What Do Technology Journalists Need Before Covering a Startup or Quoting a Tech Expert?

By Jordan Baker, Founder of Fair Public Relations and former UK national journalist

A technology journalist needs more than a confident founder, a clever acronym and a link to a product. Before covering a startup or quoting a specialist, a reporter usually needs a clear problem worth explaining, a credible person who can explain it, verifiable facts, an accurate account of the product's development stage, appropriate demonstration access and a realistic chance to challenge the claims. For AI, cybersecurity and data-sensitive products, responsible limitations and privacy safeguards are part of that material, not awkward details to conceal.

From a former journalist's perspective, the test is straightforward: could I independently understand and check the information, put it in front of my readers and answer sensible editorial questions about it? A pitch that calls a prototype a revolution while withholding the test methodology fails that test, however polished the press release. A smaller company with an honest, usable explanation can be a better source.

What does a technology journalist need before quoting an expert?

An expert quote is valuable when it helps a reader understand a development, not when it repeats a company's sales positioning. A journalist wants to know what happened, why it matters, how the source knows, and where their knowledge stops. For a software outage, the useful expert might be someone who can explain system dependencies without pretending to know the affected company's private architecture. For workplace AI, the useful source may have directly relevant technical or organisational experience rather than simply owning an AI-related business.

A reporter may ask what work the person has actually done, whether they have a financial interest in the technology being discussed, how they reached their conclusion and whether a caveat belongs in the quotation. Prepare a brief biography with current role, relevant experience and genuine areas of expertise. Give them a direct, responsive contact route and disclose conflicts when relevant.

Useful evidence is not necessarily complicated. A clearly attributed observation based on an engineer's own projects can be helpful if it is presented as experience rather than a national statistic. A claim about a measured effect, by contrast, needs a defined population, method, period and result. The important distinction is between professional judgement and an empirically supported statement.

What makes a product update genuinely reportable?

An update becomes potentially newsworthy when it changes something readers care about and a reporter can demonstrate or substantiate that change. A product may reach a meaningful level of availability, solve a previously difficult problem, address a documented customer need or introduce a genuinely distinguishable capability. A version number, refreshed interface or newly adopted 'AI-powered' description is not automatically independent news.

Take a hypothetical accessibility software company that adds support for a previously unsupported workflow. Its pitch should identify the exact new function, who benefits, where it works, the release date and any limitations. Ideally, a reporter could test the function or speak to a permissioned, relevant user. That is more useful than 'the most transformative platform in its category'.

Founders should separate three stages that are routinely blurred: an internal prototype, a limited pilot and a generally available shipped product. A pilot may still support an interesting story, but a journalist must know the participant pool and what remains unfinished. Do not use a future roadmap as proof of present performance.

How should a technology founder explain a complicated product?

Start with the problem, the user and the actual workflow. If the product is a cloud-based service, explain what an organisation does differently because it exists. Then describe how the technology contributes, what a user must provide, what the output looks like and where responsibility remains with a person. Technical mechanisms matter when they explain a benefit or limitation; they are not a substitute for a useful answer.

For example, 'we use intelligent orchestration to optimise enterprise processes' tells a journalist almost nothing. 'Our tool helps a logistics manager identify delivery exceptions in a dashboard before the daily dispatch meeting' is more concrete. The next questions are sensible: which data feeds are supported, what constitutes an exception, what happens when data is missing and how is the result verified?

Prepare explanations at two levels. The plain-language version helps a general business reporter and readers unfamiliar with the category. A technical fact sheet can explain architecture at a safe level, integration requirements, benchmarks and limitations for a specialist title. Both must describe the same actual product rather than two incompatible marketing stories.

What makes a technical spokesperson credible?

Job titles, qualifications and follower counts are not interchangeable with relevant experience. A credible source can show work connected to the issue, explain its scope, distinguish known facts from inference and answer challenging questions without becoming defensive. Their background should be accurate and easily verified through appropriate employer, professional or publication records.

An AI product founder may know how their own system was deployed but not have expertise in every branch of machine learning, law or medical ethics. A cybersecurity engineer may be qualified to describe a general defensive principle without being entitled to declare an unrelated provider compromised. Respecting those boundaries can make a source more useful, not less.

