AI will make some customers need you less, but only for the parts of your work they can now do themselves. That is our view at NTx. The useful question is not whether AI threatens your business in general. It is which specific pieces of what you sell are still essential, and whether the rest of your business could absorb losing the pieces that are not.

A new framework from two business leaders is a useful way to work through it. In HBR's published summary of "The Market Is Mispricing Resilience" (November 13, 2026), Hemant Taneja, CEO of General Catalyst, and Bob Sternfels, McKinsey's global managing partner, argue that as companies face AI-driven displacement, supply-chain shocks, and geopolitical instability, resilience is an increasingly important and underpriced growth asset. They name four dimensions that help investors assess it: demand curve, supply chain, capital stack, and system.

Their piece is written for investors, and nothing here is investment advice. Turned inward, the four dimensions become four questions an owner can answer about their own business. The questions and advice below are NTx's, not HBR's.

Overhead view of six colleagues meeting around a long wooden table with laptops, papers, and charts in a bright office.

Demand curve: could your customer's AI now do what they pay you for?

HBR defines demand-curve resilience as how essential and entrenched a company's offerings are. NTx's test for owners: which of the tasks your customers pay for could their own AI tools now do without you?

The early evidence comes from freelance and agency markets, not from proven revenue loss at privately held businesses. A peer-reviewed study in Management Science by Demirci, Hannane and Zhu (January 2025) found that on a large freelancing platform, job posts for automation-prone writing and coding work fell 21% within eight months of ChatGPT's release, compared with manual-intensive jobs, and image-creation posts fell 17% after image-generating AI arrived. Forrester reported on July 15, 2026 that the share of B2B marketers expecting agency budget increases fell 13 percentage points from 2025 to 2026. In OpenAI research with Upwork, content writing's share of Upwork postings fell roughly 31% from the first half of 2025 to the first half of 2026, though the report adds that AI may be changing the scope of work without reducing overall labor demand. Buyers are also building instead of buying: in McKinsey's State of AI 2026 survey, mostly of larger organizations, 32% of respondents said their organizations had decided against buying software they could build internally with agentic coding tools.

The broader picture is calmer: a U.S. Census Bureau working paper (CES-WP-26-25) found AI-related employment decreases in only 2% of firms. Customers stop needing particular tasks, not you all at once. In our experience, judgment, accountability, and relationships are harder for a customer's AI to replace.

NTx advice: List your revenue by line, then ask two questions of each.

  • Is this a task a customer could hand to an AI tool they already use, with an acceptable result?
  • If they did, what would they still need from us?

Lines where the customer would still need a lot from you are entrenched. Lines where they would need little are exposed, even if those customers have been loyal for years.

Two more pressures are worth watching. Clients who know you use AI are starting to ask for the savings in your price, which we covered on November 18, 2026 in Your customers know you're using AI. They're going to ask for the discount. And tools announced on November 13, 2026 let a customer's AI agent identify itself and, in a limited rollout, pay: Sierra's Personal Agent Protocol is a protocol for how an agent connects with a business, and ChatGPT Wallet launched only for ChatGPT Pro subscribers in the US. Agent orders are still rare, as we noted in The agent can now walk in, show ID and pay. Is your door open? Keep your list of exposed lines. The capital-stack question depends on it.

Supply chain: if one vendor stops, can you still deliver?

HBR ties supply-chain resilience to redundancy and visibility. NTx asks owners to trace which suppliers, software, and AI providers each customer promise runs through, and whether each one has a second path.

In the Federal Reserve Banks' 2025 Small Business Credit Survey of 6,525 employer firms, released March 3, 2026, 48% sourced at least some inputs from outside the US, and 84% of those said their inputs cost more. In NFIB's October 2026 Small Business Economic Trends survey of its members, 65% of owners said supply chain disruptions affected their business to some extent.

For a privately held business with a remote team, the supply chain is often the services the work runs on: the managed service provider (MSP) that looks after your computers, remote-access tools, the AI providers built into daily workflows, and the contractors who handle specialized steps.

NTx advice: Pick your three most important customer commitments and map each one end to end.

  • Visibility: Which outside vendors, tools, and providers touch this work? Do you know who your key vendors depend on in turn?
  • Redundancy: If one stopped working tomorrow, is there a second option, and how long would switching take?
  • Ownership: Who on your team would notice first, and who has the authority to act?

Vendor failure is not only an outage risk, as we showed in Your Vendors Are Your Attack Surface: Four Breaches That Started Somewhere Else. A vendor you cannot see into is a dependency you cannot plan around.

Capital stack: will the revenue last as long as the bills?

HBR says capital-stack resilience aligns debt with the durability of revenue. NTx asks owners to compare the length of their loans, leases, and other fixed commitments with how long each revenue line is likely to survive the demand-curve test.

Your capital stack is your bank loans, equipment and office leases, multi-year software contracts, payroll, and owner draws the household relies on. In the same Federal Reserve survey, 38% of small employer firms held more than $100,000 in debt. Those bills are fixed. The revenue paying for them may not be.

NTx advice: Put two lists side by side.

  • Your commitments: each loan, lease, and long contract, with its monthly cost and end date.
  • Your revenue lines: each one marked entrenched or exposed, with a rough view of how long it will hold at today's price.

Then look for mismatches. A five-year lease funded mostly by work a customer's AI could soon do is a risk, even if that work is profitable today. So is a loan sized to this year's revenue if part of that revenue is likely to be repriced. The fixes are ordinary business decisions: shorter terms where revenue is uncertain, less fixed cost tied to exposed lines, and more cash from entrenched lines kept as an operating reserve.

System: what shared infrastructure could stop your work?

HBR describes system resilience as strengthening the critical infrastructure businesses depend on. NTx asks owners which shared systems they assume will always work, such as power, connectivity, cloud services, and AI platforms, and what happens if one fails.

A remote team relies on home internet, the electricity behind it, the cloud platforms where files and email live, suites such as Microsoft 365, and increasingly the AI services built into all of them. Dependence on AI is now real for small firms: in the U.S. Chamber of Commerce Technology Engagement Center's 2026 small business report, a survey of 3,732 small businesses, 66% use AI, and 78% of those users say limiting access would slow their growth. Outages are costly too. The Uptime Institute's 2026 outage analysis, which covers data center operators and enterprises rather than small businesses, found power behind 45% of respondents' most recent impactful outages, and 57% said their most recent major outage cost more than $100,000.

NTx advice: Run a short tabletop exercise with your leadership team.

  • If your main cloud or collaboration platform were down for a full working day, what could your team still do, and what would stop?
  • If your primary AI provider changed its terms, its price, or its availability, which workflows would break?
  • If a region where several team members live lost power or connectivity for a few days, who would cover their work?

You cannot fix shared infrastructure yourself. You can decide in advance what you would do without it, keep critical files and contacts reachable outside a single platform, and avoid building every important workflow on one provider. Together, your four answers form the self-check: where customers still need you, where delivery depends on others, where the bills could outlast the revenue, and where the ground itself could shift.

Sources

We can help you with that.

NTx AI Consulting works with owner-led and partner-led privately held businesses that are currently using AI tools in everyday workflows, or would like to be. The complimentary consultation is a working session, not a product demo. If a deeper engagement makes sense, Phase 1 is an AI Disruption Readiness Audit. You can learn more at https://ntx.ventures/ai-consulting/.