The AI Readiness Assessment

The Story of Two Companies

In the spring of 2025, two mid-sized professional services firms decided to adopt AI. Both had roughly 400 employees. Both served clients in the same industry. Both had competent leadership teams and healthy balance sheets. Both chose Claude as their AI platform. One succeeded spectacularly. The other spent six months and $200,000 and had almost nothing to show for it.

The company that succeeded was Hartwell & Associates, a management consulting firm based in Chicago. Their COO, Jennifer Okafor, began with a question that seemed almost too simple: "Are we actually ready for this?"

Before purchasing a single license, Jennifer spent three weeks conducting what she called an "AI readiness audit." She interviewed department heads. She surveyed employees about their current technology comfort levels. She mapped the firm's most time-consuming workflows. She reviewed data governance policies. She assessed their IT infrastructure. She identified potential champions and potential resistors.

What Jennifer discovered was sobering. While the firm had modern cloud infrastructure and a relatively tech-savvy workforce, their data governance was a mess. Client information lived in email threads, shared drives with inconsistent permissions, and personal folders on individual laptops. There was no formal policy about what data could be shared with external tools. The firm's legal team had never reviewed the terms of service for any AI product. And perhaps most importantly, roughly 35% of their employees -- mostly senior partners -- had never used any AI tool, not even casually.

Jennifer did not view these findings as reasons to delay. She viewed them as a roadmap. Before deploying Claude, Hartwell spent six weeks addressing the most critical gaps: creating a data classification policy, establishing guidelines for what information could and could not be shared with AI tools, running a voluntary AI literacy workshop for senior partners, and identifying three specific workflows where Claude would be piloted first.

When they finally deployed Claude Team to their first pilot group of 20 consultants, adoption was smooth. Users understood the boundaries. IT had answers for security questions. Leadership could articulate why they were doing this and what success looked like. Within 90 days, the pilot group reported saving an average of 7 hours per person per week on research, analysis, and first-draft writing. Hartwell expanded to firm-wide deployment three months later.

The company that struggled was Brennan Group, a competing firm across town. Their CEO, Marcus Brennan, attended a conference where an electrifying keynote about AI convinced him that his firm was falling behind. He returned to the office on Monday and directed his CTO to "get everyone on Claude by the end of the month."

The CTO, working under pressure, purchased 400 Claude Team licenses and sent a company-wide email announcing the new tool. There was no readiness assessment. No data governance review. No training program. No identified use cases. No pilot group. Just 400 licenses and an email that said, essentially, "Here is Claude. Use it to be more productive."

What happened next was predictable. About 60 employees -- the early adopters -- started using Claude enthusiastically but inconsistently, with no shared practices or guidelines. Some uploaded confidential client data without understanding the privacy implications. Others used it for tasks where it performed poorly and concluded the technology was overhyped. Senior partners ignored it entirely. The IT help desk was overwhelmed with basic questions they were not prepared to answer. Three months in, active usage had dropped to 40 users. The CTO was spending more time managing complaints than capturing value.

By month six, Marcus Brennan quietly reduced the license count to 50 and shifted his narrative from "AI transformation" to "strategic evaluation phase." The firm had spent $200,000 on licenses, IT support time, and lost productivity, with no measurable return.

The difference between Hartwell and Brennan was not technology. They used the same AI tool. The difference was not budget. Brennan actually spent more. The difference was readiness. Hartwell knew where they stood before they started. Brennan assumed they were ready because they wanted to be.

---

The AI Readiness Framework

Organizational AI readiness is not a single thing. It is a composite of six distinct dimensions, each of which can independently enable or block successful deployment. An organization might be highly ready on technology infrastructure but completely unready on data governance. Or they might have excellent leadership support but poor workforce skills. The assessment must cover all six dimensions to be useful.

Dimension 1: Leadership Alignment

AI adoption without leadership alignment is like building a house without a foundation. It might look impressive for a while, but it will not survive the first storm.

Leadership alignment means more than the CEO saying "we should use AI." It means the leadership team has agreed on:

• Why the organization is adopting AI (the specific business problems being solved)
• How success will be measured (quantified metrics, not vague aspirations)
• Who is responsible for driving adoption (a named individual or team, not "everyone")
• What resources will be allocated (budget, time, executive attention)
• Where AI will be deployed first (specific departments and use cases, not "everywhere")

The most common leadership alignment failure is what organizational psychologists call "commitment without clarity." The CEO is enthusiastic. The CFO has approved a budget. But when you ask five executives what the AI initiative is supposed to achieve, you get five different answers. This ambiguity cascades through the organization, creating confusion at every level.

