Sample Bots | Master AI and Build the Future

What is an AI harness?

The layer that helps a model orchestrate tools, memory, and workflows. A bot is not the model alone. The harness is what makes the bot useful in a business.

1. Start with a model

A harness begins by calling the model through an API: a prompt in, a temperature, a structured answer out, sometimes streamed.

2. The harness orchestrates

Instead of one question and one answer, it plans, loops, delegates, and uses tools, subagents, containers, and recipes.

3. It manages memory and context

Models do not keep a lasting memory. The harness retrieves and writes an external store so the context stays in good shape.

4. It compounds

As models improve, they help extend the harness itself: programs, books, complex systems, and long-running tasks.

Who does what

The same harness, three ways in. Business customers subsidize the students. Professionals learn to build it. We create, manage, and maintain it for the enterprise.

Standard · Students

Free to learn and join, including education accounts. Subsidized by the business customers. Community updates, cohort access, and a self bot.

Free on Standard

Premium · Professionals

We help you build the harness and become a Delivery Pilot. Full course access, active 1-1 sessions, and hands-on bot setup. $1 a month.

Start Premium

VIP · Enterprises

As a managed service provider we create, manage, and maintain the harness for the business. $10 a month, and managed bots at $100 per bot.

Talk to us

Securing, adapting, and training

The proof is not a slide. It is the marketing bots in the catalog, kept alive by the same Delivery Pilot practice we use with a business.

🔒 Securing

The bot stays inside a defined role and domain. Access, content, and actions have a boundary before the bot talks to a customer or a cohort.

🔄 Adapting

When the offer, the audience, or the channel changes, the bot changes with it. The business output stays the same: the role, the domain, and the topics.

🎓 Training

Certified professionals learn the bot by running it. The cohort that sponsors the work is the same cohort that can carry the next bot.

Marketing bots, working now

The private catalog at github.com/rifaterdemsahin/bots holds 52 roles, last read 4 October 2026. These marketing roles are marked Active. They sit at the front because they are the proof a business can buy.

Active

Content creation

Writes customer-discovery content and feeds the video pipeline the business uses to find the next buyer.

Active

LinkedIn messaging

Drafts and sequences professional messages and keeps a relationship pipeline, so outreach does not depend on one person’s memory.

Active

LinkedIn maintenance

Keeps the audience inside the connection cap, turns connections into followers, and collects feedback.

Active

Content curation

Keeps posts and notes at one standard, so the public message matches the certified practice.

Active

Learning

Builds learning paths and tracks skill gain, which is how a free student moves toward certification.

Active

Video naming

Renames each clip from its first frame, so the content library stays searchable as volume grows.

Active

Narration

Local text-to-speech for posts and notes, so a marketing piece can ship with audio without a studio booking.

Active

Capacity planning

Warns when video production, customer discovery, and consulting no longer fit in one week.

Also working for the business

These roles are Active in the same catalog. Hermes, the gateway that runs the scheduled work, is Healthy. One recorded result: the expenses bot reported 23 receipt files for September 2026. That pack still needs a human review before it is closed.

Delivery — contract work, the morning checklist, and lead follow-up.
Project management — deadlines and status across the work in flight.
Strategic planning — goals, resources, and risks lined up.
Finance — cash in, cash out, and invoicing.
Automation — scripts, n8n flows, and the other agents.
Email — outbound mail through the n8n path.
Drive search — full-text and OCR across the shared business drive.
Drive scan — opens a file and writes what is inside it.
Knowledge synthesis — connects notes and surfaces the pattern.
Retrieval tuning — improves file-based retrieval without a full retrain.
Azure delivery — infrastructure as code and secrets by name.
Build crew — explore, write, review, and research as separate workers.
Multi-agent orchestration — runs independent Delivery Pilot tasks in parallel.
Hermes gateway — Healthy. Runs the scheduled jobs and the heartbeat.

Catalog, private: github.com/rifaterdemsahin/bots

Why a business sponsors this

🔁 The B2C2B loop, stage by stage: B2C — professionals and students prepare for frontier-lab AI certifications with us, and students get free access. Certified — the lab certifies them, not us, so credibility is borrowed instead of claimed. C2B — our certified graduates become the delivery pilots who build and run the bots under our managed service. B — the business gets AI transformation without the jungle, and its staff come back into the loop as learners. See the full B2C2B loop.

🎯 We turn AI learners into certified delivery pilots, and delivery pilots into enterprise results.

A business sale sponsors the students. The professionals who close the skill gap get certified and build the bots the business needs next. Delivery Pilot supports that business as a managed service provider: we secure the bot, adapt it, and train the people around it. The Delivery Pilot template is what makes that sustainable. It holds the software development life cycle together while the business mitigates the risk of AI transformation by staying on the output — the role, the domain, and the topics.

See Pricing MSP Offerings