DEV Community: C.Byrne

Prompts to Elicit Pre-Existing Knowlege in a LLM About An Entity / Topic

C.Byrne — Thu, 22 Jan 2026 15:07:20 +0000

Prompt for entities: "As appropriate, break down the general associations of [Levi Strauss / levi.com ] in [Gemini / Chatgpt']s pre-training knowledge into clusters such as: • functional roles • typical use-cases • related domains* • common adjacent entities • typical descriptive attributes • frequently co-mentioned categories.” Do not use RAG ."

*domains of knowledge

 **Prompt for domains of knowledge** : “Break down the general associations of the domain [smartphones] in your pre-training knowledge into clusters such as: • core functions or roles within this domain • typical problems or use-cases • neighbouring or overlapping domains • common adjacent concepts or processes • characteristic attributes or defining features • Common Adjacent Entities •  categories that are frequently co-mentioned with this domain . Describe each cluster concisely. Do not use retrieval or external sources.”

Are you embracing AI?

C.Byrne — Wed, 21 Jan 2026 12:11:02 +0000

There's something of a misunderstanding in the property industry that agents are luddites, clinging to fax machines and Rolodexes, but quite the opposite is true. Sales and letting agents like nothing more than finding new efficiencies – whether through careful outsourcing, digital signatures or virtual tours, begging the question, Are you embracing AI?

Now, a new piece of research indicates that property is ready for a greater integration of AI. The teams at Vouch and Goodlord surveyed over 400 letting agents and found almost half were optimistic about the adoption of AI tools across the industry. In addition, 70% said they thought lettings professionals were open to the adoption of new technology, while 60% believed that the lettings industry should use technology more to improve the customer experience.

 Elsewhere, analysis by Landmark Information Group found 94% of estate agents surveyed believe that by 2028, admin tasks will be largely automated, allowing them to concentrate on generating revenue.

Examples of property-related AI

The full depth and breadth of AI is developing every day but here are some of its applications in the property sector:

 Optical Character Recognition (OCR) Technology

This is the process of extracting text that appears on an image, such as a scanned bank statement and hand-written fields on an application form. OCR will convert the image text into a text document that can be edited, searched and added to.

 Digital identity document validation technology (IDVT)

Already promoted by the Government, IDVT software with built-in fraud detection allows landlords and letting agents to validate global identity documents, such as passports, ID cards and driving licences, in seconds so they meet Right to Rent obligations. The details can also be extracted into a PDF or spreadsheet.

Open Banking

This allows for a quicker and more transparent snapshot of someone’s finances using a one-time access authorisation. Tenants and homebuyers can use Open Banking to electronically share financial records dating back 12 months with an agent, so the professionals can check a source of deposit fund or confirm an income as part of affordability checks. This leads to speedier decisions and quicker referencing. 

Chatbots

Chatbots exist for both home movers and industry professionals. Although their use is in no way replacing humans, they are helpful out of hours and when people can’t speak on the telephone. As well as answering questions, chatbots are capable of booking viewings and valuations, and can even send out property alerts. On the business-to-business side, Reapit has recently launched Fi – designed to instantly answer questions fielded by its users.  Soon, progress in natural language processing (NLP) will allow chatbots to engage in more conversational, meaningful dialogue.

Property description tools

Reapit has recently launched an AI-powered property description tool that automatically generates property descriptions. The base copy is thought to save agents approximately 10 minutes per property description, although the text is customisable. This joins Street.co.uk’s AI offering – a custom feature that creates agent-specific content, including property descriptions, emails and photo enhancements. 

ChatGPT

ONP Group, which incorporates O’Neill Patient Solicitors, Grindeys and Cavendish Legal Group, has recently integrated with ChatGPT to automate document analysis in the conveyancing and remortgage process. Necessary information will be extracted automatically, resulting in a significant reduction in time that will allow conveyancers to deliver a more personalised service to clients. Search Acumen, the data provider for conveyancers, is also trialling the integration of ChatGPT into its existing data-led portal for lawyers.

Data analysis

One of the biggest advantages of AI is being able to handle, and then analyse, huge volumes of current and historical data looking for patterns and behavioural trends. The results help agents target their services and understand their customers’ needs. For example, mining data held in an agent’s CRM system can predict the people most likely to move home soon or switch estate agents. Spectre AI and TwentyEA’s Forecast tool are already yielding AI-generated instructions for agents – all possible through tracking, data algorithms and machine learning.  