Assessment questions for Leadership Alignment:

Dimension 2: Workforce Readiness

Your employees are the people who will actually use the AI tool every day. Their readiness -- their skills, their attitudes, their confidence, their concerns -- determines whether Claude becomes a daily productivity multiplier or an expensive icon that sits unused on their desktops.

Workforce readiness encompasses several factors:

Technology comfort. How comfortable are your employees with learning new technology tools? This varies dramatically across roles, departments, generations, and individual personalities. A workforce that recently navigated a major technology change (CRM migration, new project management tool, cloud transition) may be fatigued by yet another change. A workforce that has been on stable systems for years may find the novelty exciting but the learning curve steeper.

AI literacy. What do your employees actually understand about AI? Many knowledge workers have used ChatGPT casually but have never thought carefully about hallucinations, data privacy, or the difference between pattern prediction and factual retrieval. Others have heard enough alarming headlines about AI replacing jobs that their dominant emotion is anxiety, not curiosity. Understanding your starting point on AI literacy is essential for designing effective training.

Attitudes toward AI. Employee attitudes toward AI typically fall into four categories:

1. Enthusiasts (typically 15-20%): Already using AI tools, eager for more, may need guidance on responsible use more than encouragement to adopt.
2. Pragmatists (typically 40-50%): Open to AI if shown clear personal benefit. Will adopt willingly if well-supported. This is your swing group and the primary target for your onboarding program.
3. Skeptics (typically 20-25%): Doubtful that AI adds real value, may have tried it and been unimpressed, or may be concerned about job security. Require evidence-based persuasion -- success stories from peers, not executive mandates.
4. Resistors (typically 5-15%): Opposed to AI for philosophical, practical, or emotional reasons. Some have legitimate concerns about quality, reliability, or ethics. Others are simply resistant to any change. Mandating adoption for this group without addressing their concerns creates toxicity.

Assessment questions for Workforce Readiness:

Dimension 3: Data Governance

Data governance is the dimension that most frequently derails AI deployment, because it is the dimension that organizations most frequently assume is adequate when it is not.

Deploying Claude means that employees will inevitably type organizational information into an AI tool. The question is not whether this will happen. The question is whether your organization has clear rules about what information can be shared, under what circumstances, and with what protections.

Data classification. Does your organization have a data classification scheme? Most enterprise data falls into categories like:

• Public: Information already available externally (published reports, marketing materials, public filings)
• Internal: Information meant for organizational use but not highly sensitive (meeting notes, project plans, internal communications)
• Confidential: Sensitive business information (financial data, strategic plans, employee records, proprietary processes)
• Restricted: Highest sensitivity (personally identifiable information, protected health information, trade secrets, client data subject to NDA)

Without clear classification, employees cannot make good decisions about what to share with AI tools. And even well-intentioned employees will make mistakes. The compliance analyst who pastes a regulatory document into Claude is probably fine -- that is public information. The HR manager who pastes a performance review into Claude might be violating employee privacy policies. The distinction requires clear guidelines, not individual judgment calls in the moment.

Assessment questions for Data Governance:

Dimension 4: Technology Infrastructure

Technology infrastructure for AI deployment is simpler than most organizations expect. Claude is a cloud-based service -- you do not need GPU clusters, specialized hardware, or complex installations. But there are still infrastructure considerations that matter.

Identity and access management. Can your IT team manage Claude accounts through your existing identity provider? Claude Enterprise supports SSO (SAML 2.0) and SCIM provisioning, which means you can manage Claude access through the same identity system you use for other enterprise tools. If you are starting with Claude Team (which does not include SSO), you need a manual process for provisioning and de-provisioning accounts.

Network and security. Is Claude accessible from your corporate network? Some organizations have web filtering or firewall rules that block access to AI tools. Others have security policies that require all external services to go through a review process. Identify and resolve these blockers before deployment, not after -- nothing kills adoption momentum faster than employees being unable to access the tool they were just told to use.

Integration points. Where does Claude need to connect with your existing tools? While initial deployment typically starts with the Claude.ai web interface, you should identify future integration points -- CRM, document management, project management, communication tools -- so your technology choices are compatible with your long-term roadmap.

Dimension 5: Use Case Clarity

The organizations that succeed with AI deployment can articulate, in specific terms, exactly which tasks they expect AI to improve. The organizations that struggle have a vague aspiration to "be more productive" without a concrete starting point.

