Phase 2 Grow: Using AI to Capture Revenue, Not Just Save Time (“The Staircase Methodology”)
This is the final article in the series on the Staircase Methodology, a practical way for small and mid-sized businesses to choose, ship, and measure AI projects on a modest budget. Earlier articles covered why SMBs get stuck on AI, a real case study, and the full Phase 1 (Optimize) workflow. This one covers Phase 2 Grow.
If you are landing here first, the short version is this. The methodology is a staircase, not a leap. Phase 1, Optimize, uses AI to save time and cost on work you already do. Phase 2, Grow, uses AI to capture revenue that is currently slipping past you. Each phase runs through the same six steps: set a dated goal, brain-dump candidates, narrow them to the ones that match the goal, score and rank what is left, run the top items through four checks, then commit in writing to a measurable outcome with a check-in date.
The good news for Phase 2 is that you already know the staircase. Walk through every step in the same order you did in Phase 1. Only two things really change: the question you ask in the brain dump, and the formula you use to score. Everything else works the same way. The two printable worksheets are the same as Phase 1, with the scoring formula swapped, and I have attached them to this article so you can follow along.
Why Grow Is Phase 2, Not Phase 1
Growth-mode AI is more ambiguous, more expensive, and harder to measure than optimization. That is exactly why it comes second. Doing optimization first means you have already learned what AI can realistically do for your business, so your growth bets become informed bets rather than blind ones. If you have not shipped a couple of optimization projects yet, it is usually worth starting there, unless your business is the kind where growth wins clearly dwarf any time-saving wins, which the Phase 1 article covers in more detail.
Step 1: Set a Goal
Same idea as Phase 1, a specific and dated goal, but pointed at growth. Pick one and finish it. “In 3 months, I want to add ___ more customers, leads, or dollars from ___.” Or, “In 3 months, I want to launch ___ that I currently do not have time to do.” Or, “In 3 months, I want to understand ___ about my customers or business that I currently do not.” The blanks still do the work; a vague growth goal filters nothing.
Step 2: Brain Dump, With a Different Question
This is the first real change. In Phase 1 you asked, what hurts? In Phase 2 you ask, where am I leaving money on the table? The list looks different as a result: customers you are not following up with, data you are collecting but never analyzing, marketing you do not have time to do, audiences you cannot reach, products or services you could offer if you had more capacity. As before, list freely without filtering yet.
Step 3: Connect the List to the Goal
Same as Phase 1. Circle the items that would directly help reach your growth goal, and cross out the rest. The goal is your filter.
Step 4: Score With Hidden Revenue
This is the second real change, and the important one. Phase 1 scored each item with Bleed = Pain × Frequency. Phase 2 scores with Hidden Revenue = Estimated Annual Value of the Missed Opportunity × Probability You Can Capture It With This Project.
For each circled item, estimate two things. First, the annual dollar value of the opportunity you are not capturing today, such as dormant customers worth re-engaging or leads you never follow up on. Second, the probability that this specific project will actually capture that value, on a 0 to 1 scale, where 0.1 is unlikely, 0.5 is a coin flip, and 0.9 is very likely. Multiply them.
The probability factor is what keeps this honest. Most SMBs are dazzled by the headline opportunity number, but a $200,000 opportunity with a 10% chance of capture is a $20,000 expected value, not $200,000. That is still a meaningful number for a small business; it is just an honest one. By the end of this step you should have your top three to five candidates.
Step 5: The Same Four Checks
The four checks are structurally identical to Phase 1, so I will keep this brief and flag only what differs.
Check 1, can AI actually do this? Same as Phase 1. AI handles drafting, summarizing, structured extraction, analysis, pattern recognition, and the like; it struggles with physical tasks, regulated decisions without review, and deep judgment. “Partially,” meaning AI drafts and a human approves, still counts as a yes with a checkpoint built in.
Check 2, how much effort? Same brackets as Phase 1: Small is a few days, Medium about a week, Large about a month, Extra-large is more than a month and usually best skipped on a first attempt. The one difference is that the Bleed-based size thresholds from Phase 1 do not apply here, because Hidden Revenue is measured in dollars rather than on a 1-to-25 ranking scale. In Phase 2, Check 3 does the heavy lifting on whether a higher-effort project is worth it. As always, if fewer than two candidates remain, loop back to Step 4 and pull in more items.
Check 3, return on investment as payback. Same formula as Phase 1, with one substitution: use monthly revenue gain instead of hours saved. Payback in months = Implementation Cost ÷ Monthly Revenue Gain. The thresholds are unchanged: under 6 months is good to go, 6 to 12 months is still a yes with planning, and over 12 months means picking a different item.
Check 4, reversibility and risk. Same formula, Risk = Impact × Stickiness, with Stickiness being the worse of its technical and reputational dimensions, and the same thresholds, where 10 or above means STOP and add a human checkpoint. But Phase 2 deserves extra attention here, and this is the most important caution in the whole article.
The Phase 2 Risk Trap. Growth projects are usually customer-facing by definition. You are reaching out to dormant customers, sending marketing, making offers, putting AI-generated communication in front of real people. That is exactly the situation where Reputational Stickiness is at its worst, because once an output has reached a customer, you cannot take it back.
A growth project that races past a high Risk score without a human checkpoint is the most common way Phase 2 projects damage trust with customers. The fix is the same one from the case study in the earlier articles: read the output before it goes out, or add an approval step, which drops the reputational risk and usually clears the project on safer terms. The cost of that checkpoint is small. The cost of skipping it can be a public mistake that undoes the revenue you were chasing.
Step 6: Commit With a Metric, Then Ship
Same as Phase 1. Commit three things in writing: one specific success metric, a check-in date 2 to 8 weeks out, and an honest answer to what this project replaces. That last one matters even more in Phase 2, because growth experiments compete for attention with the revenue-generating work you are already doing.
Growth-flavored metrics tend to look like these: “re-engage at least 50 dormant customers and convert 10% within 8 weeks,” or “increase the marketing email click-through rate from 2% to 4% within 6 weeks,” or “identify the top three customer segments by lifetime value within 4 weeks.” Then ship it, measure against the metric, and iterate.
The Series Summary
That is the whole Staircase Methodology. Phase 1 teaches you what AI can do for your business while saving time and cost at low risk. Phase 2 turns that hard-won knowledge toward revenue. The steps are the same; the inputs change, and the risk discipline tightens because the stakes are more customer-facing. Climbing one step at a time, with a written commitment and a check-in date on every project, is a defensible way to close the AI gap on your own terms rather than betting the whole business on it.
The full “The Staircase Methodology for AI Adoption in US Small and Mid-Sized Businesses”, with complete references, is on SSRN for the deeper version. This series has been the practical, faster-to-apply companion to it.
First article “Why Most Small Businesses Get Stuck on AI, and How the New Staircase Methodology Helps”.
Second article “The Staircase Methodology at Work: How One Small Business Picked Its First AI Project (Case Study)”.
Previous (third) article “Phase 1 Optimize: A Practical Way to Pick Your First AI Project (“The Staircase Methodology”)”.
Note: the views expressed in this article are my own and do not represent the official positions of any past, present, or future employers, clients or stakeholders.
