Intelligence / AI agent
Custom AI agent development, one workflow, handled end to end
Pick the workflow eating your week. We build an agent that completes it across your systems, with guardrails and approval where it matters. $4,997 in fourteen days.
- Fixed price
- $4,997
- Ships in
- 14 days
- Quote in
- 48 hours
Code transfers to you. Deposit back if we miss the date.
An agent is different from a chatbot in one important way: it does things. It reads the inbound email, pulls the record from your CRM, decides what category it falls into, drafts the response, updates the deal stage, and books the follow-up. A chatbot would have described that process to you.
That capability is why agents need to be built carefully. Anything that can act can act wrongly, so this build spends as much time on guardrails, approval gates, and audit logging as on the happy path.
We build one workflow at a time, properly. A single agent that reliably handles lead qualification is worth more than five that half-work.
Best for
- Lead qualification and CRM routing
- Support triage, categorisation, and drafted responses
- Order and returns processing
- Research, enrichment, and report generation
- Any process where a person follows the same rules repeatedly
Built with
- Anthropic Claude
- OpenAI
- LangChain
- Next.js
- Supabase
- Vercel
- n8n
What's included
Everything in the $4,997 scope
Written out in full, because a fixed price only means something when the scope is specific.
Workflow mapping
We document the process as it runs today, including the exceptions and the judgement calls, then mark which steps the agent takes, which need human approval, and which stay manual. Automating a broken process just breaks it faster.
Tool and system integration
The agent connects to every system the workflow touches, including CRM, email, calendar, database, Slack, and payments. Credentials are scoped so it can only reach what the job requires.
Reasoning and decision logic
Structured prompting with defined output schemas and validation, so decisions are consistent and machine-checkable rather than free text you have to trust.
Guardrails and approval gates
Spend limits, rate limits, forbidden actions, and human approval on anything irreversible. High-stakes steps queue for a one-click confirmation instead of firing straight into your production data.
Audit logging and observability
Every run is logged: what it saw, what it decided, what it did, what it cost. When something goes wrong you can see exactly where, which is the difference between fixing it and turning it off.
Evaluation and staged rollout
Tested against historic cases, then run in shadow mode alongside your team before it takes over. You approve going live on evidence.
What you receive
The concrete list handed over on ship day.
- A deployed agent running your workflow
- Integrations with every system the workflow touches
- Guardrails, spend limits, and human approval gates
- Full audit log with an admin review interface
- Shadow-mode period before full activation
- Evaluation results against historic cases
- Runbook covering monitoring, pausing, and adjustment
- Thirty days of tuning support
What this tier does not cover
Stated up front. Anything here can be quoted as an add-on.
- More than one workflow, quoted separately
- Multi-agent orchestration, which is the custom tier
- Custom model fine-tuning
- Actions in systems without an API
- Model API usage, billed to your own account
On-time guarantee
Miss the timeline on a fixed-price build and your deposit comes back. No negotiation, no invoicing games.
How it runs
From brief to live in 14 days
The same four phases on every build. No discovery retainer, no phase two.
- 0148 hours
Scope lock
You pick a tier or send a brief. We come back with a written scope, every screen, every integration, a fixed price, and a ship date. Nothing starts until you approve it, and nothing outside it is assumed.
- 02Day 1
Asset handoff
A single onboarding form collects logins, brand assets, and access. One call if the project needs it, none if it does not. We select the stack and start the same week.
- 03The bulk of it
Build in the open
Senior delivery from day one, no junior handoff. You get a walkthrough video at every functional milestone and a live preview URL you can open at any point, so there is no reveal at the end.
- 04Launch + 30 days
Ship and hand over
It goes live on your infrastructure, in your accounts, with the repository transferred. Thirty days of support covers the bugs and questions that only appear once real users arrive.
FAQ
AI agent questions
Answered here so you can decide without a sales call.
What if the agent does something wrong?
Design assumes it will. Irreversible actions require human approval, spend and rate limits cap the blast radius, every run is logged, and there is a kill switch in the admin interface. During shadow mode it proposes without executing, so you see its judgement before trusting it.
How is this different from a Zapier or Make automation?
Zapier follows fixed rules: if this, then that. An agent handles the cases rules cannot: reading an ambiguous email, deciding which of six categories fits, noticing something is unusual. If your workflow is genuinely deterministic, we will tell you to build it in Make for a fraction of this price.
How much does it cost to run?
Usually $50 to $300 a month in model usage, depending on volume and how much reasoning each run needs. We project it at your volume during scoping and route simple steps to cheaper models to hold it down.
Can it work with a system that has no API?
Sometimes, through browser automation, but it is fragile and we price and flag it honestly rather than pretending otherwise. Where a system has no API we will usually recommend changing the system.
How long before it is actually reliable?
Fourteen days to build, then a one to two week shadow period where it runs alongside your team without acting. Most agents reach the accuracy their owner wanted somewhere in that window, and the tuning support covers the rest.
Who owns the agent and its prompts?
You do. Code, prompts, evaluation sets, and configuration all transfer on ship day, running on your API keys and your infrastructure.
Other AI app services
AI chatbot
A support chatbot that answers from your own material and escalates when it should.
- Price
- $1,997
- Timeline
- 10 days
RAG system
Retrieval over your whole document library, with citations and access control.
- Price
- $5,997
- Timeline
- 14 days
Start a ai agent build
$4,997, shipped in 14 days. Send a brief and we will confirm scope and a start date within 48 hours.
No retainer, no hourly billing. Quote in 48 hours.