Use case clarity means you have identified:

• Specific tasks that AI will be applied to (not "improve marketing" but "draft first versions of weekly client email newsletters")
• Current performance on those tasks (time required, quality level, volume)
• Expected improvement from AI (quantified time savings, quality benchmarks)
• Success criteria for the pilot (how will you know this worked?)

The best starting use cases share several characteristics:

1. High volume: Tasks that happen frequently enough to generate meaningful data on AI impact
2. Clear input/output: Tasks with well-defined inputs (a document, a dataset, a brief) and well-defined outputs (a summary, a draft, an analysis)
3. Low risk: Tasks where an AI error would be caught before causing harm (internal documents rather than client-facing; first drafts rather than final versions)
4. Measurable: Tasks where you can objectively measure time, quality, or both

Dimension 6: Change Management Capacity

Even with perfect leadership alignment, a skilled workforce, strong data governance, solid infrastructure, and clear use cases, AI deployment will fail if the organization lacks the capacity to manage change effectively.

Change management capacity includes:

• Communication channels: Can you reach all affected employees with clear, consistent messaging?
• Training resources: Do you have the people and time to train users effectively?
• Feedback loops: Can employees report problems, ask questions, and share successes?
• Iteration capability: Can you adjust the deployment plan based on what you learn during the pilot?
• Champions network: Do you have influential individuals in each department who will advocate for adoption among their peers?

---

Apply: Conduct Your Assessment

---

Reflect: Your Readiness Roadmap

The Maturity Spectrum

Based on your assessment, your organization falls somewhere on the AI maturity spectrum. Understanding where you are helps you set realistic expectations and plan effectively.

Stage 1: Unaware (Overall score 1.0-1.5) The organization has not seriously considered AI adoption. There are no policies, no training, no identified use cases, and limited leadership interest. This stage requires foundational education and awareness-building before any deployment planning begins.

Stage 2: Exploring (Overall score 1.5-2.5) The organization recognizes AI as relevant and some individuals are experimenting, but there is no organizational strategy, governance, or structure. This is where most organizations were in 2024. The priority at this stage is building leadership alignment and data governance foundations.

Stage 3: Piloting (Overall score 2.5-3.5) The organization has leadership support, some governance structure, and identified use cases, but has not yet deployed at scale. This is the ideal stage for launching a structured Claude pilot. The priority is executing a well-designed pilot and learning from it.

Stage 4: Scaling (Overall score 3.5-4.5) The organization has successfully piloted AI with positive results and is expanding deployment across departments. The priority is scaling infrastructure, training, and governance to support broader adoption.

Stage 5: Optimizing (Overall score 4.5-5.0) AI is embedded in core workflows across the organization. The priority is continuous improvement -- measuring ROI, optimizing workflows, building advanced use cases, and developing AI-native processes that were not possible before.

Common Readiness Gaps and How to Close Them

The 30-Day Readiness Sprint

If your assessment reveals gaps but your organization is motivated to move quickly, here is a 30-day sprint to build readiness:

Week 1: Align leadership. Present assessment results to executive team. Agree on business problems, success metrics, sponsor, and budget. Get executive sponsor to spend one hour using Claude personally.

Week 2: Establish governance foundations. Create data classification policy (three tiers). Draft AI usage guidelines. Have legal review Claude terms of service. Establish rules for what data categories can be shared with Claude.

Week 3: Build workforce readiness. Run AI literacy workshop for all affected employees. Identify 2-3 champions per department. Create a simple quick-start guide for Claude. Set up a feedback channel (Slack channel, Teams channel, or regular check-in meetings).

Week 4: Define and launch pilot. Finalize pilot use cases (2-3 specific tasks). Select pilot group (10-20 users across 2-3 departments). Configure Claude Team workspace. Launch pilot with clear success criteria and 90-day evaluation timeline.

This is not the only way to build readiness, but it demonstrates that readiness gaps do not require months of preparation. A focused four-week sprint can close the most critical gaps and position your organization for a successful pilot.

What Jennifer Okafor Would Tell You

At a panel discussion six months after Hartwell's successful deployment, Jennifer Okafor was asked what advice she would give to other COOs starting their AI journey. Her answer was characteristically direct:

"Do the boring work first. Everyone wants to skip to the exciting part -- the demos and the wow moments. But the organizations that succeed are the ones that answer the boring questions first. Who owns this? What data can we share? How will we train people? What does success look like? If you cannot answer those four questions in one sentence each, you are not ready. And no amount of technology will make up for that."

She paused, then added: "And make your CEO use the tool. If the person championing AI transformation has never actually used AI, you have a marketing campaign, not a strategy."