The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Before investing in an AI company, find out whether customers pay for a product that works in real conditions, whether the company can deliver it profitably and lawfully, and what rights and risks come with the security you would own. Ask for evidence that can be independently checked—not just a product demo, forecast or claim that the company is “AI-first.”
What customer problem does the product solve, and who pays for it?
Start with the buyer and the job the product does. A technically impressive model is not, by itself, evidence of a sustainable business. Ask the company to identify the customer, budget owner, use case and alternative the customer would use without the product.
- How many customers are paying for the product, and how many are still in pilots or trials?
- What evidence shows customers renew, expand their use or rely on the product in day-to-day work?
- How long does implementation take, and how soon does a customer receive measurable value?
- What outcome does the product improve—such as time, cost, accuracy or revenue—and how is that outcome measured?
- Can customers independently confirm the results, subject to appropriate confidentiality?
Separate signed, recurring paid use from announcements, free evaluations and one-off projects. Ask for retention, expansion and customer-concentration information in a form appropriate to the company’s stage; a young company may have less history, but should be clear about what it can substantiate.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsWhat does the AI system do, and how well does it work in its intended setting?
Ask which product functions actually depend on AI and which rely on conventional software, human labor or third-party services. Then examine performance where customers use it—not only on a curated demo or a benchmark that may not resemble their workflow.
#1 Best Overall
- What evaluations measure quality, and who designed or verified them?
- How representative are the test data and tasks of actual customer use?
- What are the common failure modes, and how frequently do they occur in relevant settings?
- Which outputs require human review, and what happens when the system is uncertain or wrong?
- How does the company detect performance changes after launch, including changes caused by new models, data or customer behavior?
Ask for examples of limitations and incidents as well as successful outputs. Inaccurate or misleading results, model defects and inadequate oversight can create business, legal and reputational risks, as issuer risk disclosures have noted. A polished demonstration cannot answer whether the system remains reliable in production.
Does the company have the rights and controls it needs for its data?
Map the data used to train, tune and operate the product, as well as information customers provide. For each important data source, ask where it comes from, what rights permit its collection and use, and whether those rights cover training, retention and the company’s intended commercial uses.
- Can the company explain the relevant licenses, permissions or other legal basis for its data?
- What customer data is retained, used to improve models or shared with vendors, and what can customers control?
- How are customer information and access separated and protected?
- Which outside model, dataset, cloud or compute providers are critical, and what happens if their prices, terms or availability change?
- What processes address data quality, bias, privacy complaints, security incidents and deletion requests?
Look for evidence of governance and incident response, not just a privacy or security statement. Data-rights disputes, poor-quality or biased data, privacy and contract violations, cyberattacks and dependence on providers can affect a company’s ability to operate or serve customers.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Can the business make money as customer usage grows?
Revenue growth matters only in context. Examine financial statements that are audited or otherwise reliable when available, and ask how much it costs to deliver the product as customers use it more. In an AI business, compute and inference costs can be important inputs to gross margin; include them rather than relying only on revenue growth or a headline margin.
- What are gross margins after model, inference, hosting and other delivery costs?
- Do unit economics improve, stay stable or worsen as usage expands, and what evidence supports that pattern?
- How concentrated is revenue among the largest customers, and how exposed is the company if one leaves or reduces use?
- What are cash burn, available funding and expected financing needs?
- What assumptions underlie forecasts, and which are supported by actual customer behavior rather than unverified pipeline or market estimates?
- How much capital is allocated to research, infrastructure, compliance and customer support?
Ask how management expects adoption and profitability to develop, but distinguish an objective from demonstrated results. Investor.gov advises private-placement investors to review financial statements, assess whether claims and expectations are reasonable, and understand how proceeds will be used. Its guidance is general investor education, not a substitute for reviewing a particular offering.
What is defensible, and what could make the product replaceable?
Compare the company with direct competitors, established software providers, customers’ in-house alternatives, open models and the vendors on which its product depends. A claim of technical uniqueness matters less than a durable reason customers would continue paying.
- What would a customer lose by switching, and how costly or disruptive is the change?
- Does the company have distinctive, lawfully usable data or durable access to a customer workflow?
- Could a customer build the feature internally or obtain it as part of an incumbent’s product?
- Could a model or infrastructure provider offer a similar capability, change its terms or raise its prices?