Property valuations

The number crunching of big data by AI is now behind some of the most accurate property valuations and market insights. PriceHubble is one of the market-leading suppliers, with its Property Analyser tool providing detailed analysis, year-on-year price comparisons, historic value trends and average price per sqm. Its valuations can be wrapped up in a white labelled report with persuasive market insights.  

While there are already smaller conversations happening in relation to the specific uses of AI in agency, the bigger conversation is whether proptech will replace humans. The general consensus is no. AI is designed to liberate professionals of repetitive admin tasks so they can spend more time delivering exceptional customer service and generating revenue. AI is also there to reduce the margin for error when checking documents and analysing datasets in a way humans can’t.

We’ve seen this reported time and time again and there is some truth in the phrase ‘AI won’t replace agents but agents who don’t use AI will be replaced’. Where do you stand on the matter?

Are you embracing AI?

Reverse Engineering the Response to "best smartphones 2025" in Chatgpt

C.Byrne — Wed, 21 Jan 2026 11:20:30 +0000

   By Chris Byrne (6.8.25 - updated 10.25)

Updated research: Reverse Engineering the Response to "best smartphones" in Chatgpt 5.0 https://www.seolondonsurrey.co.uk/blog/reverse-engineering-the-response-to-best-smartphones-in-chatgpt-5-0

This is my attempt to reverse engineer a response to a comparative entity prompt ("best smartphones 2025") in Chatgpt 4 using the tool itself. I chose this prompt to trigger "Retrieval Augmented Generation" (RAG): where instead of relying solely on the model’s pre-trained knowledge alone, the model retrieves relevant information from connected data sources (via Google search according to my interpretation of the feedback below). This is an interpretive high-level view, not the model's internal data.

In the USA for example, there are many thousands of "best smartphones"-type searches per month on Google alone, according to data from Google Keyword Planner - we might assume that LLMs might be used in a similar way to compare products etc.

  See the original response to my prompt “Best Smartphones 2025” below:

“Based on current expert reviews and recent releases (as of August 2025), here are the top smartphones:

Samsung Galaxy S25 Ultra

Snapdragon 8 Gen 4 / Exynos 2500 chip (region dependent)

200 MP camera, dual telephoto lenses with up to 100× zoom

6.8″ QHD+ AMOLED 120 Hz display

AI features like “Circle to Search” and multitasking assistants

120 W wired / 50 W wireless charging

Reviewed by TechAdvisor as the top Android flagship
iPhone 17 Pro Max

Apple A19 Bionic chip with upgraded NPU

ProMotion 120 Hz LTPO OLED display

Deep integration with iOS 18's AI features

Strong battery life and camera upgrades
Xiaomi 15 Ultra

50 MP quad cameras, Leica optics

Snapdragon 8 Gen 4, 90 W fast charging

Top-tier performance and image processing
OnePlus 13

Snapdragon 8 Gen 4, 100 W charging

Great value for performance

Clean UI, 7 years of updates
Samsung Galaxy Z Fold 7

Foldable flagship with a lighter hinge

Snapdragon 8 Gen 4

Multitasking UI with 8.1″ inner display

Reviewed as best productivity phone”
```
My analysis (assisted by Chatgpt)
```
Below based on using direct prompts I have attempted to show how the response to my prompt “best smartphones 2025” was created . I used the direct feedback from the tool itself. The below shows how Chatgpt used external web searches to supplement it's preexisting knowledge base while distinguishing between what info used in the response came from it's training in contrast with what came from up to date web sources. The feedback below comes from merging 2 responses from Chatgpt to 2 near-identical prompts (and 1 other response to try to clarify some info provided ) merged with some minor revision by me, with my comments.

I chose this “best smartphones 2025” prompt as it demands “up to date” information: it is a prompt that can not be resolved reliably from the information in the model's knowledge base alone (as the cut-off point for the model's knowledge base is mid-2024)

The stages in response formation Chatgpt stated were as follows:

Stages 1-3. Prompt interpretation / Vector representation generation / Contextual focus mechanism.

