AI Readiness Assessment: Why Enterprise AI Pilots Stall

Quick answer: An AI readiness assessment is a structured review of your strategy, data, technology, skills, governance and culture that shows whether your organisation can adopt AI successfully and what must be fixed first. It ends with a maturity score for each area, a ranked list of use cases and a 6 to 12 month roadmap.

Many Indian enterprises have already tried an AI pilot: a chatbot here, a forecasting model there, a document extraction tool tested by one team. Few of these pilots reach production. The reason is rarely the algorithm. It is messy data, unclear ownership, missing skills, security concerns raised too late and no agreed way to measure value.

An AI readiness assessment finds these gaps before money is spent on tools, so the first serious AI project has a fair chance of success. This guide explains what it covers, how to score it, how to run it and what a useful report looks like.

Why AI Projects Stall Without a Readiness Check

AI depends on foundations that many organisations have not built yet. The common failure patterns are:

• Data that cannot be used. Records spread across Excel files, legacy systems and paper, with no single definition of a customer, product or case.
• Use cases chosen for novelty. Projects picked because they sound impressive, not because they solve a costly problem.
• No owner in the business. IT builds a model, but no department is responsible for using it or measuring results.
• Security and privacy raised late. Legal or compliance teams stop a project after months of work.
• Infrastructure gaps. No environment to deploy, monitor and update models once they are built.
• Unmanaged AI tools. Staff already paste company data into public AI tools or connect AI agents to business systems without review.

A readiness assessment surfaces these issues in weeks, when they are still cheap to fix. The rest of this guide shows how.

The Six Areas of an AI Readiness Assessment

1. Strategy and use cases

Start with business problems, not technology. Interview department heads to list pain points such as slow approvals, high manual effort, forecasting errors or customer wait times. Score each possible use case on business value, data availability, feasibility and risk.

2. Data

Review where key data lives, its quality, completeness and history, and who owns it. Check whether data can be reached through APIs or database connections, or only through manual exports. For most enterprises this is the area with the largest gap. A working data governance framework makes the rest of the work much easier.

3. Technology and infrastructure

Assess current cloud and on-premises capacity, integration options and tools for deploying and monitoring models. Decide whether workloads can run on public cloud, need a private environment or require a mix.

4. People and skills

Map existing skills in data engineering, analytics, software and domain knowledge. Identify who will build, who will use and who will govern AI systems, and plan training for business users as well as the technical team.

5. Governance, risk and compliance

Review policies on data privacy, information security, model approval and human oversight. Any use case that touches personal data must be designed in line with the Digital Personal Data Protection Act, 2023 and the DPDP Rules, 2025, with consent, purpose limitation and data minimisation in mind. Our DPDP Act compliance checklist covers the basics. AI agents that can act inside business systems also need access controls and logging; see our note on the CERT-In advisory on AI agents.

6. Culture and change readiness

Gauge how staff feel about automation and new tools. Resistance often comes from fear of job loss or extra work. Early communication and involving users in design reduce this risk.

How to Score AI Readiness

Score each of the six areas on a simple 1 to 5 scale, with evidence for every score. A shared scale keeps the discussion objective and makes progress easy to track at the next review.

Level	What it looks like	Typical next step
1. Initial	No AI strategy, data in silos, ad hoc use of public AI tools	Write an AI use policy and pick one problem to study
2. Exploring	A few pilots, no shared platform, unclear owners	Assign business owners and define success measures
3. Defined	Prioritised use cases, data owners named, basic governance	Build a repeatable data pipeline and deployment process
4. Managed	Models in production, monitored, with approval steps	Scale to more functions and track value centrally
5. Optimised	AI part of normal planning, budgets and risk reviews	Review portfolio value and retire weak use cases

Most organisations land at level 1 or 2 in data and governance even when strategy scores higher. That gap is exactly what the roadmap should close.

