AI consulting that starts with how your business runs.
QS Digital helps owner-led businesses decide where AI fits and then builds it. Every engagement follows the same spine: Diagnose the problem, Map the right system, Build it in accounts you own, and Install it so your team runs it without us. What changes from client to client is depth and scope, not the structure.
Where should this business start, and what will it cost?
We start from your inquiry answers, research your business and the problem, and run one working session. The output is a written plan: where you are today, the opportunities ranked by impact and effort, a recommended first step, and realistic budget ranges.
You receive
A written plan you can act on with or without us
A recommended starting point and why
Budget ranges for the next phases
We need from you
Answer the follow-up assessment before the session
Bring the person who will own the work day to day
Share the systems and data the work would touch
02
Map
Exactly what will we build, how, and for how much?
For the chosen starting point we write the scope: specifications, the accounts you will own, data rules, consent needs, acceptance criteria, and a fixed quote. For simple scopes this folds into the proposal. For complex or multi-department work it is its own step.
You receive
A scope your team can read
Acceptance criteria agreed before any build
A fixed quote for the build
We need from you
Approve the scope and the acceptance criteria
Open the accounts in your name
Confirm who approves work before it goes out
03
Build
Does it work the way we agreed?
Agents, automations, clone or content systems and sequences are built inside accounts you own. You see a working demo every week, not a black box at the end. Build finishes when the acceptance criteria are met and you sign off.
You receive
A working system that meets the criteria from Map
Weekly demos and a record of decisions
Documentation written as we go
We need from you
Review demos and give decisions on a schedule
Provide source material (footage, voice, examples, approvals)
Verify your own voice or likeness where a platform requires it
04
Install
Can your team run it without us?
We hand over the keys: playbooks, prompts, SOPs, training, credentials in your vault, and a clear statement of what you control versus what depends on a vendor. Install is complete when your team runs one full cycle unassisted. Ongoing support is optional and scoped separately.
You receive
Playbooks, prompts and SOPs your team uses
Training for the people who own it
A portability plan: one primary platform, one tested backup
We need from you
Name the owner and give them the hours
Run one full cycle with us watching
Decide whether you want ongoing support
We map the work before we pick the software.
Sometimes the answer is a simpler process or a standard automation, not an AI agent or a new CRM.
Fictional business
A home-services company gets quote requests by web form, phone and social messages.
Today: 5 steps, 5 bottlenecks, no owner
!
1Requests arrive
Web form, voicemail and social messages land in three different inboxes.
Bottleneck: no single list
!
2Details collected
The office manager copies details into a spreadsheet, often without address, photos or timing.
Bottleneck: retyping, missing details
!
3Request routed
Whoever sees it first decides what happens next.
Bottleneck: no owner
!
4Reply drafted
Replies are written from scratch when someone has time.
Bottleneck: replies wait for free time
!
5Price approved
The owner prices from the spreadsheet when free.
Bottleneck: waits for the owner
Assessment summary
Draft for review
ObjectiveIn view
Every request gets an owner, complete details and a first reply.
What changes
EliminatedSpreadsheet copy
AutomationIntake, routing, follow-up task
AI-assistedFirst-reply draft
Human decisionPricing, exceptions
Tool choice and dependencies
Keep the current CRM if it can route and assign. GoHighLevel only if it can't.
Acceptance and measurement
Every test request logged, owned and answered. Measured against a baseline from your own inbox and CRM timestamps, not estimates.
Ownership and boundaries+
Systems, knowledge and data ownership
Form, phone, social page, spreadsheet. The business owns every account.
Responsibilities and boundaries
Coordinator owns each record; owner approves prices. No access to accounting.
Operational foundations
Playbooks, owners and a scorecard come before the agent.
Many AI projects stall because the work they would automate was never described. Where it makes sense, the Map phase writes down the playbooks, names an owner for each, decides which steps a person approves and which a system carries out, and picks a few measures. The diagram shows the model we use. It is QS's own plain-language model, not a reproduction of a third-party framework.
Two engines, one hand-off, held up by playbooks, owners and a scorecard
Growth engine
Wins customers and agrees the work.
Define the offer
Reach the right people
Answer and qualify every inquiry
Agree the scope
Hand-offAgreed scope passes to delivery
Fulfillment engine
Delivers the work and learns from it.
Onboard
Deliver and check
Support and improve
Holding both up
Playbooks
Repeatable work written down so anyone trained can run it.
Named owners
One accountable person per playbook, and a backup.
A scorecard
A few measures from your own records that everyone can read.
A review rhythm
A regular time to look at the measures and fix the weakest step.