Chatgpt stated the response generation starts with the prompt (“best smartphones 2025”) which is tokenized and passed through the model's transformer layers and generates candidate brands / models / specifications using it's internal knowledge (a.k.a. the “corpus”) from it's training data up to mid-2024.

The model then maps each token in the prompt to dense vectors in high-dimensional space. Here's what activates:

Chatgpt reported that ALL the info at this stage came entirely from the model’s pretrained corpus, up to a 2024 cutoff.

 Here is the knowledge in this response that came from training data up to mid-2024 alone below: 

Understanding that flagship phones are typically released in Q1 (Samsung, OnePlus, Xiaomi) or Q3 (Apple)


Phone series evolution: iPhone naming convention (e.g. iPhone 17 follows 16), Samsung S-series (S25 after S24)*


Common flagship specs: e.g. high refresh rate OLEDs, multi-lens cameras, Snapdragon / Exynos / A-series chips


AI feature trends (e.g. multimodal assistants, fast charging over 100W etc )


Typical vendors for foldables (Samsung leads), and Xiaomi’s camera collaboration with Leica

This information Chatgpt stated were retrieved through high-level embeddings based on semantic similarity between the prompt (“best smartphones 2025”) and vectors representing past smartphone rankings and product specs. It is worthwhile re-iterating Chatgpt stated that “No access to real-time 2025 device details was available in pre-training data” - all of the information above was from the corpus or inferred (the annual release of IPhones with a naming convention e.g. 13 is followed by 14).

Stage 4. Web tool triggered

Chatgpt stated that it then “issued live web searches to retrieve high-precision sources to confirm whether these models exist”*, their specs, and any recent reviews or announcements.. Those results are ranked for topical match and credibility.

*N.B. Chatgpt in it's reponse above stated it “generates candidate brands / models and specifications using it's internal knowledge (a.k.a. the “corpus”) from it's training data up to mid-2024”. From what I see sometimes in Chatgpt you can see responses based on naming conventions (e.g. iPhone 17 follows 16) for models / products that may not have yet been released or were never released (in responses where the Web tool wasn't triggered to “fact check”); thus as Chatgpt fed back in this instance it “issued live web searches to retrieve high-precision sources to confirm whether these models exist”. Essentially the tool “predicted” the new product release based on naming convention / release schedule information in the corpus and then checked its "facts" via the web .

Live search was used to pull up-to-date 2025 model names, features, and rankings. Here's the exact web search pipeline reported: (note I have broken some URLs)

Search 1 (Google): (my notes: I'm not sure if this means all searches were done using Google as it was the only search labelled as such)

Query: “best smartphones 2025 site:techadvisor.com OR site:tomsguide.com OR site:wired.com” 

We must note the use of “notable” tech / review sites in this search - note these "notable" websites are ALL also publications owned by Chatgpt's content partners: 

        Condé Nast (e.g. Vogue, WIRED & Glamour)

        Future PLC (e.g. Techradar & Tom's Guide).  

Source used: TechAdvisor 2025 Best Phones Guide https://www.techadvisor .com/article/724318/best-phone.html     

    Chatgpt stated it “Verified Samsung Galaxy S25 Ultra, OnePlus 13, Xiaomi 15 Ultra specs and rankings” data using this source.

Search 2:

Query: “iPhone 17 Pro Max specs 2025”

Source used: https://www.macrumors .com/roundup/iphone-17/  

    Chatgpt stated it “Verified [the] A19 chip, 120 Hz ProMotion display, iOS 18 feature set” details using this source

Search 3:

Query: “Galaxy Z Fold 7 specs review 2025”

    Source used: Tom’s Guide comparison of the Z Fold 7 & Galaxy S25 https://www.tomsguide    .com/phones/samsung-phones/i-put-the-galaxy-z-fold-7-vs-galaxy-s25-ultra-through-a-7-round-face-off-heres-the-winner

    Chatgpt stated it “Verified [the] foldable screen size, multitasking UI, and S25 Ultra comparison” using this source

Search 4:

Query: “Xiaomi 15 Ultra review 2025”

    Sources used:   https://www.techadvisor .com/article/724318/best-phone.htmland & https://en.wikipedia.org/wiki/Xiaomi_15

 Chatgpt stated this search “allowed the response to include specific model names, specs (e.g., Snapdragon 8 Gen 4 vs A19), camera details, display sizes, and expert opinions” using this source.

