Getting started
Is AI a threat to your business? A practical check for mid-size companies
By SourceX Editorial · Updated
Short answer
AI is a threat to a mid-size business when customers pay mainly for work software can now draft, sort or answer, and when the company's know-how lives in people rather than records. Check five things by service line: what customers pay for, routine work, where expertise lives, switching costs and record depth. Deep, owned records strengthen every answer.
Key takeaways
- AI exposure depends on what customers pay for, not on the industry label.
- Work that ends in physical delivery, licensed judgment or accountable sign-off is harder to replace than drafting or sorting.
- Companies whose expertise is written down in records can train staff, tools and partners faster than those that rely on memory.
- The same records that help a company respond to AI can sometimes be licensed for a defined use.
- Run the check by service line, because one company often has exposed and protected lines side by side.
Is AI a threat to your whole business or to some tasks?#
AI is usually a threat to specific tasks inside a mid-size business before it threatens the whole business. A firm that writes proposals, answers routine support questions and schedules crews may see the first two change quickly while the third barely moves.
The useful question is which parts of what customers pay for a competitor could deliver with AI tools at lower cost or higher speed. Those parts are exposed. Parts that need site visits, licensed sign-off, physical goods or long relationships are less so.
The check below works by service line rather than by company. Run it with the leaders who know each line's margins, customers and delivery model.
The five-question exposure check#
The five-question exposure check asks what customers buy, how routine the work is, where expertise lives, what keeps customers from switching and how deep your records go. Each answer comes with a record-position reading, which shows whether your own history helps you respond.
| Question | Higher exposure answer | Lower exposure answer | What your record position means |
|---|---|---|---|
| What do customers pay for? | Documents, answers or analysis delivered on a screen | Physical work, goods or accountable sign-off | Records of past delivery show which steps can be automated and which cannot |
| How routine is the work? | Similar requests handled the same way each time | Each job differs and needs judgment | Ticket, job or project histories reveal how much is truly repeated |
| Where does expertise live? | In senior people's heads | In playbooks, notes and linked records | Written records let you train new staff and tools faster |
| What keeps customers from switching? | Price and convenience only | Integration, trust, contracts and site knowledge | Account histories show which relationships run deep |
| How deep and owned are your records? | Thin, scattered or client-controlled | Years of company-owned records with outcomes | Deep records support internal AI and may be licensable |
How to read your answers#
Reading the check works best as a simple count per service line: how many of the five answers fall in the higher-exposure column. A line with most answers on the exposed side needs a plan now; a line with few can be watched and revisited.
The record question deserves extra weight because it shapes every response. A company that can show its own history to a tool, a new hire or a partner adapts on its own terms; one that cannot is left buying generic tools and hoping they fit.
- Mostly lower exposure: keep operating, and capture expertise in records as experienced staff retire or leave.
- Mixed: automate the routine pieces yourself before a competitor uses them to undercut your price.
- Mostly higher exposure: redesign the offer around judgment, accountability or delivery, and price for outcomes.
- Thin records on any line: start capturing decisions and outcomes in the systems you already run.
Which mid-size businesses are more exposed?#
Mid-size businesses whose output is mainly text on a screen tend to be more exposed than those whose output is physical work. Document-heavy consulting, first-line support and routine back-office services feel pressure sooner, while HVAC service, freight handling and manufacturing change more slowly at the core.
The picture is rarely clean. A mechanical contractor's estimating desk and call center may change quickly even though its technicians' work does not. A software company may find that AI speeds up its own engineering while lowering the barrier for new competitors in its niche.
