Definition & Framework

Process Intelligence and AI Readiness

What both terms mean, the frameworks that make an assessment repeatable, and why the assessment is the most sellable AI engagement an MSP has.

The definition

Process Intelligence is the practice of documenting how a business actually works before automating any part of it. For an MSP it is a paid assessment engagement: interview the executives, inventory and score the processes, map the workflows into written procedures, and produce a roadmap the client will fund.

AI Readiness is the scored output. Five domains, five-point scale each, producing domain scores and one aggregate AI Readiness Score on a maturity continuum.

Written by Tony Ferrigno, Founder and Managing Partner, AiT Advisory Group. Author of AI for MSPs: Process Intelligence & AI Readiness.

Why the assessment is the easiest AI engagement to sell

Most MSPs trying to enter AI reach for the hardest thing first. They try to sell an automation build to a client who has never bought anything like it, at a price the client has no way to evaluate.

The assessment inverts that. It is small, fixed-fee, short, and low-risk. It answers a question the client is already asking themselves, which is some version of "everyone says we should be doing AI, are we even ready." It produces a document the client owns. And it puts you in a room with the CEO and the COO for several hours, which is worth more than the fee.

It also does the qualification work you would otherwise do for free. By the end you know whether this client has processes worth automating, data worth using, and a sponsor willing to fund anything. Clients who fail that test have paid you to find out.

This is not enterprise process mining

Worth stating plainly, because the term collides. In the enterprise software market, "process intelligence" means log-based process mining: software that reconstructs workflows from system event data, priced for large organizations with a team to run it. That is a different thing solving a different problem.

DimensionEnterprise process miningProcess Intelligence for MSPs
MethodReconstructs processes from system event logsStructured executive and functional-leader interviews
AssumesProcesses already run inside instrumented systemsThe processes that matter live in people's heads and spreadsheets
Client sizeLarge enterprise with a dedicated team25 to 300 users, no analyst on staff
Software purchaseRequired, and substantialNone
DurationMonthsTwo to five weeks
OutputA dashboard someone has to maintainWritten documentation and a funded roadmap the client owns
Also not this

Most "AI readiness assessments" in the channel today are vendor lead magnets. They are free, they take an hour, and they conclude that the client should buy Copilot licenses. That is a licensing checklist wearing an assessment's clothes. A real assessment can conclude that the client is not ready, and sometimes should.

The Process Intelligence Framework

Five components, run in order. Components one and two qualify. Three and four create the lasting value. Five is what gets funded.

01

Executive Discovery and Process Assessment

60 to 90 minutes each, with the people who own outcomes.

Structured interviews with the CEO, the COO, and functional leaders. Not a survey. Not an email questionnaire.

  • The CEO interview establishes what the business is trying to become
  • Functional leaders describe what actually happens, which rarely matches the org chart
  • Contradictions between the two are findings, not noise
02

Process Inventory and Prioritization

Scored on four dimensions.

Every process worth naming gets scored on frequency, effort, error rate, and AI suitability.

  • Scoring forces comparison, which stops the loudest voice in the room from setting priorities
  • Error rate is the dimension clients underestimate most
  • The inventory is often the first time anyone has written the list down
03

Workflow Mapping and SOP Development

Where the deepest lasting value is created.

Turn the priority processes into documented workflows and written standard operating procedures.

  • Clients keep referring to this years later, whether or not they ever automate
  • Documentation is also the input an agent needs, so the work is never wasted
  • It converts knowledge held by individuals into an asset the business owns
04

Centralized Process Hub Design

SharePoint, Notion, or Confluence.

Information architecture plus a governance model, so the documentation stays current instead of decaying.

  • Structure matters less than naming an owner and a review cadence
  • Build it in a tool the client already pays for
  • A hub with no owner is a folder, and folders go stale within a quarter
05

Automation and AI Readiness Roadmap

The deliverable that makes the next investment defensible.

Synthesize everything into a sequenced plan with scores, priorities, and dependencies.

  • Sequenced by readiness, not by enthusiasm
  • Names what must be fixed before any AI work starts
  • This is the document the client takes to their board or their bank

The Five-Domain AI Readiness Framework

Each domain scored on a five-point scale. The domain scores matter more than the aggregate, because they tell the client exactly where the work is.