The Fix: How to Run the Assessment Step by Step

1. Set the scope. Decide whether to assess the whole enterprise, one business unit or a single function such as finance or customer service.
2. Form a small working group. Include IT, a data owner, a business sponsor and someone from risk or compliance.
3. Collect information. Use structured interviews, a short questionnaire and a review of systems, data sources and policies.
4. Score maturity. Rate each of the six areas on the 1 to 5 scale, with evidence for every score.
5. Prioritise use cases. Plot them on a value versus feasibility grid and pick two or three quick wins and one longer-term bet.
6. Build the roadmap. List the data, platform, skills and policy work needed before and during each project.
7. Present and agree. Review findings with leadership and confirm budget, owners and timelines.

For a mid-sized organisation this usually takes a few weeks, depending on how many departments and systems are in scope.

AI Readiness Checklist

Use these questions as a quick self-check before a full assessment:

• Do we have a written list of business problems ranked by cost or impact?
• Can we name the owner and location of the data each problem needs?
• Is there an approved policy on which AI tools staff may use and what data they may share?
• Do we have an environment where models can be deployed, monitored and updated?
• Has legal or compliance reviewed how personal data would be used?
• Is there a business sponsor willing to own results, not just attend meetings?
• Have we agreed how we will measure success and when we will stop a pilot?

If you answer "no" to three or more, start with a formal assessment before buying tools.

Self-Assessment vs Guided Assessment

Factor	Online self-assessment	Guided assessment
Effort	An hour or two	A few weeks
Evidence	Opinions of whoever answers	Interviews, system and data review
Output	Generic score	Scored gaps, ranked use cases, roadmap
Best for	Starting a conversation	Budget and investment decisions

What a Good Assessment Report Contains

The output should be short enough for leadership to read and specific enough for teams to act on:

• A maturity score for each area, with evidence rather than opinions.
• A ranked list of use cases with expected benefits, data needs and risks.
• A gap list covering data clean-up, integration, infrastructure and skills.
• A governance starter kit: an AI use policy, a model approval checklist and roles for oversight.
• A 6 to 12 month roadmap with milestones and success measures.
• A budget range based on the factors that drive cost, such as data work, licences and people.

From assessment to first project

Pick a first project with a clear owner, available data and a measurable outcome. Good early candidates include invoice and document data extraction, demand or inventory forecasting from sales history, routing of customer or citizen complaints, and internal search over policies and manuals. Our guide to generative AI use cases for Indian businesses has more ideas. Define success in business terms before the pilot starts, such as hours saved or fewer stock-outs, and agree in advance what result would justify scaling or stopping.

Public sector bodies and enterprises can also watch the IndiaAI Mission, approved by the Union Cabinet in March 2024 with an outlay of Rs 10,371.92 crore to build compute capacity, datasets and skills in India. Shared infrastructure and datasets may lower the entry cost of some projects over time.

Frequently Asked Questions

What is an AI readiness assessment?

It is a structured review of an organisation's strategy, data, technology, skills, governance and culture to judge whether it can adopt AI successfully, and what must change first.

How long does an AI readiness assessment take?

For a mid-sized organisation, a few weeks is typical. Larger enterprises with many business units may take longer or assess one unit at a time.

Do we need clean data before starting AI?

You need data that is good enough for the chosen use case. The assessment shows which data is ready now and which needs work.

Who should own AI in an enterprise?

The business function that benefits should own each use case, supported by IT for platforms and a governance group for risk and policy.

Is an online AI readiness quiz enough?

It is a useful starting point, but it reflects the opinions of whoever answers it. Investment decisions need evidence from interviews, data and system reviews, which a guided assessment provides.

How Affix Center Can Help

Affix Center runs AI readiness assessments for enterprises and public sector organisations, covering strategy, data, technology, skills and governance. Our enterprise advisory team turns findings into a practical roadmap, and our data and AI team helps build the data pipelines, models and monitoring for the first projects. To plan an assessment for your organisation, contact us.