- Automation Roadmap
The Team: Kenan Ali and 1,000 Agents
15+ years turning strategy into brands and products that made it to market. The last three, with AI. Same job, new tools.
Little Legion is Kenan Ali, Senior Director of Enterprise Transformation at Code and Theory, and for 15+ years we've worked in the gap between strategy decks and the work that actually ships at startups and across the Fortune 100. We don't hand over a deck. We build the technical, operational and strategic capacity inside the team, so the work keeps moving after we go. Then we go.
We've led teams across design, engineering, strategy and research and launched and relaunched products in markets around the world, and the lesson every time is that working prototypes and pilots compress months of debate into days of evidence. Teams stop arguing and start deciding. Everyone can finally see the thing they are deciding about. Proof, not persuasion.
Little Legion is an AI-native services firm. One person makes the judgment calls, a standing team of agents does the repeatable work.
Today we help founders and operators move AI from strategy decks into production. Weeks, not quarters.
Past clients
"I wanted to take my AI work to the next level and learn how to build agents, which felt pretty overwhelming. Kenan made it understandable and relatable, and walked me through every step. He clearly knows AI inside and out, and he's great at teaching it. The results came fast. Within a few weeks I had agents handling a time-consuming part of my business, and it's drastically cut the time it takes. I can already see how this is going to transform how I work. I can't recommend him enough."
"Kenan built me a Claude agent that runs my entire COGS and PO workflow. What took a day now runs in one pass."
"Kenan made building AI agents feel approachable and genuinely useful for my business. Complex tools feel easy, practical, and immediately impactful."
Where AI Is Taking Expert Work
We write about where AI is taking expert work, and how to stay useful as it does.
For thirty years the web paid you to go viral. Now it pays you to be useful.
Read on LinkedIn →Everyone at your company is building software almost no one will use
Read on LinkedIn →Six ways to wire an agentic strategy team, and when to use each
Read on LinkedIn →Your strategy team just got a new org chart. Half of it is AI.
Read on LinkedIn →How LLMs and answer agents killed traditional brand differentiation
Read on LinkedIn →I read (most of) NASA's $1.7 billion failure manual so you don't have to
Read on LinkedIn →We're teaching ourselves to think like machines (and that's a problem)
Read on LinkedIn →Two Ways to Build Your Legion
Both options start with the same 30-minute call. The difference is who holds the keyboard. After that it comes down to how much of the build you want to own, and how much time your team has. Your call.
Agentic Transformation
The whole build, done for you.
We run all four phases for you, end to end. You end up with a legion already running your work and a team trained to grow it. Then the Automation Roadmap, which tells you what to build next. All yours.
- Full discovery-to-deployment build, done for you
- Your team trained, and a playbook they keep
- The Automation Roadmap, handed over
1:1 Training / Consulting
You at the keyboard, us alongside.
We build your first agents alongside you on your real work, and teach you our method while we do it. You keep building after we go. Same end state, either way. You're just the one at the keyboard.
- Your first agents, built together
- Our method, taught as we go
- You own and extend the legion without us
Questions
What exactly is an AI agent, and how is it different from ChatGPT or a Zapier automation?
ChatGPT is built for conversation. A Zapier automation runs a fixed path you defined in advance. An agent owns a piece of work. You set it up once with your standards, your files and the way you want the work done, and after that it makes the judgment calls a person would, the same way every time, without you re-explaining. A little legion is a set of these specialists, each running a part of the work you're repeating every week.
What kind of work can actually become an agent, and what should stay with me?
The best candidates are the judgment-heavy things you repeat: the inbound that needs a decision, the numbers that get reconciled, the first pass at research. Purely mechanical steps are cheaper as plain automation. The relationship-driven calls stay with you. We score every workflow before we build anything.
How much does it cost?
Pricing is bespoke. The scope drives it, not a rate card. The engagement can be a single agent built alongside you, or a full transformation with your team trained and the Automation Roadmap handed over. We size it with you on the first call, so you'll know the number before anything starts.
How fast will I have something working?
Weeks, not quarters. The first pilot is deliberately narrow so you get real evidence on your own files quickly, then we widen from there one proven agent at a time, instead of a big-bang rollout that stalls.
How do I know it will do the work to my standard?
Every agent is tested against a gold set of examples you sign off on yourself, before it touches anything real in your business. One question decides it. Did it do the work the way you would, every time, without you re-checking? Nothing goes live until it's cleared that bar. Autonomy is earned, never granted.
What happens when an agent gets something wrong?
An agent is built to flag what it's unsure about rather than guess, and checkpoints keep a person in the loop on the calls that matter. You set how much each agent is allowed to decide on its own, and where it has to stop. Tighten that when trust dips. Loosen it as trust grows.
Is my data safe?
Yes. The agents run on your own accounts, your files, and your environment, so nothing has to leave your walls for this to work. You own the setup and everything it touches.
Do you build the legion for me, or teach me to build it myself?
Either. You don't need to be technical for either one. With Agentic Transformation we run all four phases and hand you a working system with your people trained on it and running it. With 1:1 Training we build alongside you, so you keep going without us. Both ways you walk away with the agents, a written playbook for how each agent runs, and the Automation Roadmap for everything after that.
Let's build your little legion
We start with a 30-minute call. We go through your operating model, your business model and the core functions you run every week. Then a proposal built for what we find. No template.
