AI & Technology
AI: The Most Efficient Way to Reach the Wrong ICP at Scale?
AI can help you move faster. Your ICP determines whether that speed creates value.
By Bill Ryan · 2026-10-12
“Our messaging isn’t resonating. Sales cycles are stalling. Customer acquisition costs are rising. Clearly, we’re on the right track. We just need more volume. Let’s use AI to do more of this.”
Nobody would say it quite that way in a leadership meeting. But when the response to declining performance is to automate more outreach without questioning who you are targeting, that is effectively the strategy.
AI is an excellent tool for reaching the wrong ideal customer profile at scale.
My co-founder and partner, Fred Soller, phrases it well: “If your ICP is wrong, AI can help you contact the wrong prospects dramatically faster. If your messaging does not resonate, AI can help thousands more people ignore it.”
Before turning up the volume, ask whether you have current evidence that you are aiming at the right customer.
Will Meinhardt at Mach 1 shares a strategy for verifying that evidence: “Before you ask AI to reach more prospects, use it to understand why your best customers bought, why others walked away, and what has changed since. That learning should shape your campaigns.”
More activity does not establish customer fit
AI can research accounts, generate content, personalize messages, and execute campaigns faster. Those capabilities have real value. They also make it easier to scale a commercial motion before validating its underlying assumptions.
The consequences extend beyond wasted time. Gartner reported that 73% of surveyed B2B buyers actively avoid suppliers that send irrelevant outreach.[1]
My takeaway: scaling irrelevant outreach can damage the relationships you are trying to build.
A message can reference someone’s company, industry, and latest announcement and still miss the problem they care about. Personalization is useful. Relevance requires understanding why the buyer would act.
Longer sales cycles, weak engagement, and rising acquisition costs deserve investigation. They can reflect problems with targeting, positioning, pricing, channels, or execution. Increasing activity before diagnosing the problem risks making it more expensive.
Your ICP can change while your account list stays the same
An ideal customer profile defines the organizations most likely to buy, achieve meaningful value from your offering, and generate profitable, repeatable revenue for your business.
Industry, company size, and technology stack help identify those organizations. The problem they need to solve, the alternatives available, and their ability to realize value establish fit. Buying triggers and urgency help determine when to pursue them.
The assumptions behind that profile need ongoing validation.
Companies recognize that new ICPs can emerge. We spend less time considering how the same market forces can weaken an existing ICP or eliminate the buying conditions that made a segment attractive.
Budgets change. Regulations change. Competitors introduce new alternatives. Capabilities that once required an outside vendor become available internally. Your own product and pricing evolve.
The account may still meet every demographic requirement in your CRM. Its reason to buy may have changed entirely.
Consider workflow software. McKinsey’s August 2026 State of AI survey found that 32% of respondents said their organizations had decided against buying at least one software product or feature because they could build the functionality internally with agentic coding tools.[2]
That finding does not establish how often ICPs drift. It illustrates a mechanism that can change customer fit: a buyer now has an alternative that your original targeting assumptions did not account for.
A segment can remain attractive for complex workflows while becoming harder to serve profitably for simpler ones. The industry has not disappeared. The opportunity within it has changed.
AI can help you discover the change
Using AI to execute an outdated ICP will not resolve that problem. Using it to investigate the ICP can help.
AI can analyze sales conversations, organize customer feedback, compare recent cohorts, and surface patterns across wins, losses, stalled opportunities, and churn.
But the quality of the evidence matters. Bain’s 2025 commercial excellence research found that slightly more than half of commercial organizations lacked adequate technology and data foundations to optimize AI.[3]
An AI-generated profile built from incomplete CRM records can give weak assumptions a convincing presentation. A profile based only on historical wins can reinforce yesterday’s strategy.
The accounts you won through founder relationships may not represent a repeatable market. The customers who signed but never achieved value may expose a fit problem that the sales data alone misses.
Use AI to generate hypotheses. Verify them through source records, customer conversations, and commercial tests.
Build a process that keeps the ICP current
A commercial process is the repeatable sequence of actions, decisions, and handoffs that moves the right customers from initial engagement through purchase, successful adoption, and expansion.
That process should also feed customer evidence back into your targeting decisions. Here is how I would make that happen.
Define the buying conditions. Document the problem, its economic impact, the buying group, the alternatives, and the conditions that make your offer valuable. Include disqualifiers. A useful ICP tells the team where to focus and when to walk away.
Analyze outcomes by segment. Compare win rates, no-decisions, sales cycles, acquisition cost, implementation effort, retention, expansion, and cost to serve. Separate recent cohorts from older ones. Look for segments that buy and succeed economically for both parties.
Talk to customers and noncustomers. Interview recent wins, losses, stalled deals, and customers who left. Ask what prompted the search, what alternatives they considered, and why they acted or delayed. AI can organize the evidence; the conversations test its interpretation.
Test a specific hypothesis before scaling it. Identify a segment, a problem, a relevant offer, and the evidence that would justify further investment. Define success and stop criteria beforehand. Replies and meetings are early signals. Qualified progression, paid commitments, and successful adoption provide stronger evidence. Allow enough time for the buying cycle.
Review fit regularly. Bain’s segmentation guidance recommends measuring segment performance and adjusting as market conditions change.[4] My recommendation is a quarterly review, with an earlier review after a meaningful market, competitive, regulatory, or product change. Include sales, marketing, product, customer success, and finance.
Give AI a validated direction
At Big Wheel Performance, our approach starts with market validation, moves into commercial architecture, and then scales revenue execution. Customer outcomes should keep informing each of those decisions.
AI can support the entire cycle. It can help you learn faster, recognize changes earlier, and execute a validated strategy more effectively.
AI magnifies structural quality. It does not compensate for its absence.
Before asking how much more AI can help your team do, ask whether you have current evidence that the work deserves to be scaled.
AI can help you move faster. Your ICP determines whether that speed creates value.
Sources
[1] Gartner: B2B buyer preferences and irrelevant outreach
[2] McKinsey: The State of AI in 2026
[3] Bain: Parsing How Winners Use AI
[4] Bain: Customer Segmentation
Bill Ryan is a Co-Founder & Managing Partner at Big Wheel Performance. He helps enterprise sales and SaaS companies build capital-efficient growth by aligning sales, marketing, and customer success into flywheels that scale beyond founder dependence.