HOW TO // AI · AI Readiness

Sample report for a fictional mid-size manufacturer. Yours is written from your answers.

Get your report

AI Readiness Report

Example Manufacturing Co.

Prepared by David Davis, HOW TO // AI · Sample

39
out of 100
Not Ready

Your infrastructure is ahead of your guardrails: fix shadow AI and oversharing first, then your service desk goal is very achievable.

Executive summary

Example Manufacturing is rated Not Ready at 39/100. That isn't a verdict on your IT team. It reflects that AI is already in use across the company without the policy, visibility, or data controls to make it safe. Your identity controls and hybrid Microsoft footprint are real strengths to build on.

Three issues matter most. Employees are using public AI tools with company data and nothing stops customer or pricing data from leaving. Your SharePoint and file shares are overshared, which becomes an instant data leak the day an AI assistant is switched on. And nobody owns AI decisions, so there's no inventory, no approved-tool list, and no one to say yes safely.

The good news: most of the fixes are policy and configuration work inside the Microsoft 365 tenant you already own. Close the risk gaps in the first 30 days, and a service desk pilot, your stated goal, is realistic inside 90 days with results you can show leadership.

Scorecard

  • Data Readiness38
  • Security & Privacy50
  • Infrastructure & Platforms50
  • Skills & Talent31
  • Use Cases & Value38
  • Governance & Operating Model25

0-39 Not Ready · 40-59 Emerging · 60-79 Ready · 80-100 Leading. Scores are calculated from your answers using the HOW TO // AI readiness framework.

Dimension by dimension

Data Readiness

38

Core systems exist, but data ownership is informal and unstructured content is overshared.

Strengths

  • Core business apps are in place with some integrations, so structured data is findable.

Gaps

  • Nobody owns data quality, so AI answers will inherit today's errors.
  • SharePoint and file share permissions are known to be too broad and haven't been cleaned up.

Security & Privacy

50

Identity is solid; data protection and AI-specific controls lag behind actual usage.

Strengths

  • SSO, role-based access, and periodic access reviews give you a strong base for any AI rollout.
  • Some sensitive data already carries labels.

Gaps

  • Employees use public AI tools with work data and there's no approved alternative.
  • Vendor reviews don't ask whether AI features retain or train on your data.

Infrastructure & Platforms

50

A hybrid Microsoft and Azure footprint that can support AI once governance catches up.

Strengths

  • Hybrid SaaS and Azure workloads mean AI services are within reach without new platforms.

Gaps

  • Enterprise AI is still in evaluation, so people fill the gap with consumer tools.
  • Cost dashboards exist, but no budgets or alerts are ready for usage-based AI spend.

Skills & Talent

31

Curiosity is there, but no training, no AI credentials, and no budget to close the gap.

Strengths

  • Most of the team has tried AI tools, so adoption won't start from zero.

Gaps

  • No one holds an AI-related certification.
  • No one owns building AI automations, which your service desk goal requires.
  • Upskilling is approved case by case with no budget.

Use Cases & Value

38

A clear goal exists, but no prioritized list, no pilot, and no way to measure value.

Strengths

  • You have a concrete, measurable goal: reducing service desk and documentation time.

Gaps

  • Ideas aren't prioritized or owned by the business.
  • Value is judged by anecdote, so wins won't survive a budget review.

Governance & Operating Model

25

The weakest area. No policy, no decision owner, and no inventory of AI in use.

Strengths

  • An AI acceptable use policy is already in draft.

Gaps

  • IT decides on AI by default, without security, legal, or business input.
  • SOX and state privacy obligations haven't been assessed against AI use.
  • There's no inventory of AI tools or AI features inside your SaaS apps.

Top risks

  1. 1. Overshared files become AI answers

    Any AI assistant connected to Microsoft 365 can surface files a user can technically open, such as pricing sheets, HR records, or board decks, even if they'd never find them on their own.

    Fix: Run SharePoint and OneDrive sharing reports, remove 'Everyone' and broad links on your top 20 sites, and assign an owner to each site before enabling any Microsoft 365 AI assistant.

  2. 2. Customer and pricing data in public AI

    Your stated concern is already happening. Data pasted into consumer AI tools can be retained under terms you haven't reviewed, which is a SOX and customer-contract problem.

    Fix: Give everyone an approved enterprise AI assistant with data protection, then block or warn on unapproved AI sites through your web filter. An approved tool has to come first, or usage just moves to personal phones.

  3. 3. No inventory of AI in use

    AI features are switching on inside SaaS tools you already pay for. Without an inventory you can't answer an auditor, a customer questionnaire, or a breach investigation.

