Skip to content
The AI workforce

Don't start with the AI. Start with the org chart.

The hybrid workforce does not begin with a model or a framework. It begins with an honest look at your organisation — which roles are hiring, which are churning, which are drowning in procedure — and then putting an AI worker where the pressure already exists.

Lens 1

Where are the people?

Walk your org chart by function, department and role type. Wherever a function holds a lot of people and there is clear room for improvement, start there. Inside that function, find the work that is predictable, documented and repeated — and the work you would otherwise be hiring for.

In a typical partner business, the concentration is here
  • Service desk and NOC — highest headcount, highest repetition
  • Engineering and project delivery — scarce, expensive, easily consumed by presales support
  • Account management — absorbs QBR prep, renewals and reporting
  • Operations and billing — reconciliation and provisioning drift
  • Marketing and sales — small team, large expectations
Lens 2

Where are you already hiring?

Open reqs, attrition and backlogs are the business telling you where capacity is short. That is precisely where an AI worker joins fastest — the need is already funded and understood.

Signal 01

Open positions

Headcount budget already exists. An AI worker can take part of the req, or all of it.

Signal 02

High attrition

The role churns and knowledge walks out the door. Continuity is what AI is good at.

Signal 03

Procedure-driven work

If the steps are written down, the role can be described.

Signal 04

Interaction and data entry

Email, chat, phone calls and entering data into systems — native territory.

The roster

Seven roles, mapped to a partner business.

Each worker is a multi-agent, multimodal system governed by a job description and managed by a person on your team. The technology is plural. The responsibility is singular.

Worker Channel role What it does People remain responsible for
Nick Channel marketing & demand generation Turn approved expertise and source material into organised content and campaign assets. Editorial direction, source accuracy, brand strategy, sensitive claims, final approval.
Jules AI business development rep Build account context and coordinate personalised outbound within approved sales policies. Account strategy, messaging and channel policy, relationship judgment, qualification, sales ownership.
Pepper AI lead qualification & response Respond to inbound interest, collect context, support qualification and coordinate the next step. Qualification policy, sensitive conversations, commercial commitments, exceptions, opportunity ownership.
Tony AI solutions & presales engineer Bring approved technical and licensing knowledge into discovery, demonstrations, options and follow-up. Architecture, complex judgment, security representations, pricing and ROI approval, deal ownership.
Joy AI deal desk & sales operations Connect meetings, records, follow-up, subscription state, risk visibility and document coordination. Deal strategy, relationship ownership, forecast judgment, pricing, negotiation, contractual commitments.
George AI customer success & renewal manager Keep onboarding, customer context, engagement, health, expansion and renewal preparation visible. Customer relationship, success strategy, sensitive escalations, commitments, renewal and expansion decisions.
Mike AI service desk & support assistant Triage requests, resolve approved common issues, gather context and improve escalation quality. Support policy, complex diagnosis, high-risk actions, customer-impacting decisions, incident ownership.
One governed job description per worker

Every worker is defined the same way you would define a role: responsibilities (what it owns), boundaries (what it may never do without approval) and measures (how performance is judged). That document is what makes an AI worker governable rather than a novelty.

In detail

How these roles behave inside a partner business.

Each block is the same job description, translated into channel work.

N Nick
Channel marketing / demand generation

Turn approved expertise and source material into organised content and campaign assets.

Work may include
  • Develop and repurpose content from approved sources
  • Organise assets around audiences, themes and campaigns
  • Prepare work for review and publication
Evidence to require
  • Source-to-asset workflow
  • Review process
  • Approved portfolio
  • Quality measures
J Jules
AI business development rep

Build account context and coordinate personalised outbound activity within approved sales policies.

Work may include
  • Research approved accounts and contacts
  • Prepare outreach around an approved hypothesis
  • Maintain follow-up state and escalate replies
Evidence to require
  • Research-to-message workflow
  • Approval controls
  • Personalisation quality
  • Verified engagement
P Pepper
AI lead qualification & response

Respond to inbound interest, collect context, support qualification and coordinate the next step.