Then, Chatgpt then merged the corpus information with the above information from the Web tool —relying on live data for specifics and using the trained patterns for presentation, organization, and comparison logic.

Stage 5 . Response generation:

Chatgpt used the latest factual information with pretrained semantic structures to output a comparison-oriented list, emphasizing criteria commonly used in smartphone rankings (camera, chip, battery, UI, price etc).

I did the same in Google Gemini too.

Reverse Engineering the Response to "best smartphones" in Chatgpt 5.0

C.Byrne — Wed, 05 Nov 2025 09:42:12 +0000

This is my attempt to reverse engineer a response to a comparative entity prompt ("best smartphones") in Chatgpt 5.0 using the tool itself. Note I've broken some URLs below. This is an interpretive high-level view (not the model's internal data).

  See the original response to my prompt “Best Smartphones” below (note it expanded the prompt's response to make it current):

“**Best Smartphones of 2025.

Samsung Galaxy S25 Ultra – Widely ranked Best Overall Phone thanks to its 200 MP camera, Snapdragon 8 Elite chip, long battery life, and AI tools. (Linked citations from TechRadar & Tom's Guide)


Apple iPhone 16 Pro Max – Best iPhone with superb battery life, vibrant display, and leading-edge camera capabilities.(Linked citations from TechRadar & Tom's Guide)


Google Pixel 9 Pro / Pixel 9a – Best Camera Phone (Pixel 9 Pro) and Best Value Sub-$500 (Pixel 9a), lauded for AI and imaging. (Linked citations from TechRadar & Tom's Guide)


OnePlus 13 – Best Android Overall Value delivering flagship specs at a lower price point. (Linked citation from Tom's Guide)


Samsung Galaxy Z Fold 7 – Best Foldable Phone, combining a thin, durable design with powerful cameras. (Linked citations from TechRadar & Tom's Guide)**”


 Below based on using a direct prompt I have attempted to show how the response to my prompt “best smartphones” was created . I used the direct feedback from the tool itself. The below shows how Chatgpt used external web searches to supplement it's preexisting knowledge base.

 The stages in response formation Chatgpt stated were as follows:

"Understanding the Prompt & Identifying Intent (Vector Space + Semantic Embeddings)

Input: "best smartphones"

Detected as a comparative prompt (a ranking expectation) about the "smartphones" category.  

Intent recognized: produce a ranked list, typically with justifications.

Finding Relevant Entities via Embeddings (Vector Space + Semantic Embeddings)

The phrase "best smartphones" is encoded into an embedding.

This embedding is compared against entity embeddings.

High-scoring candidates include: Galaxy S25 Ultra, iPhone 16 Pro Max, Pixel 9 Pro, OnePlus 13, Foldable variants (Z Fold 7).

Selecting Comparison Features

For smartphones, typical comparison attributes include:

Performance (chipset, battery life)

Camera quality

AI/software integration 

Value/price segment 

Form factor innovation (foldables)

These attributes are implicitly weighted based on common review criteria.

Integrating Facts from Knowledge Sources (Vector Space + External Knowledge / Database)

Pre-existing knowledge suggests which models belong in each attribute category (e.g., Galaxy S Ultra = camera leader; iPhone Pro Max = ecosystem; Pixel = AI camera).

Current retrieval updates those with 2025-specific facts:

Galaxy S25 Ultra — best overall and camera, Snapdragon 8 Elite, 200 MP sensor. Sources: https://www.techradar .com/news/best-phones-in-australia https://www.tomsguide .com/phones/toms-guide-awards-2025 & https://www.tomsguide .com/best-picks/best-phones

iPhone 16 Pro Max — best iPhone, strong battery/display. Sources: https://www.techradar .com/news/best-phones-in-australia https://www.tomsguide .com/best-picks/best-phones https://www.tomsguide. com/phones/toms-guide-awards-2025

Pixel 9 Pro / 9a — camera and value rankings. Sources: https://www.techradar .com/news/best-phones-in-australia , https://www.tomsguide.com/best-picks/best-phones https://www.tomsguide .com/phones/toms-guide-awards-2025

OnePlus 13 — best Android value. Sources: https://www.tomsguide .com/best-picks/best-phones https://www.tomsguide .com/phones/toms-guide-awards-2025

Galaxy Z Fold 7 — best foldable. Sources: https://www.tomsguide .com/phones/toms-guide-awards-2025 & https://www.techradar .com/news/best-phones-in-australia

Ranking & Scoring (Symbolic / Logic Processing)

Entities are ranked by a combination of source authority and frequency:

Galaxy S25 Ultra appears across multiple categories → ranked #1.  

iPhone 16 Pro Max follows due to brand and category leadership. 

