Put AI to work in your investing strategy

A conversation with Chris Patterson, Head of AI Solutions at CIBC, on using AI to ask smarter questions and make sharper investing decisions

AI tools are changing how investors discover ideas, digest information, and make decisions. But where do you start if you’re a novice? And how can an intermediate investor get more rigorous — basically, more focused, fact-driven and consistent — faster. We sat down with Chris  Patterson, Head of AI Solutions at CIBC and passionate investor, to discuss how he uses AI in his own process, what it’s great at, where it can go wrong, and practical prompts anyone can try.

Chris Patterson: Recently, I was rebalancing to create a more stable income portfolio when yields on cash-like products fell. I wanted dividend payers with sustainable payouts. AI explained dividend sustainability in plain English — it looked at and broke down payout ratios, cash flow coverage, and how economic cycles can affect earnings. That gave me the confidence I needed to go ahead and make the changes.

Where AI really helped was translating dense materials — financial statements, management commentary — into plain language that helped answer some of the specific questions I asked.

Chris Patterson: I guided it. That’s key. AI is predictive — it will do its best with the question you ask. The more specific your question, the better the answer. If you simply ask, “Is this a buy or a sell?” you’ll often get a vague or sentiment-driven response. If you ask, “How well was this company’s dividend covered by earnings and free cash flow over the last 3 years?” you’ll get something far more useful.

Chris Patterson: Plenty. Suppose you’re looking at a commodity stock in a cyclical sector, such as an oil stock. You can ask, “If the oil price averages X, how could that affect revenue, margins and dividend safety?” AI can infer from historical results and management’s forward-looking commentary — as long as you supply or reference those sources. It’s not a crystal ball, but it’s excellent at framing plausible ranges and their implications.

Chris Patterson: It’s very good at digesting earnings releases and call transcripts. I’ll ask it to summarize the quarter in plain English, flag any changes in guidance and highlight risks that management or analysts emphasized. If a stock moves in a way that seems unintuitive versus the reported numbers, AI can often point to something in the Q and A or broader news sentiment that explains the move. It’s a time saver and can surface things you might miss.

Chris Patterson: Start top down. Perhaps ask at the sector level first. For example:

• “Given today’s macro environment, which sectors might suit a risk profile that’s [conservative, moderate, aggressive] and a time horizon that’s [insert your time horizon]?”
• “What are the typical drivers of performance in [sector], and which metrics matter most?”

Once you’ve picked a sector, drill down:

• “Provide a shortlist of companies or diversified ETFs or mutual funds in [global or region] within [sector], with high-level fundamentals: valuation, growth and income characteristics.”
• “Explain the key risks to this sector over the next 12 to 24 months and what would change the investment outlook.”

The point is to let AI build you a map — then you choose the route.

Chris Patterson: It’s usually not the best question. You’ll often get a hedged response or something influenced by current chatter. A better prompt is: “Give me the bull and bear case for [company or ETF] based on fundamentals and credible sources. List assumptions and what would invalidate each case.” That educates you while keeping the decision yours.

Chris Patterson: Try these, and tailor for your situation. 

• “I’m a [conservative, moderate, aggressive] investor with a [time horizon]. Outline 3 or 4 sectors that could fit, with the key drivers, risks and typical metrics to watch. Keep it plain language.”
• “Summarize the last 2 annual reports and the most recent earnings call for [company or ETF]. What changed in guidance, what risks were discussed, and what should an investor monitor next quarter? Cite sources.”
• “For [sector], list the top 5 factors that most influence revenue and margins. Explain how changes in [interest rates, input costs, trade policy] typically flow through to earnings.”
• “Provide the bull and bear case for [company or ETF]. State assumptions, catalysts and what would invalidate each case. No price targets; focus on business drivers.”
• “Build 3 simple scenarios — base, upside, downside — for [company or ETF] over the next 2 years using conservative assumptions. Explain the operational or macro conditions behind each scenario.”

Chris Patterson: Definitely, I avoid:

• Forcing answers: “Give me five 10% yielders no matter what.” That invites hallucinations.
• Vague advice: “What should I buy today?” Ask for frameworks and drivers instead.
• Blind trust in math-heavy outputs: Always verify calculations.
• Treating AI as an oracle: It’s a copilot — strong at synthesis, weak at certainty.

Chris Patterson: Be careful. It’s easy to slip from investing into gambling, especially with instruments designed for leverage. AI won’t change the risk profile of those products. If you use them, do so within a disciplined plan and risk limits — and understand that many users lose money on short-dated speculation. For most investors, focusing AI on research, scenario analysis and portfolio education is a better use.