AI Is Getting Better at Explaining Your Business—But Can You Trust It?
For a Malaysian SME owner, asking an AI tool to analyse sales, customer records or operations can feel like having an instant business analyst. You upload information, type a question and receive a polished explanation, chart or recommendation. The danger is that a confident answer can still be incomplete, based on the wrong data or impossible to verify.
This is why the story of QueryStory matters. The startup is building an AI analytics platform designed to connect answers to the underlying data, show the queries used and provide a confidence indicator. Its central idea is simple: businesses should not only receive an AI-generated answer; they should be able to understand how that answer was produced before acting on it. TechCrunch reports that QueryStory focuses on making AI-generated analysis more transparent and reviewable.
What Happened
QueryStory was co-founded by Shapor Naghibzadeh, a former Google SysOps engineer who worked on analysing complex cyberattacks. His experience taught him that reliable decisions depend on verified knowledge gathered from many different systems. He later co-founded Chronicle, a cybersecurity company that was developed within Google’s X Labs and focused on helping organisations query complex security data. These details were reported by TechCrunch in its profile of QueryStory.
The company emerged from stealth in August 2026 with a product aimed at enterprises holding large, proprietary databases. Its platform lets users ask questions, generate analyses and assemble the results into a narrative. QueryStory raised a US$6 million seed round in late 2025 from Brightmind Partners and New York Life Ventures, according to TechCrunch. The funding and investor information are sourced from the TechCrunch article.
In a demonstration using space-activity data, the platform created visualisations and analysis in hours, compared with several weeks for a previous developer-led project. It also displayed the SQL queries behind the analysis and included a confidence indicator explaining why the system considered its conclusions reliable. Users could flag work for human review, with the review recorded in the platform. TechCrunch describes this demonstration and QueryStory’s review features.
Why This Matters for Malaysian SMEs
You may not operate a large enterprise database, but your business still depends on information spread across multiple places. A café might keep orders in a point-of-sale system, supplier details in WhatsApp, stock counts in a spreadsheet and customer feedback on social media. A construction subcontractor may have project updates in email, attendance records in a cloud folder and payment status in accounting software. When these sources do not match, an AI tool can produce a neat answer that hides a messy foundation.
Consider a simple question: “Which products should we promote next month?” An AI assistant could combine sales figures with stock information and suggest several items. But if returns are recorded separately, online orders are missing, or some products have unusually high sales because of a one-off event, the recommendation may be misleading. You need to know which records were used, which period was examined and whether a person checked the result.
The same issue applies to staffing. You might ask AI to identify the busiest operating hours and recommend a roster. If the system ignores public holidays, delivery peaks or part-time staff availability, the suggested schedule could create service problems. In Malaysia, where businesses often manage different outlets, languages, payment channels and seasonal demand patterns, context is particularly important.
“The thing that we are selling is the trust in the answers,” QueryStory CEO Shapor Naghibzadeh told TechCrunch. For an SME, that trust should come from traceable data and human checking—not from confident wording alone.
What to demand from an AI business tool
| Capability | Why it matters to you | Practical question to ask |
|---|---|---|
| Source tracing | Shows where figures and conclusions came from | Which sales records and dates were used? |
| Query visibility | Allows you or a reviewer to inspect the calculation | What formula or database query produced this result? |
| Human review | Prevents important decisions from being fully automated | Can a manager approve or reject the analysis? |
| Confidence explanation | Highlights incomplete or uncertain information | What could make this recommendation inaccurate? |
| Audit history | Creates a record of decisions and changes | Can we see who reviewed this report and when? |
How You Can Apply the Lesson Now
Start by choosing one repeatable business question rather than asking AI to run your entire company. Examples include identifying slow-moving stock, comparing outlet performance, summarising customer complaints or checking overdue invoices. Define the data sources before connecting them. This may reveal that your records need cleaning first, which is often more valuable than generating another report.
Next, create a review habit. Ask the system to show its source records, calculation steps and assumptions. Have one person compare the result with the original spreadsheet or software report. If the decision affects staff schedules, customer eligibility, supplier selection or compliance, require approval from a named manager.
You should also separate low-risk and high-risk uses. AI can help draft a weekly sales summary or group common support questions. It should receive closer scrutiny when recommending credit terms, changing employee shifts, rejecting a customer request or making a decision involving personal data.
The Bigger Picture
QueryStory reflects a wider change in business technology. General-purpose AI chat tools make it easy for employees to ask questions, but different people can receive different answers based on their prompts, access permissions or selected files. Over time, this can create multiple versions of the truth inside one organisation. TechCrunch highlighted this risk as part of its reporting on QueryStory.
For SME owners, the answer is not to avoid AI. The better approach is to treat AI output as an analysis that needs evidence. A useful system should preserve links between the answer and the records behind it, make assumptions visible and allow a person to challenge the result. This also makes staff training easier because your team can learn how a conclusion was reached instead of memorising instructions for a mysterious tool.
Data protection should remain part of the decision. Before uploading customer, employee or supplier information, check where it is stored, who can access it and whether the provider uses it to train other systems. Keep permissions narrow and avoid placing unnecessary personal information into an AI prompt. For Malaysian businesses, this is a sensible way to support responsible handling of customer and employee records.
The most valuable AI assistant for your business will not necessarily be the one with the most impressive answer. It will be the one that helps you verify the answer, identify uncertainty and take the next action with confidence. Build that discipline now, and AI can become a practical layer over your existing business systems rather than another source of confusion.
Source: TechCrunch, “QueryStory wants you to believe what AI is telling you,” published August 26, 2026.
Ready to Streamline Your Operations?
Technology moves fast. Your operations should keep up. AutoRunBiz builds AI systems that run your daily workflows — from WhatsApp order capture to accounting. Book a free 15-min ops audit →