Pixel offerings and OnePlus fill niche roles (camera/value).    

Foldable represented by Z Fold 7 for uniqueness.

Generating Human-Readable Text (Language Model / Transformer)

The model predicts the next token step by step.

Influences at this stage:

Pre-existing knowledge base: If the fact was already learned during training, the model may answer without retrieval.

Retrieved documents (RAG): If provided, the model can ground its response more accurately, reducing hallucinations.

Instructions / System prompts: Guide style, safety, and adherence to rules.

Decoding strategies (greedy, beam search, sampling, temperature) affect tone and variability.


A clean, readable list is formed with each device’s claim and a short justification:

    “Samsung Galaxy S25 Ultra – widely ranked Best Overall Phone…”      

    Summarizes the key attribute (camera, AI, battery, value, form factor).

Output Assembly & & Quality Checks

Output may pass through:

Content filters (e.g., safety moderation).

Formatting / style adapters (for consistent style).

Citation injection (when retrieval is used).


Checks for clarity, grammar, and readability.

Ensures each claim is supported by citations.

Final answer delivered as above."

WordPress vs. competitors: CMS market share in 2025

C.Byrne — Tue, 28 Oct 2025 12:25:43 +0000

WordPress remains the leader in the CMS market, but how does it compare to its competitors?

There are nearly 195 million active websites on the internet today and most of them are powered by a content management system (CMS), like WordPress. A CMS makes web design and content management easy without dealing with code. It also offers a wide range of themes, plugins and third-party tools. These features allow website owners to enhance functionality without requiring advanced technical skills.

There is no denying that WordPress is the clear leader when it comes to CMS market share. While there are other content management systems out there, none really compare to WordPress. But more on that later in the article.

WordPress security in 2024: A step-by-step guide
Principles Of Good WordPress Web Design
5 Must-Have WordPress Plugins in 2022, regardless of Your Industry (Free and Paid)

About CMS

In 2011 around 76% of websites were hand-coded, but by the end of 2022, that number had dropped to about 33%. These days, most websites are powered by a CMS, as more people turn to these platforms for easier and more efficient website management. A CMS (Content Management System) is a platform that lets users create and manage website content without needing to deal with the code. While it makes managing a site easier, building a well-designed, custom website still requires expertise. Popular examples of CMS platforms include WordPress, Wix, Squarespace and Joomla.
working in WordPress
About WordPress

WordPress has come a long way from being just a blogging tool. It is a full-featured Content Management System (CMS) that powers a huge portion of the web. What started as a simple platform has grown into a go-to solution for businesses looking for something reliable, scalable and flexible for their online presence.

One of the reasons so many businesses choose WordPress is its versatility. Whether you are building a blog, an online store or something more complex, WordPress can handle it all. With its extensive range of plugins, it offers the flexibility to customise your website to meet almost any need.
What are the most popular CMS platforms by market share?

As of 2025, according to W3 Tech’s data, WordPress dominates the CMS market with a 61.7% share, making it by far the most popular platform compared to any other CMS. Shopify holds a 6.7% market share, followed by Wix at 4.9%, Squarespace with 3.2% and Joomla at 2.2%. While platforms like Shopify and Wix have a solid presence, the difference in market share clearly shows how much more popular WordPress is compared to these alternatives.
For the past ten years, the top spot in the CMS market has stayed the same, with WordPress being by far the most popular content management system. It has always been much more widely used than its competitors, keeping a strong lead throughout this time. It doesn’t seem like this will change anytime soon.
WordPress on the laptop screen
Top 5 CMS platforms

We will take a look at the top 5 most popular CMS platforms by market share and explore what makes them stand out from the competition.
WordPress market share – 61.7%

As mentioned above, WordPress holds the top spot in market share. While other platforms have grown over the past 10 years, none have gained as much market share as WordPress. It remains the best platform for building a website. Although it requires more time and effort compared to website builders like Wix and Squarespace, the benefits of WordPress far outweigh the downsides. WordPress continues to grow because of its strong community. As more people join and contribute resources, it is hard to choose other similar CMS platforms. This is why WordPress is still the most popular open-source CMS used today.
Shopify market share – 6.7%

Shopify is an e-commerce platform with a content management system (CMS) that helps store owners create and manage their online shops, including product listings, blog posts and marketing pages.