In words: A growth engine attracts and qualifies customers and agrees the work. A fulfillment engine delivers it. The two meet at a hand-off. Both run from written playbooks, each with a named owner, and are read through a short scorecard reviewed on a regular rhythm. People (roles, hiring, onboarding) staff both engines.
For the people side (role scorecards, work-sample hiring and onboarding plans), see People in the frameworks below. We help design roles and hiring steps; we are not a recruiter and do not source candidates. A longer explanation of documenting a process is in what to document before building an AI agent.
The frameworks behind the work
Three established approaches, used where they fit.
They solve different problems, so we keep them separate rather than blending them into one branded method. Diagnose, Map, Build, Install is our engagement process, shown above; these are the methods applied inside it.
Marketing
DigitalMarketer: Build → Optimize → Accelerate.
Build. Who you serve, your offer and value proposition, the Customer Value Journey, and a baseline scorecard.
Optimize. Conversion and the early customer experience. Fix the main bottleneck before buying more traffic.
Accelerate. Expand what is working, with delivery capacity and owners in view, and set the next priorities.
Operations
An operating system approach, influenced by Scalable's methods.
Two engines. A growth engine that wins customers and a fulfillment engine that delivers, with a clear hand-off.
Playbooks and owners. Repeatable work written down, each playbook with one accountable owner.
Scorecard and rhythm. A few measures everyone can read, reviewed on a regular planning and review cycle.
People
Role clarity and practical hiring, drawn from Agency Rocket Fuel's Hiring Engine.
Define the role. Outcomes, responsibilities and measures before the job is posted.
Hire with a work sample. A test project built from a real task, assessed the same way for everyone. A person decides.
Equip. Access, playbooks, training and a definition of success for the new hire.
From playbook to AI workflow
An agent is only as dependable as the process it runs.
We decide what to automate first using the numbers: cost, capacity, margin and the cost to serve a customer. Then each piece of the business design becomes part of the build.
How business design becomes a dependable AI workflow
Business design
What it becomes in the build
What you get
Clear outcome and scope
A bounded task the system is allowed to do
Less activity that misses the goal
Playbook and examples
Explicit instructions, inputs, outputs and checks
Repeatable work you can review
A named owner
Someone responsible for exceptions and changes
Problems don't disappear between tools
Tool and data access
The smallest set of integrations and permissions that does the job
It fits how you already operate
Quality checks and test cases
Known success, failure and escalation paths
Confidence based on evidence
Scorecard and review
Logs, operating measures and a maintenance rhythm
You can see whether it's helping
An agent skill can carry out part of a playbook, but having one doesn't make the process reliable. No agent is error-free or free of questions. The aim is predictable work with clear limits, and a person's attention where it matters.
Scope determines the next step. We quote the number before you commit, and we only publish starting prices once they are approved.
What you receive and control
Own the work. Rent the engines.
You control
Accounts opened in your name, and the keys to them
Your approved source content: footage, voice recordings, scripts, knowledge files
Prompts, workflows, SOPs and documentation we build for you
Your finished content and agreed exports, once paid in full
Depends on a vendor
The AI models and hosted platforms that run the system
Your avatar and voice clone, which live inside the platform that created them
Subscriptions and usage costs, which we document and review with you
We make systems portable where it is practical: one primary platform, one tested backup, documented interfaces and reusable instructions. Record once, rebuild anywhere. We do not promise exported voice or avatar model files, or a system that runs with no maintenance.
Questions
Straight answers.
By telling us about the project through the contact page. It takes a few minutes. Pat reads every message personally and replies with a next step or a question. Nobody is asked to buy anything to have that conversation.
No. The four phases are the spine, not a mandatory sequence. A single hour of consulting, a bounded pilot, or a direct proposal for a clearly scoped project are all valid paths. Diagnose exists because most prospects do not yet know where to start.
Hourly consulting is $500 per hour. Beyond that, scope determines the next step, and we tell you the number before you commit to it. We publish other starting prices only once they are approved, so you will not find guesses here.
You own the accounts we open in your name, your raw footage and voice recordings, scripts, knowledge files, prompts, workflows, SOPs, and your finished content once paid for. You rent the AI models and hosted platforms that run them. Your avatar and voice clone live inside the platforms that create them and usually cannot be exported as files; if you switch platforms we rebuild from the recordings you own.
Diagnose needs the assessment plus one working session. Build needs a weekly demo review and timely decisions. Install needs the owner on your side to run one full cycle with us. We will tell you the hours up front for your scope.
We agree data rules for your business before any build: which provider and plan handles which class of data, what is never entered into an AI tool (payment data, for example), and how retention works. Those rules go into your plan and your workspace documentation.
Tell us about the project. We will tell you the right next step.
A few minutes to describe it. Pat reads every message and replies personally with a next step or a question.
