Chris Bentley, Rick Mills & Philip Maynard | Fractional AI, Finance & Technology

Q: Do you advise, or do you implement?

Implement. Advice is part of it, but the engagement is judged on what is running afterwards. If all you need is a second opinion, say so and it stays short.

Q: How do you handle security and access?

Least privilege, named accounts, and access revocable the day an engagement ends. Credentials live in your secret store. Systems are built in whatever environment best fits your security, ownership, cost and maintenance requirements. Where AI models process your data, you are told which provider, what leaves your environment, and the retention terms before anything is connected.

The team

Three operators. Three different lanes.

Each of them has run this work inside a real business, not advised on it from outside. Engage one, or engage the combination the problem actually needs.

Chris Bentley

AI & Digital Operations

Turns disconnected systems, manual workflows, and unreliable data into durable business operations.

Bring Chris in when…

Your tools do not communicate, reporting takes too long, manual work is multiplying, or an AI prototype needs to become a reliable operating system.

What he does

AI agents and workflow automation
E-commerce and digital operations
Data, analytics, and reporting
Systems and API integration
Cloud infrastructure and security
Technology and vendor strategy

AI, specifically

Scheduled agents that run a routine on their own and report by exception
Document and data extraction feeding systems that already exist
Reporting that assembles itself from the source systems each morning
Prototype to production — monitoring, rollback, and an owner after launch

Example engagement

Problem: A personal AI assistant had to run continuously for a CEO who will never open a terminal.
Owned: The machine, cloud accounts, and security posture it runs on; the operating constraint the whole design had to satisfy.
Built: Hardened single-host deployment, hourly self-checks that reapply the system's own patches, one-message rollback for every feature.
Result: Running continuously since February 2026 — see the case study.

[NEEDS VERIFIED EXAMPLE]

Add one client engagement outside Headsets.com, with: initial problem · what Chris owned · what was implemented · business result · ongoing role. Publish only with the client's approval, or anonymise the client and keep the numbers exact.

Talk with Chris LinkedIn [NEEDS VERIFIED URL]

Rick Mills

Fractional CFO & Operations

Builds finance and operating disciplines that give owners timely, decision-ready numbers and processes that can scale.

Bring Rick in when…

The books close too slowly, leadership does not trust the numbers, ownership lacks visibility, or growth has outpaced the company's operating discipline.

What he does

Close and reporting cadence
Forecasting and management reporting
Finance process redesign
Operational planning and accountability
Controls and workflow documentation
AI-enabled reporting and finance automation

AI, specifically

Close checklists that chase themselves — open items surfaced daily, by owner
Management reporting drafted automatically, then reviewed by a human before it goes out
Variance explained in writing against the plan, not just flagged in a cell
Transaction and document coding assisted, with the exceptions routed for approval

Example engagement

[NEEDS VERIFIED EXAMPLE]

Initial problem: …
What Rick owned: …
What was implemented: …
Business result: …
Ongoing role: …

Also confirm the market-facing title. Internal records describe Rick's current lane as IT alongside Chris; this page markets him as Fractional CFO & Operations. Confirm which is correct before launch.

Talk with Rick LinkedIn [NEEDS VERIFIED URL]

Philip Maynard

AI Products & Technology Operations

Builds and owns practical AI and technology used by customers and frontline teams.

Bring Philip in when…

You want AI inside the customer journey, need a working internal agent, or need someone to own implementation after the demo.

What he does

Customer-facing AI agents
Customer service and support automation
Internal AI tools and workflows
Agentic AI implementation
Technology operations and adoption
Ongoing testing, iteration, and ownership

AI, specifically

A support agent that answers from your product data and hands off cleanly when it should not
Product guidance in the buying flow — the questions a good salesperson would ask
Internal agents your team uses daily, with the escalation path designed in
Measured after launch — deflection, escalation, and what it got wrong

Example engagement

[NEEDS VERIFIED EXAMPLE]

Initial problem: …
What Philip owned: …
What was implemented: …
Business result: …
Ongoing role: …

Nothing about Philip's history is on file here. Current title, tenure, and at least one shippable example are all needed before launch.