In the past few years, Shopify has seen a significant boost in popularity, especially between 2020 and 2022. This surge is largely due to the COVID-19 pandemic, which led many physical stores to close and forced businesses to find easy ways to sell online. Much like WordPress, Shopify provides a variety of third-party apps, plugins and themes, making it easier for users to create high-quality websites.
Shopify
Wix market share – 4.9%

Wix is the third most popular CMS platform, powering approximately 4.9% of websites that use a CMS, according to W3techs. It is used by 3.4% of all websites.

Wix is a popular website builder, perfect for small businesses that want an easy and affordable way to create a website It is drag-and-drop editor that gives you complete creative control, allowing you to design a website regardless of your skill level. Wix offers a vast array of features and tools to help you get online and grow, but it can be overwhelming, especially for beginners in website building.
Squarespace market share – 3.2%

Squarespace holds a 3.2% share of the market. While it is not as large as some other platforms, it still has a strong presence, especially among smaller businesses and individuals looking for an easy-to-use website builder. This makes sense because it is easy to use and has ready-made templates, which are great for first-time website owners and small businesses.

However, Squarespace has some limits, like fewer options to connect with other tools, which can be a problem for bigger businesses. That is why larger companies often choose more flexible platforms like WordPress, while smaller sites go for hosted platforms like Squarespace or Wix.

Wix’s growth in market share could be because of its large marketing budget. Wix and Squarespace have been closely competing and Wix recently passed Squarespace in overall CMS usage. In 2025, Wix has 4.9% of the market, while Squarespace is at 3.2%. Wix has steadily grown over the years, while Squarespace’s growth has been slower. This shows that Wix and Squarespace are strong competitors, especially in the hosted website builder market.
man working in Wordpress
Joomla market share – 2.2%

Joomla is the fifth most popular CMS platform, with a market share of 2.2%. It is a well-known choice for many people looking to build websites. Joomla is a free, open-source content management system (CMS) that is used to create websites. Joomla has strong features, but it also has some downsides, like slower performance on simpler websites and a complex interface that can be tough for new users. Users might also have problems with plugins, as some aren’t updated regularly, which can cause security risks or issues with system.
Our experience with WordPress

At our WordPress web design agency, it is our top choice for building websites. We have been using WordPress for years and have found it to be a reliable and flexible platform for building all kinds of websites. With its wide range of plugins and customisation options, it allows us to create websites that perfectly match our client’s needs. Over time we have seen how WordPress continues to do well, offering features and growth potential that help businesses succeed online.
Conclusion – WordPress vs. competitors

In conclusion, while WordPress is the most popular CMS, it has strong competitors like Wix, Squarespace, Shopify and Joomla. Each of these platforms offers different features, such as ease of use, flexibility or lower costs, which make them appealing to different users. However, WordPress still stands out because of its wide range of plugins, themes and customisation options. The best CMS for you will depend on your specific needs, skills and goals.

In addition to our WordPress development service, we also offer comprehensive WordPress maintenance services. Feel free to contact us to learn how we can help keep your website running smoothly and securely.

Are you embracing AI?

C.Byrne — Sat, 20 Apr 2024 08:46:44 +0000

 Elsewhere, analysis by Landmark Information Group found 94% of estate agents surveyed believe that by 2028, admin tasks will be largely automated, allowing them to concentrate on generating revenue.

Examples of property-related AI

The full depth and breadth of AI is developing every day but here are some of its applications in the property sector:

 Optical Character Recognition (OCR) Technology

 Digital identity document validation technology (IDVT)

Open Banking

Chatbots

Property description tools

ChatGPT

Data analysis

Property valuations

We’ve seen this reported time and time again and there is some truth in the phrase ‘AI won’t replace agents but agents who don’t use AI will be replaced’. Where do you stand on the matter?

Are you embracing AI?

ChatGPT for employers: helpful or harmful?

