The Agent Era
How WKT Wins Next
Technology-powered solutions for training that matters
Leadership Lunch & Learn — February 11, 2026
1 / 16
← → arrow keys · click · swipe
Before we start...
Think about something you've wanted to do — but couldn't.
- Not enough people
- Not enough budget
- Not enough time
- Not the right tools
Hold that thought.
2 / 16
The Shift
Every ~20 years, everything changes
2000s
The Internet
Connected information
→
2010s
Mobile
Connected people
→
2025+
AI Agents
Connected workers
We're at the beginning of the third one.
3 / 16
The Shift
This isn't ChatGPT
"Hey AI, write me an email"
AI agents are not chatbots. They:
- Read your files, emails, and databases
- Take action — send messages, update systems, build things
- Remember who you are and what matters to you
- Work autonomously — even while you sleep
They're not tools. They're team members.
4 / 16
How It Works
Four layers, working together
🧭 Soul + Memory + Skills
Who the agent is — personality, knowledge, capabilities, tools
⚙️ OpenClaw — The Agent Platform
Open-source framework — messaging, tools, memory, scheduling, agent networking
🧠 Claude by Anthropic — The Brain
The intelligence that reasons, writes, understands, and decides
🖥️ Your Hardware — The Container
Runs on YOUR machine — your data never leaves your infrastructure
5 / 16
How It Works
The Brain: Claude
Built by Anthropic — an AI safety company. Claude is designed to be:
- Helpful — genuinely capable, not just clever-sounding
- Harmless — refuses dangerous or unethical requests
- Honest — tells you when it doesn't know something
Think of Claude as a brilliant hire who knows everything — but has no memory of you.
That's where everything else comes in.
6 / 16
How It Works
The Body: OpenClaw
Claude is the brain. OpenClaw gives it hands, eyes, and a voice.
- Messaging — Telegram, Signal, Discord, Slack, email
- Tools — files, code, web search, APIs, databases
- Memory — persistent files that survive between sessions
- Scheduling — proactive check-ins, automated tasks
- Agent networking — secure communication between agents
Open source — anyone can inspect the code.
7 / 16
How It Works
Creating a Soul
Every agent starts with plain text files that define who they are:
SOUL.md
Personality, values,
communication style
+
USER.md
Who they serve,
context, preferences
+
IDENTITY.md
Name, role,
backstory, purpose
You literally write who the agent becomes. Editable. Auditable. Yours.
Example: Kay — named after Eisenhower's wartime aide. Sharp, direct, resourceful. Not a chatbot voice — a real personality.
8 / 16
How It Works
How an Agent Remembers
AI models start every session blank. No built-in memory.
We solve this with files.
📝 Daily Notes
Raw logs —
what happened today
→
🧠 MEMORY.md
Long-term memory —
curated, like a journal
→
🧭 SOUL.md
Core identity —
never changes
The agent wakes up, reads its files, and picks up where it left off.
Names, preferences, projects, lessons learned — all retained.
9 / 16
How It Works
Trust, But Verify
A powerful agent without controls is dangerous. Here's how authority works:
🟢 Free to Act
Read files, search the web, organize data, research, check calendars
🟡 Ask First
Send emails, post publicly, contact people — anything that leaves the machine
🔴 Never
Share private data, run destructive commands, bypass security, act without oversight
🔒 Infrastructure
Allowlists, group policies, audit logs — everything is logged and traceable
You define the boundaries. The agent respects them.
10 / 16
In Practice
Agents Working Together
Agents don't work alone. They coordinate — securely, across machines.
⟷
Kay
Work Agent
MacBook Pro
⟷
Belle
Home Agent
Home Server
⟷
Each agent on separate hardware, connected over encrypted networks.
They delegate, coordinate, and report back.
11 / 16
In Practice
10 Days. One Agent.
- Built and deployed 4 websites — no developers, no agencies
- Organized 138 tasks across 5 databases
- Researched markets and delivered executive briefs
- Managed email, calendar, and messaging across platforms
- Coordinated with a second agent on separate hardware
- Built this presentation
Hours, not weeks. One agent, not a department.
12 / 16
Live Demo
Let's See It In Action
Live
Watch a real task happen in real time.
No scripts. No pre-recorded demos. Just a real agent doing real work.
13 / 16
The Challenge
The Multiplier Effect
Think about your current workflow:
- How much time goes to repetitive coordination?
- How much knowledge is trapped in your head?
- How many tasks wait because you're the bottleneck?
60
→
660
Not replacing people. Multiplying what each person can do.
14 / 16
The Challenge
Your Turn
Remember the constraint you were holding?
🔄
What do you do every day that follows the same pattern?
📋
What knowledge do you have that isn't written down?
⏳
What takes you hours that could take minutes?
🚀
What would you do with 3× your current capacity?
If you can describe it, an agent can probably do it.
15 / 16
The Agent Era
Is Here
We unleash confidence by improving competence.
And that includes our own.
What constraints just disappeared for you?
16 / 16