Scope
A 60-min call to nail down the question the prototype is meant to answer.
AI
AI Prototype.
A working demo wired into your real data — proof an AI feature is worth building before you commit to building it.
A working prototype — agent, RAG bot, or AI feature — wired into your data with a clear demo.
What you get.
Every deliverable below is included in the scoped engagement — no upsell at handoff.
Working prototypeA runnable demo wired into your actual data — not a slide deck and not a synthetic-data toy.
Choice of architectureAgent, RAG bot, classification, or a single embedded AI feature — chosen based on what fits your problem.
Data wiringWe connect to your real source — Postgres, S3, Notion, Drive, SaaS API — under read-only credentials.
Demo walkthroughA recorded walkthrough plus a live demo session with your team and any stakeholders.
Build-or-not memoA 2-page memo on whether this is worth productionising, what the costs look like, and what the risks are.
Source codeRepository handed off with README, environment setup, and notes on what we'd change for production.
How it works.
The same four-step flow we use across every engagement, scoped to this gig.
Wire data
Read-only connection to your real data source within 48 hours.
Build
We assemble the prototype iteratively, showing progress mid-week.
Demo
Live demo, written memo, and a clear yes/no on productionising.
Tools we use.
The stack we default to for ai prototype work. Always open to fitting yours.
Claude
OpenAI
Gemini
LangChain
LlamaIndex
Python
TypeScript
Why work with us on this.
Three reasons clients pick Hepha Works for ai prototype.
Senior practitioners only
The person who scopes the work is the person who delivers it. No invisible subcontractors, no junior handoffs.
Written scope, fixed price
You see a written scope and a number before any work starts. No timesheet surprises, no scope-creep arguments.
Honest read on outcomes
We won't say it'll work if we don't think it will. If the gig isn't right for your situation, we'll tell you that on the call.
Frequently asked.
The questions we get most before kicking off ai prototype engagements.
What does "real data" mean?
Your actual production data, with read-only credentials and any necessary anonymisation. Toy datasets miss the failure modes that matter.
Can you sign an NDA?
Yes — we sign mutual NDAs before any sensitive data is touched. We can also work entirely on your infrastructure if required.
What if the prototype reveals the idea isn't worth building?
That's a successful outcome. The build-or-not memo will say so plainly, with reasoning.
Which model should I use?
We'll recommend based on cost, latency, and quality for your specific task. Often it's not the model you'd guess.
Can this turn into a full build?
Yes — the natural follow-on is the LLM App / Agent Build gig or a Custom Project engagement.
Related engagements.
Other gigs that pair well with AI Prototype.
Ready to start?
Send a brief and we'll come back with a written scope and a number within 1–2 business days.