How it works

A structured process, not a leap of faith.

Every Hire-X deployment follows the same five phases. Each phase has defined activities, a concrete output, and a decision point where you choose whether to continue. Nothing reaches your customers or your data at scale until your team has tested it and signed it off.

The process

The five phases

Each phase has defined activities, a concrete output, and a decision point where you choose whether to continue.

  1. Discovery

    Typically week 1

    We start by understanding the work, not by selling software.

    Activities: a kick-off call to agree scope and success criteria; documentation of the current workflow, including its exceptions; an audit of the tools involved; definition of the role — what the AI employee will do, what it will not do, and who it reports to; mapping of integration requirements.

    Output: a written role definition and integration plan you can review.

    Decision point: you approve the role definition before configuration begins. If Discovery shows the process is a poor fit for automation, we say so, and you can stop here.

  2. Configuration

    Typically week 2

    We build the role to the agreed specification.

    Activities: platform integrations set up and tested end to end; capabilities and limits configured; escalation rules designed for every identified edge case; communication channels connected (Teams, Slack, email); for Custom-Trained deployments, ingestion of your training data.

    Output: a working AI employee in a test environment, plus documented escalation rules.

    Decision point: a configuration walkthrough with your team. Testing does not start until you are satisfied the rules match how you actually work.

  3. Testing

    Typically week 3

    Your team tries to break it before your customers can.

    Activities: an internal pilot on a controlled, representative workload; quality review of outputs against the acceptance criteria agreed in Discovery; deliberate edge-case testing; escalation paths validated with real scenarios; adjustments made from your team's feedback.

    Output: a test report showing output quality against the acceptance criteria, with every issue and its fix listed.

    Decision point: go-live requires your explicit sign-off on the test results. If the outputs are not good enough, we iterate or we stop.

  4. Go-Live

    Typically week 4

    A gradual rollout, not a switch-flip.

    Activities: production workload increased in stages with safety thresholds; real-time monitoring and rapid adjustment through the first live week; training for the human manager and the wider team; dashboard access, reporting, and audit trail set up.

    Output: an AI employee running at full production workload, with your manager trained to supervise it.

    Decision point: you control the pace of rollout. Volume only increases when the previous stage has run cleanly.

  5. Ongoing Optimisation

    Continuous

    Deployment is the start, not the end.

    Activities: weekly performance reports in the first month; monthly reviews with specific recommendations; continuous improvement based on real output data and your feedback; quarterly business reviews for Custom-Trained customers; proactive suggestions when new capabilities become relevant to you.

    Output: a standing record of performance, changes made, and results.

    Decision point: monthly reviews are where you expand, adjust, or wind down. There are no minimum lock-in periods on monthly plans, so continuing is always a choice.

Deployment models

Three ways to work with us

Ready-to-Deploy

£500 – £1,200 / AI employee / month

£1,500 – £3,000 one-time onboarding

Pre-configured AI employees for common roles. The fastest path when your process is close to standard. Includes up to three tool integrations, pre-built role templates, standard escalation workflows, and email support.

Best for: standard roles and common tools.

Most popular

Custom-Trained

£1,200 – £2,500 / AI employee / month

£5,000 – £15,000 one-time onboarding

AI employees trained on your own data, processes, and institutional knowledge. Includes deep business discovery over two to four weeks, custom training on your data, bespoke workflow design, unlimited integrations, a dedicated account manager, priority support, and quarterly business reviews.

Best for: unique processes and proprietary knowledge.

AI Strategy Consulting

Custom, project-based

£5,000 – £20,000 per project

For organisations that want to build the capability in-house. Includes an AI readiness assessment, architecture recommendations, build-versus-buy analysis, change management support, and enablement for your internal team.

Best for: engineering teams planning an in-house build.

Human guardrails

Every AI employee has a named human manager

Each AI employee is assigned to a specific person in your organisation. That person approves high-stakes actions, receives exception alerts, and can override any decision at any time. Oversight levels are agreed in Discovery and configured per task type.

Task type Oversight
Routine, repetitive tasks Fully autonomous, with periodic spot-checks
Customer communications Human approval before first launch
Financial actions Approval required above an agreed threshold (£1,000 by default, configurable)
HR and personal data changes Approval required for every modification
Escalations and exceptions Automatic handoff to the named human manager
Security & compliance

Built for UK businesses with regulated data

UK GDPR and DPA 2018

ICO Registered. Data subject rights, retention policies, and lawful basis are documented for every deployment.

UK data residency

Your data stays on UK-region cloud infrastructure. No cross-border transfers.

Encryption

Data is encrypted at rest and in transit using AES-256. Credentials are held in dedicated secrets management.

Tenant isolation

Every customer runs in a dedicated, isolated environment. Your data is never used to train systems for other customers.

Audit trails

Every action is logged with timestamps, inputs, outputs, and human approvals.

Access control

Single sign-on via your existing identity provider (Okta, Azure AD, Google Workspace), with least-privilege data access — the AI employee only sees the data its role requires.

AI and your team

Augmentation, not replacement

We design deployments to remove repetitive work, not people. The AI employee takes the routine load; your team keeps the work that needs judgement, relationships, and empathy — and stays in control through escalation paths and review checkpoints. We also help you communicate the change to your team, with training and documentation, because a deployment your people distrust will fail regardless of how well it is configured.

Questions

Things worth knowing before you decide

Low-stakes mistakes are caught in the audit trail and corrected during spot-checks. Higher-stakes actions require human approval before they happen, so a consequential error cannot go through without sign-off. When an error occurs, we investigate it, document the root cause, and update the configuration to prevent recurrence.

Yes. You can pause an AI employee at any time, and active work is handed back to your team cleanly. There are no minimum lock-in periods on monthly plans beyond the initial three-month term; after that, cancellation needs 30 days' written notice.

Yes. Every deployment runs in a dedicated, isolated environment with tenant-level isolation at the infrastructure layer. Your data is never commingled with another customer's, and your AI employee is never trained on other organisations' data.

We maintain a data portability policy as standard. All customer data would be returned in machine-readable format within 30 days and then securely deleted. Custom-Trained customers also receive their model configurations and training datasets so they can continue independently.

Through least-privilege access. The AI employee only sees the data its role requires. Sensitive categories — personal data, financial records, commercially confidential information — are identified during Discovery and handled under controls agreed with you.

Yes. AI Strategy Consulting exists for exactly this. Some customers start with a deployed AI employee to move quickly, then transition to an in-house build as their confidence grows.

Typically two to four weeks for Ready-to-Deploy. Custom-Trained deployments include a two-to-four-week discovery phase, so the full timeline is longer. Discovery gives you a realistic timeline for your specific case before you commit.

Start with Discovery.

The first phase is a conversation about your workflows. If the fit is poor, we will tell you before you spend anything on configuration.

Start an enquiry