Prepare examples of subjects the person can discuss and subjects they should decline. Include any restrictions around client confidentiality, employer approval or commercially sensitive work. An expert who can answer in ordinary language and follow up promptly is often more valuable to a newsroom than someone with a grand biography who can only provide pre-approved marketing sentences.

What evidence supports 'AI-powered', accuracy and time-saving claims?

Different claims require different proof. 'Uses a third-party language model for draft suggestions' is a factual product description that should be checked against the implementation. 'Produces accurate legal advice' is a much stronger assertion, potentially affecting user decisions, and requires substantially more evidence and caution. 'Reduces work by 60 per cent' raises immediate questions about the measured task, sample, baseline, variation and period.

Document what was tested, with whom, how the comparison was constructed, which failure modes were excluded and whether results come from production users or a selected demonstration. Include the tool version and date. A figure measured under laboratory conditions should not be presented as a universal outcome for customers. If an independent assessment is claimed, identify the assessor and actual scope.

The US Federal Trade Commission's final order relating to Workado's AI content-detection accuracy illustrates the importance of substantiating a specific performance claim. The decision concerned claims about one vendor's detector, not a conclusion that all AI systems are inaccurate. The UK's Information Commissioner's Office also distinguishes statistical accuracy from the legal duty to keep personal data accurate; that difference matters when AI systems generate inferences about people.

NIST's AI Risk Management Framework and its generative AI profile are voluntary risk-management resources, not government certificates proving a tool works. Describe what a company has actually evaluated rather than borrowing the framework's name as a substitute for evidence.

What metrics do technology reporters actually need?

Start with definitions. 'Users' may mean anyone who ever registered, monthly active users, weekly active users, trial accounts, paying customers or seats provisioned by a company. They are not the same measure. A claim about adoption needs the relevant denominator, observation period, geography, method of collection and any exclusions.

Likewise, a benchmark may measure average speed on a particular task rather than production reliability. An energy-saving statistic may compare modelled usage with observed usage. A customer survey may include only people who replied. None of these distinctions invalidates the result, but concealing them can mislead an editor and reader.

Make a one-page metrics register for the claim you hope to pitch. For each figure, record what is counted, the start and end dates, who collected it, whether the company may publish it and who can answer methodological questions. If the underlying result cannot be reproduced or explained, present it as an internal observation with careful wording or leave it out.

What do product demos, press access and embargoes need?

A journalist reviewing software may need temporary working access with appropriately permissioned test data, setup instructions, supported devices, an explanation of paid features and a named technical contact. If the product is available only to selected customers, say so. If an environment is staged, mark it clearly as a demonstration, not a live customer account.

Avoid furnishing fictional user records that appear to be genuine customer data. Do not provide administrative credentials, security-sensitive infrastructure details or private conversations. Where a demo could expose personal information, build a dedicated safe environment and agree with the responsible product and security teams what can be shown.

An embargo is a request that an editor agrees not to publish material before a specified time. It does not arise automatically because a startup writes 'embargoed' on an email, and each newsroom may have its own practice. Ask before sharing sensitive pre-release information and be prepared for a journalist to decline an embargo or non-disclosure terms. Most independent reporters will not sign a broad NDA simply to evaluate a routine pitch; confidential trade secrets should not be sent speculatively.

How should security and privacy claims be checked?

Security is a set of practices and controls, not an adjective a publicist can safely add to a headline. A founder might be able to document multifactor authentication, access restrictions, vulnerability handling or a completed independent assessment. Those facts must be described at their actual scope and date. 'Unhackable', 'totally compliant' and 'bank-grade' without a precise, verifiable meaning are poor substitutes.

CISA's Secure by Design guidance puts responsibility for safer software and transparent security practices on manufacturers. It does not certify that a named startup is secure. A journalist can responsibly report a company's documented security approach, but a quotation or article is not a penetration test or independent audit.

For products using personal data, determine what information is collected, how customer permission is handled, what can lawfully be demonstrated and what limitations need to be disclosed. The ICO's AI and data protection guidance is relevant for UK processing, but its guidance is under review following legislative changes, so businesses should check the current applicable requirements rather than copy a general compliance claim into a press release. US and Canadian rules also differ.

How do technology, finance and health newsrooms differ?

A technology reporter may focus on the architecture, user problem, competition and evidence for a product's claim. A financial publication may need to understand payments, authorisation, consumer harm and regulatory status. A health reporter may have additional concerns around clinical evidence, patient confidentiality, safety and professional responsibilities. The same 'AI platform' phrase can conceal very different risk and evidential needs.