- How does the product compare on the customer’s actual criteria, such as reliability, integration, cost and support?
Ask management to identify competitors and explain the basis for its differentiation. Investor.gov also recommends examining competitors when evaluating an issuer. A prominent investor, a particular model choice or an “AI-first” label is not a substitute for checking the product, customer evidence and dependencies.
What legal, security and governance obligations apply to this use case?
Legal exposure depends on what the product does, who uses it and where it operates. Identify the jurisdictions and customer industries involved, then ask what obligations apply to the company and its customers. A general checklist cannot establish compliance for a particular product; qualified counsel should assess the relevant facts and jurisdictions.
- Who reviews consequential or high-impact outputs, and how are errors escalated?
- Who is responsible if an output causes harm, and how do customer contracts allocate that responsibility?
- What intellectual-property, privacy, security and recordkeeping issues have been assessed?
- What incident-response process, insurance and contractual protections are in place?
- How does the company monitor changes in legal requirements and customer obligations?
Consider whether the company can meet the obligations associated with its actual use cases, not merely whether management describes its product as compliant. Litigation, regulatory scrutiny, security problems, rights disputes and reputational damage can all affect a company, though the relevant risks vary by issuer.
What exactly are you buying, and what could affect your ownership or exit?
First identify the security and the documents that govern it. Public shares and private offerings expose an investor to different disclosure, liquidity and contractual conditions. Review the current documents for the specific issuer and transaction rather than assuming that a company’s general description explains your rights.
If you are considering publicly traded shares
- Review the latest company filings, including its description of the business, risk factors, financial condition and share structure.
- Ask how cash needs or future fundraising could affect dilution and the company’s ability to execute.
- Consider the price you would pay in relation to your own assumptions about the company; this article does not establish a fair value.
If you are considering a private offering
- Read the offering documents and determine whether the security is equity, debt, a convertible instrument or another structure.
- Understand the valuation basis, fees, use of proceeds, voting and information rights, liquidation or conversion terms, and potential dilution.
- Check transfer restrictions and how an investor might realize value; do not assume a timely sale or public listing will occur.
- Ask which offering exemption is being used and verify claims about regulatory status in the actual documents.
Investor.gov’s Private Placements under Regulation D – Updated Investor Bulletin warns that private placements can be highly illiquid, may provide less disclosure and can result in the loss of the entire investment. It also explains that a Form D is a notice filing, not SEC approval or registration. Decide whether you can bear the possibility of holding the investment for an indefinite period or losing the full amount.
Who is making the claims, and what incentives do they have?
Check the backgrounds of executives and key technical leaders, seek references where appropriate, and corroborate material claims using primary documents or independent customer evidence. Understand how promoters, placement agents or other representatives are compensated and whether their interests differ from yours.
Best Value
- Can management answer specific questions about customers, product performance, finances and risks with consistent evidence?
- Are claims about customers, partnerships, forecasts or regulatory status documented and verifiable?
- Are compensation, fees and conflicts clearly explained?
- Do you understand the security, its risks and how you might exit?
Investor.gov cautions private-placement investors about high-pressure pitches and other warning signs. Its bulletin says, “If an issuer fails to adequately answer your questions, consider this a warning against making the investment.” Treat that as investor-education guidance, not a binding legal rule. Urgency, guaranteed-return promises, implausible claims, missing information or evasive answers are reasons to stop and verify before proceeding.
How should you compare two AI companies?
Use the same evidence standard for each company, adapting expectations to stage, business model, security and jurisdiction. A comparison is only as useful as the underlying documents and answers; private companies may disclose less and be harder to value than listed companies.
- Paid adoption, renewals, expansion, customer concentration and independently supported outcomes.
- Performance in real workflows, failure modes, human oversight and time to customer value.
- Data and intellectual-property rights, privacy and security controls, and reliance on outside models or compute.
- Gross margin after AI delivery costs, cash burn, capital needs and financing risk.
- Differentiation, switching costs, competitive alternatives and resilience to technology changes.
- Legal and governance exposure based on product use and operating geography.
- Valuation, security rights, dilution, fees, liquidity and plausible exit paths.
If information is unavailable, record that as an unknown rather than filling the gap with an assumption. No single label, model, investor or forecast establishes that one AI company is safer or a better investment than another.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
