Build your first agent with us.
Your assessment report identifies the agent opportunities for your business - what each agent would do, which jobs it handles, what it would save you. The Agent Build is where we scope it, build it, and hand it over ready to run.
Already a client? Email [email protected] to book a scoping call.
How it works
Assessment first
Complete a HoursBack Assessment. Your report identifies two to four agent opportunities tuned to your business - what each one would do and what it would save. The Agent Build picks one and ships it.
Scoping call
A 30-minute call to lock the agent's job, inputs, outputs and tone. You share 2-3 example outputs you'd accept.
Build and iterate
Ten business days of build and test in our workspace. You sign off five test runs before anything ships.
Handover
We send the export file, the handover document, and book the 30-minute walkthrough. You import, run, and own it.
What you get
Everything below is included from the £2,500 starting price. No add-ons, no tiered shopping list - final price is confirmed after a short scoping call so you know exactly what you’re paying for.
One built-and-loaded agent
A Claude Project or Custom GPT configured to do one specific job - scoped from the agent opportunities identified in your Assessment report. You choose the format at the scoping call.
A knowledge base loaded with your data
Your documents, templates and samples, loaded and tuned so the agent reads your material the way you would.
A portable export file
Delivered to your inbox as a file you import into your own Claude Pro or ChatGPT Plus account in under five minutes.
A plain-English handover document
Two pages covering what the agent does, what good input looks like, where it needs your judgement, and how to tweak it as your business changes.
A 30-minute handover call
We walk you through the agent on Zoom, run two live tasks together, and answer every setup question while we're on the line.
Five signed-off test runs before handover
You see the agent's outputs against real inputs and approve them before anything ships. If they miss, we iterate until they match, no extra charge.
A 30-day support window
One tweak pass included (up to three hours) so you can adjust tone, format or knowledge base once you've lived with the agent for a few weeks.
Aaron Leary
Founder & CEO, Bedrock Learning
Client result
Splendidly finished. The Account Management Team have had their minds blown. 197 reports done in 60 seconds - would have taken us 10 weeks manually. Genuine game changer.
What’s not included
Scope is deliberately narrow. Everything below is a separate product, a separate engagement, or a no.
Live account access
We never log in to your Claude Pro or ChatGPT account. The build happens entirely in our workspace.
Teaching you to build agents
Agent builds are not a DIY job. We don't teach customers to build them. If you want hands-on coaching on the simpler items in your report (automations, prompt setup, tooling fixes), the £349 Implementation Kickstart is the right product.
Ongoing maintenance past the support window
After 30 days, further tweaks are a fresh scope or you handle them using the handover document.
Third-party integrations
The delivered agent runs inside your Claude or ChatGPT account. Custom connections to CRMs, job-management tools or accounting systems are a separate engagement.
Agents handling regulated personal data
Health records, financial advice, safeguarding. If that's the use case, we'll say so at the scoping call and either reshape the scope or decline the build.
We never ask for your login.
You import a portable export file into your own Claude Pro or ChatGPT Plus account on your own login. The build happens entirely in our workspace using synthetic data and anonymised samples you supply.
- We build in our own Claude and ChatGPT workspace. You never share a password, session cookie, or two-factor code.
- We train on synthetic data plus anonymised samples you send us. You decide what “anonymised” means for your business.
- You receive a portable export file. You import it into your own account in three clicks and own the agent from that moment.
- 30 days after handover, we delete the samples you supplied from our workspace unless you’ve given written consent for case-study use.
Who it’s for
Best fit if you’ve finished a HoursBack Assessment and your report flagged an agent opportunity. Agent builds are not a DIY job and we don’t teach you to do them. If you want to start with the simpler wins from your report (automations, prompt setup, tooling fixes), the £349 Implementation Kickstart is the right product.
Want to see a real agent?
We dogfood our own method. HoursBack’s marketing runs through an agent we built the same way we’ll build yours.
Our marketing strategy agent, built the same way we’ll build yours. Loaded with our positioning docs, ICP notes, and brand voice rules. It drafts briefs, critiques copy, and keeps the tone consistent across every piece of content.
Start with the Assessment.
The Agent Build is a post-assessment upgrade. The Assessment is where we figure out which agent is worth building first. Book that, then we talk about the build.
Book an assessment first →Already a client? Email [email protected] to book a scoping call.