A generic productivity app should not be described as a regulated health technology because some customers happen to be clinicians. Equally, a clinical decision-support product cannot be presented as a routine scheduling tool if its public claims imply diagnostic value. Establish the real use case, applicable jurisdiction and any independent professional or regulatory review before approaching the relevant newsroom.

A business selling technology into finance should distinguish software features from the financial product, advice or regulated service being delivered by others. Writers who understand that distinction can direct questions to qualified people rather than allowing a founder to improvise beyond their expertise.

What makes a source quote editorially useful?

A useful quotation contains a specific observation, explanation or carefully qualified prediction that fits the reporter's actual question. It should sound like a person with firsthand knowledge, not a slogan. 'Companies must embrace digital transformation to thrive' is so general that it adds little. 'A retailer moving inventory decisions to automated forecasts still needs a way for staff to challenge unexpected outputs' gives the reader a concrete issue to understand.

A quotation should not smuggle a product advertisement into an independent story. A reporter might quote a founder's expertise without mentioning the company's platform at all. That can still be appropriate expert positioning, but it should not be sold internally as a product review.

Offer comments promptly, identify the speaker accurately and state any relevant uncertainty. If figures are involved, supply the source alongside the comment. Avoid putting words in the mouth of an employee who has not approved them. The goal is information a newsroom can use without having to unpick promotional claims.

Which genuine campaigns show usable expert positioning?

Fair PR's work with Trinity SES offers an example of technology-related business expertise rather than consumer gadget publicity. The three-month expert-led engagement involved Francisco Gaffney commenting on leadership, technology and organisational questions. Seven distinct linked historical articles appear across July to October 2025; the broader date range should not be presented as seven publications falling exactly inside those contracted three months.

Spectrum Search provides another example from technology recruitment, where a specialist source can explain workforce and market developments. The documented 18-month campaign reports 86 published media opportunities. That represents the case study's published-opportunity measure, not 86 full-length app reviews or evidence that software customers signed up.

Allister Frost case study covers AI and workplace change, with five recorded publications during a three-month expertise campaign. It concerns commentary on how people and businesses respond to AI, not proof that an AI model vendor received favourable product reviews. These differences illustrate why the expert's knowledge, the story type and the outlet's readership matter.

What should an expert press kit and technical fact sheet include?

The useful kit is not necessarily large. Begin with one accurate product overview describing the audience, problem, availability, pricing model and stage of development. Add a clearly attributed spokesperson biography, contact details, rights-cleared headshot and a short list of topics the person can genuinely discuss. Make the core material easy to copy and verify.

Add a technical fact sheet covering the main function, supported integrations or platforms, relevant metrics with methods, significant exclusions and the current version or launch status. Where a benchmark matters, link to the method or underlying assessment rather than presenting only the favourable headline. A customer example requires explicit permission for the details, image and quotation.

Provide appropriately licensed product images, a clear demonstration route and a quick process for follow-up questions. Keep confidential customer information, protected intellectual property and unnecessary security-sensitive implementation detail outside the press pack. It should make independent reporting easier without forcing a journalist to accept unsupported claims.

How should founders handle difficult journalist questions?

Treat questions about limitations, failures, security or competitors as part of normal editorial scrutiny. A founder can say that a figure is being checked, offer a documented clarification or decline to reveal confidential material while explaining the boundary. Inventing a confident answer to avoid appearing unprepared usually creates a larger problem.

Prepare an internal question-and-answer sheet for sensitive but foreseeable issues: what the product cannot yet do, why a result differs between customer groups, what data is retained, how errors are corrected and where human oversight sits. This helps the team find accurate facts quickly without imposing a scripted, evasive tone on interviews.

If a journalist makes a factual error, correct it with a concise source and specific explanation. Do not demand that an editor remove a legitimate critical view simply because it is commercially inconvenient. Independent reporting is not a sponsored presentation and no ethical PR process can promise complete control over the final story.