Work may include
  • Answer approved common questions
  • Collect need, timing and routing context
  • Schedule, hand off and record the interaction
Evidence to require
  • Conversation demonstration
  • Qualification logic
  • Human handoff
  • Response measures
T Tony
AI solutions & presales engineer

Bring approved technical and licensing knowledge into discovery, demonstrations, buyer questions and follow-up.

Work may include
  • Gather technical discovery context
  • Prepare approved demonstrations and responses
  • Organise requirements, objections and open questions
Evidence to require
  • Discovery-to-demonstration workflow
  • Approved knowledge use
  • Escalation path
  • Outcome evidence
J Joy
AI deal desk & sales operations

Connect meetings, records, follow-up, subscription state, risk visibility and document coordination across the deal.

Work may include
  • Track milestones and next actions
  • Prepare follow-up and maintain approved records
  • Surface missing information, stalled activity and risk
Evidence to require
  • Meeting-to-follow-up workflow
  • Record controls
  • Risk example
  • Continuity measures
G George
AI customer success & renewal manager

Keep onboarding, customer context, engagement, health, expansion and renewal preparation visible.

Work may include
  • Coordinate approved onboarding steps
  • Organise goals, commitments and health context
  • Surface risk, opportunity and incomplete actions
Evidence to require
  • Onboarding or health workflow
  • Human escalation
  • Review experience
  • Approved success measures
M Mike
AI service desk & support assistant

Triage requests, resolve approved common issues, gather context and improve technical escalation.

Work may include
  • Classify issue, priority, customer and context
  • Use approved knowledge and runbooks
  • Update the record and escalate with structured evidence
Evidence to require
  • Intake-to-escalation workflow
  • Runbook use
  • Human review
  • Approved service measures
Beyond the templates

Seven templates. Not seven limits.

The predefined roles are the on-ramp, not the catalogue. The real capability is the process: we have mastered how AI workers get built, the building blocks exist, and we create roles unique to your business, tools and processes — deployed in your environment.

  • Subscription & Renewal Sentinel
  • Channel Revenue Assurance Agent
  • Microsoft Partner Operations Agent
  • Google Partner Operations Agent
  • Employee Lifecycle Agent
Design a custom role
How we govern this

Reports to a human manager. Always.

A worker is a multi-agent, multimodal system governed by a job description and managed by a person on your team. Its authority depends on the role, the evidence, the risk and your approved control model.

Multi-agent

Research · drafting · data and records · quality check · coordination.

Multimodal

Email · chat · voice · documents · your systems of record.

Job description

Responsibilities · boundaries · measures.

Your environment

Your cloud, identity, governance and data controls.

Before you pick the first role

Common questions.

Can we begin with one worker?
Yes — that is the recommended path. One focused responsibility inside one function is the clearest way to create evidence before expanding.
Do we have to pick from the seven reference roles?
No. The reference roles provide a starting pattern, but they are not the limit. We design custom workers around your business, tools, controls and processes using the same building blocks.
Why start on the revenue side?
Because that is where the signals are usually strongest in a partner business: measurable outcomes, interaction-heavy work, documented process and standing headcount pressure. Nothing prevents starting in service delivery or operations when your signals point there — and for many MSPs they do.
Do we need the complete roster?
No. Additional workers make sense only when their responsibilities create incremental value and your organisation is ready to manage them. Buying seven workers at once is how AI programmes fail.
Where do the workers run?
In your environment: your cloud, your identity, your governance and infrastructure. Deployment specifics, access and ownership are defined per engagement.
Are the workers autonomous?
Authority depends on the role, evidence, risk and approved control model. Every worker reports to a human manager, and autonomy expands only as performance earns it.
Run the scan

Which function has the people, the pressure and the process?

Bring your org chart and your open reqs. We will run the two-lens scan with you, match the closest role, and define the first job description an AI worker should hold.