88% of Executives Plan to Increase AI Budgets

The £5 Million Question

Board meeting. CTO presents AI strategy. Charts, examples, jargon. £5 million approved. I ask one question: what specific business outcome is this supposed to achieve?

Then I asked one simple question: what specific business outcome is this £5 million supposed to achieve?

Silence. The CTO looked at the CEO. The CEO looked at the CFO. The CFO did not know what to say. They had all agreed that AI was important. They had all agreed that they needed to invest. But nobody could articulate what the £5 million was supposed to deliver.

This is happening in hundreds of organizations right now. According to PwC, 88 percent of executives plan to increase AI budgets due to agentic AI. But most executives cannot tell you what they are buying.

This is dangerous. Lots of money with no clear strategy is just waste with a different name.

The Three Types of AI (and What They Actually Do)

If you do not understand the three types of AI, you cannot evaluate AI investments. So let me explain them clearly.

Type 1: Automation AI

This AI replaces human work. A person used to do X. Now a machine does X. Examples: document processing, data entry, invoice processing, customer service.

The outcome: cost reduction. You spend less on labor because machines do the work. This is valuable. Cost reduction is real. But it is not growth. It is not new revenue. It is doing what you did before, cheaper.

Budget impact: This should be a capital expenditure with a clear ROI. If a task costs £100,000 per year in labor and AI automation costs £20,000 per year, the payback is one year. This is easy to justify.

Type 2: Augmentation AI

This AI works alongside humans to make them better. A person still does the work, but AI handles parts of it. Examples: writing assistance, analysis support, research acceleration, design generation.

The outcome: productivity improvement. Your people get more done. They produce higher quality work. They spend less time on routine stuff and more time on judgment.

Budget impact: This is harder to measure than automation. How much faster is your analyst? How much better is their analysis? But if it is real, the payback is significant. A consultant who used to take one week for a project now takes three days. You can do more projects with the same team.

Type 3: Agent AI

This AI makes decisions and takes actions autonomously. You set the parameters. The AI operates within those parameters. Examples: autonomous customer service decisions, automated approvals, autonomous campaign optimization.

The outcome: both cost reduction and new capabilities. You reduce human decision-making (cost) but also you make more decisions per day (new capability). You can serve more customers. You can respond faster.

Budget impact: This is the highest-risk category. You are giving the AI decision authority. If the AI makes bad decisions, you have a liability problem. But if it works, the upside is large.

What You Should Know Before You Write a Check

Here are the five things an executive needs to understand before committing AI budget.

Question 1: Which Type of AI Are We Buying?

Is this automation? Augmentation? Agent? Or combination?

Different types have different ROI profiles, different timelines, different risks. You need to know what you are buying.

Practical tip: In your next AI presentation, ask the CTO: which of the three types of AI is this? If they cannot answer clearly, you are not ready to commit budget.

Question 2: What Is the Specific Business Outcome?

Not "improve customer service." Specific. "Reduce customer service cost by 30 percent" or "increase customer satisfaction from 75 percent to 85 percent."

If you cannot define the outcome in numbers, you cannot measure success. And if you cannot measure success, you are just spending money.

Question 3: What Is the Baseline?

Before you deploy AI, understand the current state. How long does it currently take? How many people does it currently require? What does it currently cost? What is the current quality?

If you do not know the baseline, you cannot measure improvement. And you certainly cannot tell if the AI actually helped.

Question 4: What Is the Risk?

For automation AI: what if the AI fails? For augmentation AI: what if nobody uses it? For agent AI: what if the agent makes a bad decision?

Every AI project has risks. You need to understand them. You need mitigation plans. If the CTO cannot articulate the risks, they have not thought this through.

Question 5: What Is the Timeline to Value?

Is this a 30-day project or a 12-month project? Is value delivered upfront or gradually?

The best AI projects have clear milestone. Pilot phase. Measurement phase. Scale phase. If you do not have milestones, you have an open-ended commitment.

The Executive AI Literacy Framework

You need to understand five things. Not deeply. But well enough to ask good questions and evaluate investments.

Understand Generative AI vs Predictive AI

Generative AI creates new content. It writes. It generates code. It generates images. It generates strategies.

Predictive AI forecasts. It predicts customer churn. It predicts loan defaults. It predicts equipment failures.

Most current excitement is around generative AI. But predictive AI is where business value has been for years. You need to know the difference because they have different applications and ROIs.

Understand Hallucinations and Accuracy

Generative AI sometimes makes things up. It will generate plausible-sounding text that is completely false. It "hallucinates."

This matters for different applications differently. If you are using AI to brainstorm marketing ideas, hallucinations do not matter (bad ideas are just ideas). If you are using AI to generate compliance documents, hallucinations are a liability disaster.

Know your application. Know whether accuracy matters. Design safeguards accordingly.