Talk with Philip LinkedIn [NEEDS VERIFIED URL]

All three work with AI. Their primary operating lanes are different. Chris makes the operation run. Rick makes the numbers trustworthy. Philip puts AI in front of customers and frontline teams. Most engagements start with one of them.

[NEEDS APPROVED HEADSHOTS]

Drop three square JPGs into /assets/team/ as chris-bentley.jpg, rick-mills.jpg, philip-maynard.jpg (600×600 or larger). They appear automatically and the monogram fallback disappears. Until then each card shows initials, which the brief rules out for launch.

Where we help

Start with the sentence you would actually say.

Most of these conversations begin as a complaint, not a brief. Find yours, and it points at the person who should pick up the phone.

“We have too much manual work.”

Workflow automation
Integrations between systems that do not talk
Reporting that builds itself
AI agents for the repetitive middle of a process

Usually → Chris Bentley

“We cannot trust or access our numbers quickly.”

Close process and cadence
Forecasting
Management reporting
Finance operations and controls

Usually → Rick Mills

“We want AI in the customer experience.”

Customer-facing agents
Support automation
Product guidance and buying help
Internal escalation workflows

Usually → Philip Maynard

“Our systems work, but nobody truly owns them.”

Fractional technology leadership
Infrastructure
Security
Vendor and platform management

Usually → Chris Bentley, often with Philip

“We are not sure where AI is actually worth using.”

Workflow discovery
Opportunity prioritisation by value, not novelty
One small pilot, in production
Measurement and ongoing operation

Start with Help me choose

Why this team

We already work together, inside a real business.

Not a network of strangers assembled per project. The same three people, with a working history and a shared standard for what counts as finished.

Operators, not presentation-only consultants

Everyone here has been accountable for a system after it went live, on a day it broke. That changes what you build.

Strategy and implementation from the same people

The person who scopes it is the person who ships it. Nothing gets handed to a delivery team who were not in the room.

AI where it creates value; ordinary process work where it does not

A lot of “AI problems” are a missing integration, an unclear owner, or a process nobody wrote down. We will tell you when that is the case.

Systems designed to stay understandable

Configuration your team can read and change. Documentation that matches what is running. No component only one person can maintain.

One specialist, or a coordinated team

Start with the person whose lane the problem sits in. Pull in the others when it turns out to cross lanes — which it often does.

Range across the whole operation

Technology, finance, customer operations, data, and infrastructure. Enough coverage that the seams between them do not become your problem.

Case study · Smarty

Proof that we operate what we build.

Smarty is a personal AI chief of staff built for the CEO of Headsets.com and kept running by this team. It is one case study, not a product for sale — but it is the clearest evidence of how we work: shipped, monitored, repaired, and still running.

30+Scheduled routines running every day without anyone starting them
5Interaction channels — text, phone, browser voice, email, iMessage
∞Persistent memory and live business context carried between every conversation
↯Voice-triggered research and task dispatch — ask on a call, get the answer back as a message
✓Monitoring, reversibility, and self-repair — every feature has a one-message undo

What it does on a working day

06:30Morning briefing, before the day startssent
07:45Meeting preparation for the calendar ahead1 brief
09:15Operational monitoring across the systemsok
11:00Inbox prioritisation — what actually needs a reply6 flagged
15:30Follow-up tracking on open commitments3 chased
16:00Task dispatch to the people who own themsent

The same pattern — scheduled work, exception reporting, a human in the loop for anything consequential — is what we build for companies.

Explore how Smarty works →

Production principles

How we build AI that survives the demo.

The demo is the easy part. These six rules are what separate a system still running next year from one that quietly stopped and nobody noticed.

01

Configuration is understandable and editable

Thresholds, schedules and rules live where a person can read and change them — not buried in code that needs a developer and a deployment.

02

Automations are reversible

Every automated behaviour can be switched off and switched back on cleanly. If turning something off is a project, it should not have been turned on.

03

Failures are monitored and repaired

Something watches the automation and tells someone when it stops. Known failure modes get an automatic repair, so the same outage does not happen twice.