The table below shows how the split tends to look across common mid-size segments. Treat it as a starting point for your own check, not a verdict on any one company.
| Segment | Tasks under pressure | Work that holds up | Records that matter |
|---|---|---|---|
| B2B and vertical software | First-line support, routine code and documentation | Product judgment, customer-specific integration | Tickets linked to issues, code reviews and releases |
| Engineering and architecture | Drafting, building-code research, first-pass reviews | Stamped design decisions and site coordination | RFIs, submittals, review comments and change orders |
| Management consulting | Research, slide production, standard diagnostics | Facilitation, implementation and accountable results | Proposals, project reviews and playbooks |
| Home services and trades | Call handling, booking, estimate write-ups | Diagnosis and repair on site | Jobs linked to estimates, invoices and callbacks |
| Logistics and distribution | Order entry, quoting, exception emails | Physical handling, routing under constraints | Orders, exceptions and resolution notes |
| Manufacturing | Quote preparation, document control, reporting | Production, inspection and process engineering | NCRs, CAPAs, maintenance and warranty records |
Why your records decide how well you respond#
Records decide how well a company responds because AI tools are only as useful as the examples and context they work from. A company with years of support tickets linked to fixes, or estimates linked to final job costs, can test and tune tools against its own history before trusting them.
Records also hold the reasoning that leaves when senior staff retire or depart. Proposals with win and loss notes, RFIs with responses and quality reports with root causes are how a mid-size company keeps judgment in the building.
The same records may have outside value. AI developers building tools for an industry need examples of real work done well, and some license operational records for a defined use after personal and confidential details are removed.
Illustrative: a consulting firm runs the check#
Illustrative: a fictional management consulting firm focused on operations improvement runs the check on its three service lines: diagnostic reports, on-site implementation support and a training program for frontline supervisors.
Diagnostic reports score high on exposure. The deliverable is a document, much of the analysis follows a pattern and clients compare on price. Implementation support scores low, because it depends on site presence and accountable results. Training sits in between.
The firm's record position is uneven. Proposals, project reviews and playbooks sit in SharePoint and Notion with partner comments, but client deliverables belong to clients under most engagement letters. The managing partner rebuilds diagnostics around internal tools grounded in the firm's own playbooks, prices implementation on results and asks counsel which internal records could ever be licensed.
Mistakes that distort an AI threat assessment#
The most common mistake is treating AI exposure as an industry verdict rather than a task-by-task check. Industry headlines are too broad to act on, and they hide the service lines that are safe or even strengthened.
- Assessing the company as one unit instead of line by line.
- Assuming physical work is untouched while ignoring the office functions around it.
- Buying tools before checking whether your records can support them.
- Letting departing experts leave without capturing their decisions in records.
- Deleting old archives to save storage before anyone has reviewed what they show.
Where SourceX fits in an exposure review#
SourceX looks at the last question in the check: the depth and ownership of a company's records. It rates them with the SourceX Enterprise Data Value Framework on drivers such as uniqueness, domain expertise, human-generated signal, recency and rights, and notes reproducibility, preparation cost and privacy burden. A fit check uses metadata only, and the company decides whether to go further.
Frequently asked questions
Will AI replace my business entirely?
Wholesale replacement is rarely how the threat arrives. More often, specific services become cheaper for customers to do themselves or for competitors to offer, which squeezes price and volume line by line. Firms whose whole product is routine documents or answers face more direct pressure, so review exposure by service line on a regular schedule.
Who should run the AI exposure check?
The CEO should own it, with input from the leaders closest to customers and delivery: heads of service lines, sales, operations and finance. A CTO or IT lead adds a view on which tools are realistic. Keep the group small enough to finish one working session per service line.
Should we buy AI tools now or wait?
Start with the exposed tasks, and test tools against your own historical records before committing broadly. A pilot that uses closed tickets, past estimates or completed proposals shows whether a tool handles your real work. Waiting carries its own risk if competitors move first.
Can AI be an opportunity rather than a threat?
Yes, for companies that use it to handle routine work faster and redirect experts to judgment-heavy work. Records can also become an asset: they support internal tools, and some companies license de-identified operational records to AI developers for a defined use while keeping ownership.
How does AI exposure affect company valuation?
Buyers and lenders may ask how AI affects each revenue line. A documented exposure check, a plan for exposed services and evidence of well-kept records give management a stronger answer in diligence than general reassurance, and they show the board the risk has been examined.
Related resources
See if your company qualifies
A short company assessment. No data uploads are needed.