Understand Data Privacy and Security

Most generative AI tools are cloud-based. If you send your data to a cloud API, who owns it? Can it be used to train models? Is it encrypted? What if there is a breach?

These are not technical questions. These are business questions. Your legal and compliance teams need to evaluate. You need to understand the implications.

Understand Cost Dynamics

Some AI platforms charge per token (per bit of text processed). Some charge per user. Some charge per transaction. Some are subscription.

These cost structures matter. If you are processing millions of documents, per-token pricing might be expensive. If you have hundreds of users, per-user pricing might be expensive. You need to understand how costs scale.

Understand Governance and Accountability

If an AI system makes a decision that goes wrong, who is responsible? This is a business question, not a technology question. But you need clarity before you deploy agents.

What decisions can the AI make autonomously? What decisions need human approval? What audit trails exist? These need to be defined before deployment.

The Vendor Evaluation Questions

When a vendor pitches you an AI solution, ask these questions:

What problem does this solve? Be specific. Not "improve efficiency." What problem?

How would you measure if it worked? What metrics?

What is your evidence that it works? Case studies? Benchmarks?

How much does it cost? What is included? What are the lock-in terms?

What if it does not work? Can you exit? What is the cost?

What are the security and privacy implications? How is data handled?

What training and support is included? How long is the learning curve?

Practical tip: If the vendor cannot answer these questions clearly, do not buy. The vendors worth buying from can answer these questions in 15 minutes.

The Budget Allocation Framework

If you are allocating £5 million to AI, how should you split it?

I recommend: 40 percent to automation projects (clear ROI, quick wins), 40 percent to augmentation projects (longer-term productivity), 20 percent to experimentation (learning what works).

This allocation balances quick wins with longer-term capability building with learning. You get some immediate ROI. You build sustainable capability. You learn what works for your organization.

The Governance Model That Works

Who decides which AI projects get funded? Who measures results? Who has authority to shut down a project that is not working?

The best organizations I see have an AI steering committee. Not a technology committee. A business committee. They include representatives from business units that have skin in the game. They review projects quarterly. They measure ROI. They reallocate budget based on results.

This is not a technology governance structure. It is a capital allocation structure. AI projects compete for budget like any other investment. The best ones get more. The bad ones get discontinued.

The Correlation With Bottom-Line Impact

I have seen data showing a correlation between strong AI governance and actual business impact. Companies with poor governance spend more and get less. Companies with strong governance spend less and get more.

The difference is discipline. Clarity on what you are buying. Measurement of whether it works. Willingness to stop funding things that do not work. Allocation of resources to things that do.

The Board Conversation You Need to Have

If you are a board member, ask your CEO: do we have a clear AI strategy? What are we investing in AI? What are the expected returns? How are we measuring? What if it is not working? How do we reallocate?

If the CEO cannot answer these clearly, you do not have an AI strategy. You have a bet. And bets are how you waste money.

Richard Batt has delivered 120+ AI and automation projects across 15+ industries. He helps businesses deploy AI that actually works, with battle-tested tools, templates, and implementation roadmaps. Featured in InfoWorld and WSJ.

Frequently Asked Questions

How long does it take to build AI automation in a small business?

Most single-process automations take 1-5 days to build and start delivering ROI within 30-90 days. Complex multi-system integrations take 2-8 weeks. The key is starting with one well-defined process, proving the value, then expanding.

Do I need technical skills to automate business processes?

Not for most automations. Tools like Zapier, Make.com, and N8N use visual builders that require no coding. About 80% of small business automation can be done without a developer. For the remaining 20%, you need someone comfortable with APIs and basic scripting.

Where should a business start with AI implementation?

Start with a process audit. Identify tasks that are high-volume, rule-based, and time-consuming. The best first automation is one that saves measurable time within 30 days. Across 120+ projects, the highest-ROI starting points are usually customer onboarding, invoice processing, and report generation.

How do I calculate ROI on an AI investment?

Measure the hours spent on the process before automation, multiply by fully loaded hourly cost, then subtract the tool cost. Most small business automations cost £50-500/month and save 5-20 hours per week. That typically means 300-1000% ROI in year one.

Which AI tools are best for business use in 2026?

It depends on the use case. For content and communication, Claude and ChatGPT lead. For data analysis, Gemini and GPT work well with spreadsheets. For automation, Zapier, Make.com, and N8N connect AI to your existing tools. The best tool is the one your team will actually use and maintain.

What Should You Do Next?

If you are not sure where AI fits in your business, start with a roadmap. I will assess your operations, identify the highest-ROI automation opportunities, and give you a step-by-step plan you can act on immediately. No jargon. No fluff. Just a clear path forward built from 120+ real implementations.

Book Your AI Roadmap, 60 minutes that will save you months of guessing.

Already know what you need to build? The AI Ops Vault has the templates, prompts, and workflows to get it done this week.