DomainWhat it measuresCommon finding
Technology InfrastructureWhether the systems can support AI workloads and integrate with each otherIntegration gaps nobody knew existed until someone asked
Data Quality and AvailabilityWhether the data an agent would need is accessible, current, and trustworthyThe data exists but lives in three places that disagree
Business Process MaturityWhether processes are consistent and documented enough to automateProcess debt, usually more than anyone expected
Workforce ReadinessWhether people will adopt, resist, or quietly route around new toolsEnthusiasm at the top, unspoken concern below it
Governance and ComplianceWhether policy, approval, and audit structures exist to control AI useNo AI policy at all, and often no plan to write one

The fifth domain is the one that turns an assessment into recurring revenue. Almost no SMB has an AI governance policy. Finding that gap is not a problem you point at, it is a service you propose.

The Process Priority Matrix

Business impact on one axis, automation potential on the other. Four quadrants, four different recommendations.

High impact · High automation

Primary AI targets

Immediate focus. These become the agent candidates that carry the business case.

High impact · Low automation

Fix the process first

Process improvement and data quality work before any AI. Saying so builds more trust than promising automation you cannot deliver.

Low impact · High automation

Quick wins

Automate for efficiency. Valuable early, because momentum is worth more than magnitude in the first ninety days.

Low impact · Low automation

Deprioritize

Operational maintenance only. Revisit annually. Saying “leave this alone” is a finding too.

Process debt

Process debt is the accumulated cost of informal workarounds, undocumented exceptions, and legacy procedures that nobody owns. It does not show up on a balance sheet. It shows up when someone leaves.

Naming it is what creates urgency, and quantifying it is what validates the fee. The framing that lands in a boardroom is specific and countable:

How it sounds in the room

"You have 23 undocumented processes that represent significant knowledge-loss risk if your key people leave."

That sentence does more work than any slide about AI. It reframes the engagement from a technology project the client can defer into an operational risk they already have. Note that it also stands entirely on its own merits. Even a client who never buys a single agent is better off having the list.

Pricing the engagement

Three packages. Price against the risk the assessment removes, not against the hours it takes.

AI Readiness Assessment
Up to 50 users, roughly two weeks
$4,500 to $12,000
Process Intelligence Engagement
Up to five priority processes, roughly three weeks
$8,000 to $22,000
Full Program
Up to eight priority processes, four to five weeks
$18,000 to $45,000

Client size moves those bands considerably. A small client of 1 to 25 users sits at the bottom of each range. An enterprise client above 300 users can reach $55,000 to $100,000 or more for a full program.

Bill 50 percent at engagement start and 50 percent at report delivery. Put the annual governance retainer, $800 to $3,000 per month, in the original proposal. Never hold it back as a post-engagement upsell, because a client who has already received the report has less reason to buy the thing that keeps it current.

The value argument that works

The assessment prevents failed AI projects costing $50,000 to $500,000. It identifies automation opportunities worth $5,000 to $50,000 per month. It creates governance structures that reduce regulatory exposure. Against that, a $10,000 to $25,000 engagement is not a cost decision. It is insurance with a roadmap attached.

Why assessment engagements fail

Five failure modes. Four of them are visible in the first week if you know what to watch for.

Executive Sponsor Disengagement

The sponsor books the kickoff then delegates every interview. The report lands with nobody senior invested in it. Catch this by making the CEO interview a condition of starting, not a scheduling preference.

The Predetermined Conclusion Problem

The client has already decided what the assessment should say, usually that they need a tool someone already sold them on. Your findings become an obstacle. Surface it early by asking what they expect the report to conclude.

The Cold Room

Staff attend interviews and say nothing useful, because they think the exercise is about cutting headcount. Nobody describes their workarounds to someone they think is auditing them. Address it before the first interview, not after.

The Stalled Delivery

The report is delivered, everyone agrees it is good, and nothing is funded. Almost always a sequencing failure: the roadmap named ten things instead of the one thing to do next quarter.

The Embarrassment Discovery

The assessment surfaces something the sponsor did not want documented. Handle it privately and ahead of the written report. How you manage this determines whether you ever work with that client again.

What comes after the report

The roadmap is the bridge. High-impact, high-automation processes are agent candidates, and agent candidates are where Managed Intelligence begins. The governance gaps from the fifth domain justify the retainer. The documented SOPs from component three are the input those agents need.

That sequence is the point. An assessment sold on its own is a good engagement. An assessment sold as the front door to a recurring practice is a business.

Frequently Asked Questions

Common questions about this engagement

Sell the assessment, then sell what it finds

AiT Advisory Group runs the Process Intelligence and Automation Readiness engagement with you, so your team learns the method on a live client instead of a template.