Every enterprise executive in the country is being asked the same question: "What's our AI strategy?" Most don't have a good answer yet, and that's fine. The bigger risk is having a bad answer too quickly.
What You Need to Know
- Generative AI is a genuine capability leap, not another blockchain. It produces human-quality text, code, and analysis at machine speed.
- The technology is real, but enterprise readiness is not. Consumer tools like ChatGPT don't address data security, integration, governance, or domain accuracy.
- The winning move right now is structured exploration, not rushed deployment. Companies that invest in understanding their data and processes will be best positioned in 12-18 months.
- AI won't replace your team, but it will reshape how your team works. The organisations that prepare for this shift now will adapt fastest.
- Most enterprise AI value comes from boring problems: document processing, knowledge retrieval, workflow automation. Not flashy chatbots.
100M
ChatGPT users in first two months, the fastest consumer app adoption in history
Source: UBS Global Research, January 2023
$4.4T
potential annual value from generative AI across industries
Source: McKinsey Global Institute, The Economic Potential of Generative AI, 2023
The Capability Is Real, But Context Matters
Let's be clear about what's happened. In November 2022, OpenAI released ChatGPT, a general-purpose language model with a conversational interface. By January 2023, it had 100 million monthly users. Microsoft followed with AI-powered Bing. Google scrambled to launch Bard. The technology industry hasn't moved this fast since the smartphone.
The underlying models (large language models, or LLMs, trained on massive text corpora) represent a genuine leap. They can generate human-quality prose, write functional code, summarise complex documents, and reason through multi-step problems. This isn't incremental. It's a new category of capability.
But there's a critical gap between "impressive demo" and "enterprise value." And most of the current conversation lives firmly on the demo side.
Three Questions Enterprise Leaders Should Be Asking
1. Where Does AI Create Value in Our Specific Business?
Not "what can AI do?" but "what problems do we have that AI might solve better than current approaches?" The answer is usually less glamorous than the headlines suggest.
Across our enterprise engagements, we consistently see the highest ROI in:
- Document processing and extraction - insurance claims, contracts, compliance reports
- Knowledge retrieval - finding answers across thousands of internal documents
- Workflow acceleration - routing, triage, first-draft generation
- Quality assurance - anomaly detection, consistency checking
These aren't exciting. They're valuable. And they're the foundation for more ambitious capabilities later.
2. Is Our Data Ready?
The single biggest predictor of AI success isn't the model you choose. It's the state of your data. Generative AI models are only as good as the context they're given. If your knowledge lives in 47 SharePoint sites, three legacy systems, and someone's email inbox, no model will magically solve that.
73%
of enterprise AI initiatives are delayed by data quality issues
Source: Gartner, Top Strategic Technology Trends for 2023, October 2022
The organisations that will be best positioned for AI in 18 months are the ones investing in data organisation, knowledge management, and system integration now. This is unsexy work. It's also the most valuable thing you can do.
3. What's Our Governance Framework?
Generative AI introduces new categories of risk that most enterprise risk frameworks don't cover:
- Hallucination - models produce confident, plausible, and completely wrong outputs
- Data leakage - consumer AI tools may train on your inputs
- Bias and fairness - models inherit biases from training data
- IP and attribution - who owns AI-generated content?
These aren't theoretical. They're practical issues that need practical governance. The companies building governance frameworks now, even simple ones, will move faster when they're ready to deploy.
What to Do Right Now
You don't need an AI strategy yet. You need a plan to develop one. Here's a practical starting framework:
Month 1-2: Explore and understand
- Let your team experiment with consumer AI tools (with clear data policies)
- Identify 5-10 processes where AI could add value
- Audit your data estate. Where is your knowledge, and how accessible is it?
Month 3-4: Prioritise and scope
- Rank opportunities by business impact × data readiness
- Identify your top 2-3 candidates for structured AI exploration
- Start conversations with potential AI partners (not vendors selling tools, but partners who understand enterprise delivery)
Month 5-6: Build capability
- Run a focused discovery on your top opportunity
- Build internal AI literacy across leadership and key teams
- Establish basic governance guardrails
The Foundation Mindset
The companies that compound AI value don't treat each capability as a separate project. They build shared infrastructure (document processing, knowledge bases, integration patterns) that makes every subsequent capability faster and cheaper. Start thinking about your AI foundation from day one.
What Not to Do
Don't panic-buy AI tools. The vendor market is shifting weekly. Any tool you buy today may be obsolete or irrelevant in 12 months. Invest in understanding your problems, not in solutions looking for problems.
Don't ban AI in your organisation. Your team is already using ChatGPT. Pretending otherwise creates shadow AI risk. Set sensible policies and embrace the learning.
Don't wait for AI to be "ready." It's ready enough to start exploring. The companies that wait for certainty will fall behind the companies that learn by doing.
Don't let vendors set your agenda. Every technology vendor has become an "AI company" in the last 90 days. Most of them are wrapping a ChatGPT API in their existing product. Evaluate capabilities, not marketing.
- Should we build our own AI or buy solutions?
- It's too early for most enterprises to commit to either. Start with exploration: understand your data, your problems, and the market. The build vs buy decision will become clearer once you know what you actually need.
- Is generative AI a security risk for enterprise?
- Consumer tools like ChatGPT can be, yes. Data entered into these tools may be used for training. Enterprise-grade AI solutions with proper data governance, private deployments, and access controls address these risks. The key is not avoiding AI, but deploying it with appropriate controls.
- How much should we budget for AI exploration?
- A structured AI discovery (mapping opportunities, assessing data readiness, and prioritising use cases) typically costs $20-50K and takes 4-6 weeks. It's the single best investment you can make before committing to any AI build or purchase.