LLM integration · retrieval over internal documents · agents · evaluation
We build AI into the systems a company already runs — with the sources attached, the failure cases measured, and a description of the data flow that your security and compliance people can actually read. If a task is better solved without a model, we say so in the framing stage.
A model is the smallest part of the work. The rest is the seam: your ERP, CRM, ticketing or document store, the permissions people already have, and the places where a wrong answer costs money.
Typical tasks
Answers grounded in your own corpus — policies, contracts, manuals, tickets. We build the ingestion, the chunking, the retrieval and the evaluation set, and we show retrieval quality separately from generation quality.
Typical tasks
Multi-step work with tools: reading a queue, calling internal APIs, filling forms, escalating to a person. Every step is logged and reversible; nothing irreversible happens without a human decision.
Typical tasks
A model that answers fluently is not a model that answers correctly. We build the test set from your own hard cases, measure retrieval and answer quality separately, and keep the numbers visible after every change.
Typical tasks
Where the data lives, what leaves the perimeter, what is retained and what is logged. On-premise or in your own cloud account where the requirement demands it, with a data-flow description your compliance team can review.
Typical tasks
Models and documents change. We hand over with documentation and monitoring, or stay on for a fixed monthly engagement that includes evaluation re-runs and incident handling.
Typical tasks
Lengths are the usual range for a single process with a cooperative data owner. They are not a quote: the framing stage exists to replace them with a real plan for your case.
01
One process, written down as it happens today, with the volumes, the exceptions and the cost of a mistake. Ends with a written scope and a decision on whether AI is the right tool at all.
Needed from you
Process owner, two or three real examples, a list of systems involved.
02
Where the documents live, who may see what, what may leave the perimeter, which model providers are acceptable and what the auditors will ask for.
Needed from you
Access to a representative sample, security contact, any existing policy.
03
A narrow, real slice in production conditions: retrieval, prompting, evaluation set, UI in the tool people already use. Ends with measured quality on your own cases.
Needed from you
A named user group, test environment, one person who can approve access.
04
Failure cases from the pilot become the regression suite. Permissions, logging, rate limits, refusal paths and hand-over to a human are tightened.
Needed from you
Feedback from the user group, security review, decision on hosting.
05
Wider release, training for the people using it, monitoring and alerting, documentation for whoever runs it after us.
Needed from you
Rollout plan, support contact, agreement on what is measured monthly.
06
Fixed monthly engagement: monitoring, re-evaluation after model or document changes, and a named engineer who answers.
Needed from you
Nothing beyond the retainer decision.
Models
Models from the main providers, and open-weight models hosted in your own environment when data may not leave it. The choice follows the constraint, not the other way round.
Retrieval
Vector and keyword search over your documents, hybrid ranking, permission-aware filtering, citation back to the source page.
Orchestration
Explicit pipelines and tool-calling agents with step logging, retries and a human hand-over point.
Evaluation
Labelled question sets from your own operation, retrieval and answer metrics tracked per release, failure cases kept as tests.
Integration
APIs and web services against your existing systems; no rip-and-replace, no second copy of your data where it can be avoided.
Delivery
Containerised services in your cloud account or ours, infrastructure as code, environment separation for test and production.
Data handling in one paragraph
Your documents and records stay in the environment you choose: your cloud tenant, your own servers, or ours if you prefer and the classification allows it. We describe what leaves the perimeter, keep the audit trail of prompts, retrievals and outputs, and do not use your data to train anything. Where a requirement forbids external model calls, we run open-weight models inside your perimeter instead — with the accuracy trade-off stated in writing.
Two to four weeks to turn an idea into a written scope, a data inventory and a go/no-go decision. Fixed price, no commitment afterwards.
fixed fee, quoted per scope
One process, one user group, measured quality on your own cases. Ends with a working slice and the numbers to decide about production.
fixed fee per stage
Hardening, permissions, monitoring, rollout and documentation, delivered in stages with a demo at the end of each one.
fixed fee per stage
Monthly engagement after launch: monitoring, re-evaluation, incident handling and small changes.
monthly retainer
The scope and the price of a stage are agreed in writing before it starts. Prices depend on the number of systems, the state of the documents and the hosting constraint — a discovery sprint is what turns that into a number we can stand behind.
Tell us what has to change, which systems are involved and who would use the result. If an NDA is needed before details are shared, say so in the form and we will sign yours or send ours first.
Discuss a project
Tell us what has to change in the process, what systems are involved and who would use the result. We reply within one business day; an NDA can be signed before any detail is shared.
Prefer email? hello@northboundai.com · or call +1 (555) 010-0600.