04

Missing data is disclosed, not guessed

When a source has not updated, the output says so and names the gap. A system that quietly fills holes with stale numbers is worse than no system.

05

Consequential actions need a human

Money, customer commitments, anything hard to reverse: drafted by the system, approved by a person. Automate the preparation, not the decision.

06

Every system has an owner after launch

Named, with a documented handover, and either your team or ours. Ownership is decided before go-live, not discovered during an incident.

Mike Faith Founder & CEO, Headsets.com
Nashville, Tennessee

“I described it. They made it hold.”

Mike founded Headsets.com in 1997 and still runs it. He uses this team's work inside the business every day, and trusts them with systems, operations, and decisions that matter. Smarty is his; he is the customer here, not the vendor.

How we work

A small first engagement, then earn the next one.

No discovery phase that produces a deck. The first piece of work is scoped so you can judge it on a result.

Start with the business problem

What is slow, unreliable, or costing people's time. Not which technology you should adopt.

Identify the right expert or combination

One lane, or two working together. We will say if the problem is not ours to take.

Define a small, concrete first engagement

A defined scope with a defined outcome, sized so it can be judged quickly.

Implement — not merely recommend

The thing gets built, connected, and put in front of the people who will use it.

Measure, maintain, and improve

Watch it in real use, fix what it gets wrong, and decide together what happens next.

Model 01

Project

A defined outcome with a start and an end. Best when you know what needs to exist and need someone to build it properly.

Model 02

Interim leadership

Someone in the seat while you hire, restructure, or recover. Full accountability for the function, for a defined period.

Model 03

Ongoing fractional

A recurring share of a senior operator's week. Best when the work is real but does not justify a full-time executive.

Questions

The things people ask before the first call.

What does “fractional” actually mean?

You get a senior operator for a share of their time instead of a full-time hire. In practice that is anything from a few days on a defined project to a recurring day or two a week for as long as it is useful. You are buying judgement and delivery, not headcount, and you can change the shape as the work changes.

Can we hire just one person?

Yes, and most engagements start that way. Pick the lane your problem sits in and talk to that person directly. If it turns out to cross into someone else's lane, they will tell you and you can decide whether to bring them in.

Do you advise, or do you implement?

Implement. Advice is part of it — you will get a clear opinion on what is worth doing and what is not — but the engagement is judged on what is running afterwards. If all you need is a second opinion, say so and we will keep it short.

Who owns the systems and work product?

You do. Code, configuration, documentation and data belong to your business, in your accounts and repositories wherever possible. Nothing is built to be hostage to a relationship with us, and handover documentation is part of the work, not an extra.

[CONFIRM] Exact IP terms to match the engagement agreement once the contracting structure is settled.

How do you handle security and access?

Least privilege, named accounts, and access that can be revoked the day an engagement ends. Credentials live in your secret store, not in a document or a personal machine. Systems are built in whatever environment best fits your security, ownership, cost and maintenance requirements — your cloud, your tenancy, or hardware you own.

Where AI models process your data, we tell you which provider, what leaves your environment, and what the retention terms are before anything is connected.

How does an engagement begin?

One conversation, usually about 30 minutes, about what is actually going wrong. If we are the right people, you get a written scope for a small first engagement with a defined outcome and a defined cost. If we are not, we will say so on that call.

Contact

Tell us what needs fixing.

One form, routed to whichever of the three should answer it. You do not need to know who that is — describe the problem and we will work it out.

Talk with ChrisAI & Digital Operations Talk with RickFinance & Operations Talk with PhilipAI Products & Technology Help me chooseDescribe it, we will route it

[NEEDS VERIFIED COMMERCIAL STRUCTURE]

This section must state plainly who a client signs with — each individual independently, a shared entity, or Headsets.com — before the site goes public. Leaving it unstated is what the brief specifically rules out. Once decided, replace this block with the plain-language answer and update the FAQ on ownership to match.

Name

Work email

Company

What needs fixing?

Preferred area

Engagement optional

Goes to the person whose lane this sits in. No list, no sequence, no newsletter.