Custom AI
Custom AI projects, scoped and quoted before work starts
Every product in the Store has a fixed scope and price. For enterprise rollouts, multi-agent teams or anything bespoke, send a Custom AI request and you get a reply by email and, if it fits, a fixed written quote. This route is for projects above $14,900, where the fixed-scope products stop.
- Project size
- Above $14,900
- Price
- A fixed written quote, agreed before any work starts
- Running costs
- Paid by you, in your own accounts, with no markup
- Web app or portal?
- Start with a $3,875 Discovery Sprint
Why businesses choose custom AI
Catalog products solve common problems with a fixed scope. Custom AI is for work that is specific to how your business runs, or that has to reach more people, systems and teams than a single product allows.
It follows your process
The system works to your rules, approval steps, terminology and exceptions, instead of asking your team to change how they work to fit a tool.
It connects the systems you already use
It reads from and writes to your CRM, helpdesk, finance system, document store or phone system, within the access you grant, so information stops being copied between them by hand.
It takes routine work off your team
Rule-based tasks such as sorting, checking, drafting and routing are prepared by the system, so your people spend their time on the decisions that need judgment.
It works across teams
A larger rollout can include single sign-on, role-based access, audit logging and separate staging and production environments, so several teams use it with the right controls.
You keep control of the accounts
AI models, hosting and software run in your own accounts. You see usage and costs directly, and you can remove our access at any time.
You own what we build for you
Once you have paid in full, you own the deliverables we created specifically for you. We keep our own pre-existing tools, templates and generic components.
What custom AI projects look like
Most custom work falls into one of four kinds. Each starts from your process and is scoped before any price is fixed.
An enterprise rollout of a catalog product
Take a product from the store, such as a support agent or a company knowledge hub, and roll it out across several teams with single sign-on, role-based access, audit logging and separate staging and production.
A team of agents working together
Several agents share one workflow, for example research, content, SEO and email, or prospecting, qualification and CRM updates, with clear handoffs and a person approving the steps that matter.
A bespoke workflow across several systems
A process no catalog product covers, such as reading incoming documents, checking them against your rules and posting the result into your finance or operations system for a person to approve.
A custom web app, client portal or internal tool
Software built around your own users and data. These usually start with a paid Discovery Sprint that ends in requirements, wireframes and a fixed quote for the build.
Teams of agents scoped through this route
Show 21 more team examples
Catalog products available as an enterprise rollout
Each rollout adds: SSO, role-based access, audit logging, staging/production and multi-team rollout.
- AI Customer Support Agent
- AI Sales Agent
- AI SDR Agent
- AI Company Knowledge Hub
- AI Voice Receptionist
- AI Voice Customer Support
- AI Content Engine
- AI SEO System
- AI Recommendation Engine
- AI Customer Success Agent
- AI Churn Prevention
- AI Revenue Intelligence
- AI Data Analyst Agent
- AI Business Intelligence Dashboard
- AI Document Extraction
- AI Contract Intelligence
- AI Accounts Payable Agent
- AI Financial Close Assistant
- AI HR Assistant
- AI Resume Screening
- AI Governance System
- AI Agent Security Layer
- AI Gateway
- AI Agent Orchestration System
- AI Evaluation System
- AI Agent Observability
- AI Code Review Agent
- AI DevOps Agent
- AI Multilingual Support
- AI Website Localization
- AI Website Personalization
- AI Executive Assistant
- AI Business Operations OS
- AI Support Supervisor
- AI Travel Concierge
- AI Contract Review Workflow
- AI Corporate Training Agent
- AI Proposal System
- AI Ecommerce Growth System
- AI Market Intelligence
- WhatsApp AI Ecommerce Agent
- AI Internal Search
Problems custom AI suits, and problems it does not
AI does well with repeated work that follows rules and has a person nearby to check it. It is a poor fit where the rules are unclear or the stakes need a professional decision every time.
A good fit
- Work that repeats often and follows rules you can write down.
- Information that already lives in documents, systems or past conversations.
- Processes that cross several systems or teams.
- Tasks where a person can review the output and step in when needed.
- Needs beyond a product's limits on channels, systems, documents or workflows.
Not a good fit
- A problem a catalog product already solves. A fixed-scope AI product costs less and you can order it today.
- Decisions that need a licensed professional, such as financial, legal or medical advice, with no person reviewing them.
- A process nobody has agreed on yet. Map it with an expert first.
- One-off tasks that rarely repeat.
- Goals stated only as a promised number, such as sales or traffic. We scope what the system does, not the results it will produce.
What to plan for before a custom build
These questions shape the scope and the quote. You do not need every answer before you write to us, but each one comes up during scoping. Read our security practices and data processing terms if your team needs them early.
