A hands-on half-day for Synapxe staff. No coding required. You leave with a working assistant and a design for your own team.
A short self-paced pre-read, then one hands-on half-day. You go from what is an agent? to building one, and finish with a design for a workflow from your own team.
How AI Works: five short interactive explainers, sent before the session.
Chatbots to agents, then prompting, then build the Guideline Assistant, then design your own agent.
A working assistant, a reusable CRAFT prompt and an Agent Design Canvas for your own work.
Do these before the session, about 10 to 15 minutes each, so live time goes on building.
Documents, correspondence and operational workflows
Reports, vendor assessments and compliance
Policies, meetings and cross-department communication
Leave, onboarding and staff communications
Amazon Quick is AWS's no-code platform for building AI assistants, dashboards and automations. Amazon Bedrock is the engine underneath; you work in Quick, not in Bedrock.
The files and data an agent is allowed to read
A saved, reusable assistant you describe in plain English
Multi-step automations on demand, on a schedule or on an email
Plans its own investigation and returns a cited report
Everything you will open, in the order you will open it. Six modules, about four hours. The two starred modules are the hands-on core.
Goals, a quick poll, then sign in and find your way around.
Why a chatbot hits a wall, what makes an agent, and how much autonomy to hand over.
CRAFT: Context, Role, Action, Format, Trust. The five slots your F3 agent's instructions are written in.
A Space with one guideline, a Chat Agent built from a plain-English brief, a cited answer, and an honest "not covered" when the document is silent.
Design an agent for a task in your own function. Start from one of nine real Synapxe workflows, or bring your own. The canvas is what you take away.
Recap, pick the one workflow to try first, and keep one habit: check the citation before you act.
Every page is also in the menu button, top-left.
| Time | Module | What happens |
|---|---|---|
| 09:00–09:20 | F0 · Welcome & getting into Amazon Quick | Goals, a quick poll, sign in, find your way around, meet the shared use case |
| 09:20–10:00 | F1 · From Chatbots to Agents | Why chatbots hit a wall, what makes an agent, levels of autonomy |
| 10:00–10:40 | F2 · Prompting your agent | CRAFT: write your agent's instructions |
| 10:40–10:55 | Break | Start your lab here, not earlier |
| 10:55–11:55 | F3 · Build the Guideline Assistant | Space, then Chat Agent, then a cited answer, then a generated briefing document |
| 11:55–12:45 | F4 · Agent Design Canvas | Design an agent for your own team's workflow and take away a one-page canvas |
| 12:45–13:00 | F5 · Wrap-up & what's next | Recap, pick your first workflow, Q&A |
For all staff, no technical background needed. By the end you can:
F3 builds on a document every team has had to interpret, so the hour goes on technique rather than subject matter. You set up a Space with one synthetic guideline, create a Chat Agent whose brief carries its own honesty rules (answer only from the document, cite the section, say "not covered" if it is not there), then compare a covered question with an uncovered one. The lesson: an agent grounded in your documents, not the open internet.
F4 uses that build as the template. You choose from nine workflows drawn from real Synapxe work, or bring your own, and the canvas captures purpose, steps, capabilities, guardrails and the business case. It downloads as Markdown and is the artefact you leave with.
Not part of the live run. Open these to go deeper, or afterwards as a refresher.
A hands-on follow-up for staff who identified a promising use case and want to take it further. Where Foundations builds awareness and a first idea, this Deep Dive is about refining and implementing it — deeper agentic patterns, a robust agent design, a full build with verification, and a clear path to hand over to the DNA / IT team.
Half day — morning or afternoon
For a subset who want to mature a use case
Continue from Foundations, or bring a new problem
By the end, participants can:
| Time | Module | Description |
|---|---|---|
| 09:00–09:15 | D0 · Welcome bring your use case | Frame your use case — continue from your Foundations canvas or bring a new problem |
| 09:15–09:55 | D1 · Agentic workflow patterns | All four patterns (Chaining · Parallelisation · Routing · Orchestration) — matched to your use case |
| 09:55–10:35 | D2 · Instructions & knowledge | Hands-on: write the agent's instructions and set up the Space and reference documents it needs — guardrails, escalation, what NOT to do |
| 10:35–10:50 | Break | |
| 10:50–11:45 | D3 · Build & verify (capstone) | Build a working assistant on Amazon Quick; the deliberate-error verification / trust-but-verify moment; iterate |
| 11:45–12:15 | D4 · Agentic Automation (design) | Concept + design: triggers, approval nodes, human-in-the-loop → produce an automation plan (Markdown) |
| 12:15–12:50 | D5 · Use-case refinement Agent Design Canvas | Full canvas: value / feasibility / risk, success metrics, and the path to implementation |
| 12:50–13:00 | D6 · Showcase & next steps | Share designs, implementation roadmap, and how to engage the DNA / IT team |
Timings are adjustable — use the Session time switch above to see the same agenda as a morning or an afternoon run.
Anyone who prefers a ready-made starting point can adopt one of these instead of their own idea.
Procurement / Finance — extract terms, compare vendors, flag risks, recommend
Governance — check a document against a policy or circular, cite the clause
HR — apply the rules in a staff guideline to a request and draft the response
Finance — read the numbers, explain the variance, draft a memo
On Amazon Quick, using a Space, a Chat Agent, and reference documents:
The shift from an assistant you invoke to an agent that runs on a trigger or a schedule — the L1 → L2 → L3 autonomy ladder. Three questions frame any automation: What triggers it? · What runs? · Where does a human stay in the loop?
The Deep Dive is built around your use case, so most work happens on Amazon Quick and in your own documents. The canvas below anchors D2, D5, and the implementation plan.
One copy-paste block to keep open during the session and after it. Fill the five slots and you have a prompt worth acting on — and the instructions for a Chat Agent. Worked examples and the full annotated version live in the CRAFT exercises.