    Fix: Pull a list of AI tools from web proxy and SaaS logs, add AI features in your top 15 SaaS apps, and keep it in a simple register with an owner and data type for each.

  4. 4. No one owns data quality

    AI trained on or answering from your ERP and CRM will repeat today's data errors with more confidence, which erodes trust in the first pilot.

    Fix: Name a data owner for ERP, CRM, and the service desk knowledge base, each responsible for a short list of quality rules.

  5. 5. No AI credentials on the team

    Without a shared baseline, the team will struggle to evaluate vendors, secure new AI features, or build the service desk automation you want.

    Fix: Put every IT staff member through a fundamentals certification this quarter and send one engineer down an AI developer path on Azure.

Quick wins

Start a team AI certification sprint

Owner: IT Director · Start this week; 6 weeks to first passes

  1. Pick one fundamentals cert for the whole IT team (Azure AI Fundamentals fits your stack).
  2. Give everyone 2 hours a week of protected study time for 6 weeks.
  3. Track practice-test readiness weekly and book exams as people hit the bar.

Set a small AI training budget

Owner: IT Director · 1 week

  1. Set aside a per-person annual amount for AI training and exams.
  2. Protect a recurring learning block on calendars.
  3. Ask each person for one AI use they'll try in their own job this month.

Baseline your service desk before AI

Owner: Service Desk Manager · 2 weeks

  1. Pull 90 days of ticket volume, first-response time, and time-to-resolve.
  2. Tag the top 10 ticket types by volume.
  3. Record time spent writing internal documentation.
  4. Agree on the two metrics leadership will judge the pilot by.

Publish the AI acceptable use policy

Owner: IT Director with HR and Legal · 3 weeks

  1. Finish the draft with three data tiers: never, only in approved tools, and fine anywhere.
  2. Name the approved tools explicitly.
  3. Have legal and HR review, then publish with a 10-minute training.
  4. Require acknowledgment through your existing policy tool.

Stand up a small AI council

Owner: Executive Sponsor · 2 weeks to first meeting

  1. Invite IT, security, legal, finance, and one business leader.
  2. Meet every two weeks for 30 minutes with a one-page intake form for new AI tools and ideas.
  3. Make the first decisions: the approved assistant and the pilot use case.

Your 90-day plan

Days 1-30

Close the risk gaps

  • Publish the AI acceptable use policy and run the 10-minute training.

    IT Director → Every employee knows what data can go where.

  • Roll out an approved enterprise AI assistant and restrict unapproved AI sites.

    Security Lead → Shadow AI drops and usage becomes visible.

  • Clean up sharing on the 20 highest-risk SharePoint sites.

    Microsoft 365 Admin → Oversharing risk contained before any AI assistant is connected to company content.

  • Build the first AI inventory and stand up the AI council.

    IT Director → One list and one decision-maker for AI.

Days 31-60

Build the foundations

  • Extend sensitivity labels and DLP to cover customer and pricing data.

    Security Lead → Sensitive data is labeled and protected in email, endpoints, and AI tools.

  • Name data owners for ERP, CRM, and the knowledge base.

    Executive Sponsor → Clear accountability for the data AI will use.

  • Map SOX and state privacy requirements to AI use.

    Compliance Lead → Known obligations before the pilot touches production data.

  • Complete the team certification sprint.

    IT Director → A shared AI baseline across the IT team.

Days 61-90

Prove value with a pilot

  • Launch a service desk pilot: AI-drafted ticket responses and knowledge articles for the top 10 ticket types.

    Service Desk Manager → A working pilot on real tickets with humans approving every response.

  • Compare pilot metrics to the baseline every two weeks.

    Service Desk Manager → Measured time savings leadership can trust.

  • Set Azure budgets and alerts for AI usage.

    Cloud Engineer → No surprise invoices as usage grows.

  • Present results and a scale-or-stop decision to the AI council.

    IT Director → A funded next step based on evidence.

Skills gap and certification plan

Your team needs a shared AI baseline now and one builder by the time the pilot starts. Because you're a Microsoft 365 and Azure shop, stay on the Microsoft path and add one vendor-neutral cert for the people who own risk.

Get your whole team exam-ready

HOW TO // AI CERT for Teams gives every person practice questions, timed mock exams, an AI tutor, and a readiness score for all of these certifications, with volume pricing.

See team pricing →

A final word

You're closer than the score suggests. Most of the gap is decisions and configuration, not budget or new platforms. Get the guardrails in place this month and your team will be the one leading AI adoption, not cleaning up after it.

David DavisCCIE with 25+ years in enterprise IT, author of 100+ courses on Pluralsight and LinkedIn Learning, and co-founder of ActualTech Media.