Build a technology journalist readiness matrix

Claims to check before pitching a technology journalist
Claim What a reporter needs Proof owner Disclosure boundary
Product available Release status, version and supported regions Product lead No unshipped feature as live
Faster workflow Method, baseline, sample and test dates Analytics lead Disclose limits and selected cohorts
AI performance Task definition, evaluation and failure modes Technical lead No blanket accuracy or safety claim
Security Scoped controls or verified assessment Security lead No confidential logs or architecture

A useful internal matrix separates the claim from the supporting material and the disclosure risk. For example, 'available to all customers' should be checked against the release record and geography; 'saves time' against an appropriately designed test; 'secure' against precisely scoped verified controls; and 'uses AI' against the actual implementation. Each claim needs an owner who can confirm the latest truth.

Mark the readiness of each item as verified and safe to disclose, needs further proof, or not suitable for external use. This is a preparation method, not a promise that a publication will cover every green item. The non-visible publication pack for this article includes an accessible HTML table the website team can reproduce below this section.

The process is especially valuable when sales, engineering and marketing describe a product differently. Resolve contradictions before the reporter discovers them. A published story based on the wrong launch status or an obsolete benchmark can damage credibility well beyond the initial article.

How quickly should a tech expert respond to a reporter?

The right response time depends on the journalist's deadline. Breaking technology stories may need a knowledgeable comment within hours, while a long-form feature or product test may allow days. Ask for the actual deadline, what the reporter needs and whether the source is available. Do not agree to interviews or demonstrations that cannot be delivered.

A strong response is fast and accurate. Keep the first answer short enough to be useful, with the most relevant insight near the top and supporting facts available on request. If several specialists need to validate a technical claim, tell the journalist when a checked response is realistically possible rather than sending a premature assertion.

From my experience, the easiest sources to work with are often those who understand the editor's job: they respond, distinguish evidence from opinion, permit reasonable scrutiny and do not insist that every contribution contain a product plug. The size of the startup is not what makes someone useful to a newsroom.

Frequently asked questions

Can a startup get coverage before launching a working product?

Sometimes, but the basis for coverage must be truthful. A founder with relevant experience may offer valuable expert commentary, or a real market development may be newsworthy independently of the product. A prototype can be described as a prototype, and an announced pilot as a pilot. Do not imply a working public launch, real customers or proven performance without evidence. Our guide to tech media coverage without funding announcements explores other credible news routes.

What proof should an AI company provide for an accuracy or efficiency claim?

Explain the exact task and product version, the test sample and dates, the baseline or comparator, the metric used, known limitations and whether the result has been independently checked. Distinguish a controlled test from live customer performance and disclose material exclusions. A claimed percentage with no method is not a convincing fact merely because it looks precise. An accuracy assessment also cannot be assumed to establish privacy or safety.

Will a technology journalist sign an NDA before seeing our product?

Do not assume so. Newsrooms generally need independence and may decline confidentiality obligations for an ordinary pitch. A reporter may agree to a specific embargo in advance, but it is not automatic. Share a safe demo or non-confidential information first and keep trade secrets, personal data and sensitive infrastructure details out of speculative outreach. Discuss any proposed embargo or access restriction explicitly before sharing material.

Can a technical expert guarantee a quote, review or link?

No. A source can provide useful, timely, accurately attributed material, but a journalist and editor decide whether to use it and what context it warrants. An expert comment is not automatically a software review, endorsement, hyperlink or recommendation. A credible PR process prepares the source and pursues appropriate earned opportunities without claiming to control a publisher's judgement.

Make your technology story ready for scrutiny

A strong technology pitch is built on what can be checked, not the intensity of its adjectives. Fair Public Relations helps suitable tech businesses and genuine technical experts prepare journalist-friendly stories and responsible commentary. Explore Technology PR and contact Fair PR to discuss the product, evidence, spokesperson and likely editorial opportunities. We will assess genuine media potential rather than promise a particular review or headline.

Written by Jordan Baker

Founder of Fair Public Relations and former UK national journalist. Jordan writes about journalist needs, credible expert sources and sensible earned-media strategy.

View Jordan Baker's media expert profile.

You might also like these blogs

Jordan Baker

Jordan Baker is the founder of Fair Public Relations. A former journalist with more than a decade of PR experience, he has worked with over 3,000 brands and high-profile names across the UK and US, including Amazon, Jameson and Britney Spears.

Previous
Previous

Can Technology PR Help SaaS and AI Businesses Build Trust and Win Customers?

Next
Next

How Much Does Technology PR Cost for Startups, SaaS and AI Businesses?