Your data
Which documents, records and past conversations the system should use, who owns them and how current they are. The system can only be as accurate as the information it is given, and we do not write source content you have not provided.
Access
You add our team as a user, or create an API key you control, for each system involved. Credentials are never sent by email, and you can remove our access at any time.
Integrations
Which systems the AI reads from and which it is allowed to change. We agree each connection with you during scoping, and write access is given only where the workflow needs it.
Security
Larger rollouts can include single sign-on, role-based access, audit logging and separate staging and production. Our security page and data processing terms set out how we handle your data.
Human review
Decide where a person approves before the system acts, for example before a refund, a contract change or a message to a client. Approval rules are agreed during scoping and written into the acceptance criteria.
Running costs
AI model usage, hosting and software plans are billed by those providers directly to your own accounts, including during build and testing. Nothing runs through us, so there is no markup.
Support after launch
Optional monthly maintenance keeps the system updated after launch. For AI builds it is priced as a share of the build price, so you can see the cost before you commit.
Scope and a fixed quote
Nothing is built until scope, milestones, acceptance criteria and a fixed price are agreed in writing. New requirements after that are quoted separately in writing before any extra work starts.
How a custom project works
From your first message to a signed quote, everything is agreed in writing. No work starts, and nothing is charged for the build, until you accept a fixed quote.
-
Send the form
Describe the process, the systems involved, who will use it and the result you want. Add the closest catalog products if you know them.
-
We check the catalog first
If an existing product or bundle already covers most of it, we say so. That route costs less and you can order it straight away.
-
We reply by email
With questions, a suggested catalog route, or an invitation to a scoping call.
-
Scoping
We agree scope, milestones, acceptance criteria, access and running costs with you.
-
Fixed written quote
You get a fixed price and timeline in writing. Work starts only once you accept it.
Building a web app or client portal? Start with a Discovery Sprint
$3,875
A paid scoping sprint of about two weeks: two recorded workshops, a requirements document, wireframes for up to 10 key screens and a fixed quote for the build. The sprint price is credited if you go ahead.
See what the Discovery Sprint includesWant the solution designed and written up first?
$2,900
Solution Design and Documentation: discovery sessions, solution architecture, an integration map, a detailed specification, a milestone plan and a fixed quote. Delivered in 5–15 working days after discovery, and credited in full if you sign within 30 days.
Ask for it in your enquiry and we include it in our reply.
Tell us about your project
We reply by email. A few clear sentences about the process, the systems and the people involved are enough to start. Fields marked * are required.Questions about custom AI projects
How much does a custom AI project cost?
Custom projects usually start above $14,900, where the fixed-price products in the store stop. The exact price comes as a fixed written quote after scoping, based on the systems, users, workflows and rollout involved. Running costs such as AI model usage and hosting are separate. They are billed by the providers to your own accounts, with no markup from us.
How long does a custom project take?
It depends on the scope, so there is no standard delivery time for custom work. The written quote includes a timeline and milestones agreed with you before any work starts. If you have a date you need to meet, put it in the form, and we plan the scope and the milestones around it from the first reply.
Do I pay for scoping?
Sending an enquiry is free and we reply by email. If the project fits, scoping leads to a fixed written quote. If you want a full specification before you commit, Solution Design and Documentation costs $2,900 and is credited in full if you sign within 30 days. For a web app or portal, a paid Discovery Sprint is credited if you go ahead with the build.
What if my project is $14,900 or under?
Look in the store first. Many needs are covered by a fixed-scope product, an industry or integration edition, or a bundle, each with a fixed price shown before you buy. If a product covers most of what you need, add a Custom requirements note at checkout and you get a written Change Quote for the rest.
Can a custom project start from a catalog product?
Yes, and it is often the best starting point. Pick the closest products in the form so we can see what you have in mind. Common custom projects include an enterprise rollout of an existing product with single sign-on and audit logging, or several products working together as one team of agents.
Will a person review what the AI does?
Where you want one, yes. During scoping we agree which actions need a person to approve them, such as a refund, a contract change or a message to a client, and who that person is. The system then prepares the work and waits for approval instead of acting on its own. Approval rules are part of the agreed scope.
Who owns the system and the accounts?
The AI models, hosting and software run in your own accounts, and you can remove our access at any time. When you have paid in full, you own the deliverables we created specifically for you. We keep our pre-existing tools, templates and generic components. Our Terms of Use set out the details.
Is my information kept confidential?
We use your enquiry to reply and to prepare a quote, as described in our Privacy Policy. Our Terms of Use include a confidentiality section covering non-public information about your business, customers and systems. Please do not include passwords or API keys in the form. Access is only set up later, by adding our team as a user or creating a key you control.
Have a process the catalog does not cover?
Describe it in a few sentences. We check the catalog first, reply by email and, if it fits, send a fixed written quote before any work starts.