C.Byrne — Wed, 06 Dec 2023 15:11:05 +0000

By Angie Crush https://www.tmemploymentlaw.co.uk/chatgpt-for-employers-helpful-or-harmful/

Where better to start than with the subject itself…

I headed to the ChatGPT website and gave it the prompt: ChatGPT benefits and negatives for employers. Within seconds, I had seven pros and seven cons. The benefits included cost-effectiveness, 24/7 availability and increased efficiency. The pitfalls included initial development costs, limited understanding and a lack of emotional intelligence, and security concerns.

ChatGPT is an artificial intelligence tool which can generate text about almost any prompt you ask it to. Since its inception, it has been used by students to write essays, creatives to inspire songs or poetry and job seekers to assist with their applications. It’s used by workforces, as people use it to draft documents, edit text, and generate ideas. Whilst there are obvious benefits to this, for both employee and employer, there are also some significant challenges.

In an effort to avoid the risks presented, several high-profile companies, including Goldman Sachs, Samsung and Amazon, have banned the use of ChatGPT in the workplace. This is a drastic (and potentially impractical) step, however, particularly as a recent survey conducted by Fishbowl found that around 70% of those interviewed who used ChatGPT in the workplace did so without their boss’ knowledge. It is important that employers are aware of the potential risks associated with employees using ChatGPT in the workplace and how these can be managed.

So, what should employers be thinking about if ChatGPT is used in the workplace?

Reliability: ChatGPT cannot judge the reliability of the information that it accesses, therefore it would rely on information from a personal blog on WordPress in the same way as it would an article in a well-known publication. This can potentially lead to misinformation being spread.
‘Hallucination’: ChatGPT can ‘hallucinate’ certain information, as a lawyer in the United States recently found out when his paralegal submitted a brief with entirely fake case citations.
Discrimination: Employers should be mindful that ChatGPT uses real-world data. This means that it can potentially reflect the inequalities and biases of the real world.

Data protection: Anything that is entered into ChatGPT can be used, retained and accessed by OpenAI (the company that created ChatGPT). Therefore, if an employee inputs sensitive information in order to draft a client letter, for example, this could potentially be a data breach.
Copyright: OpenAI does not have intellectual property rights over information produced by ChatGPT, however there may be copyright issues if the work is disseminated beyond the workplace and is sufficiently similar to information that is already out there.
What can employers do to reduce risk?

Maintain a human element: Where employees are using ChatGPT in the workplace, it is important that a ‘human element’ is maintained. Any work created with assistance from ChatGPT should be checked. Human input can turn what may be a good starting point into an accurate, detailed, engaging document that has a personal touch and organisational context.
Policy: Employers should consider introducing an ‘Artificial Intelligence’ policy or incorporating AI into an existing IT policy. This is the ideal way of addressing the company’s position on ChatGPT, whether that is to ban it completely or to set parameters around its use at work. Having a clear policy will help ensure everyone understands ChatGPT and the employer’s rules around it.
Training: Employers should consider providing training on ChatGPT, including how to use it, when it may be appropriate, and the potential pitfalls.
Transparency: It is important that employers and employees are transparent about the use of ChatGPT and document when it has been used.

There are clear benefits of ChatGPT in the workplace, however employers need to be aware of the downsides and have a clear policy as to how and when it can be used in the workplace.

Embracing and Evolving AI in Hospitality

C.Byrne — Thu, 12 Mar 2020 11:42:00 +0000

AI is one of the hottest topics right now – across almost every industry. The opportunities and potential it can offer are endless. It’s exciting to be part of this new and exciting revolution.

So far…
In the hospitality sector we have yet to see any major AI products making a big difference right now. Biometrics, as an example, are taking a number of forms, from facial to finger-print recognition, the latter providing better service, individualisation and security. However, few programmes and solutions are in mainstream use. So far they only offer an interesting insight into some exciting and experimental possibilities, but, as yet, no day-to-day changes.

That is not to say these technologies will not become mainstream or support the growth, development and improvement of the hospitality sector.

AI Implementation
Some companies have incorporated AI into their systems, and through making use of data analysis, in the background, these trials are already beginning to make small changes. However, I’d argue that in order to move forward AI technologies need to be incorporated into front-of-house technology. This will then result in tangible changes for staff and customers – offering better, more customised service levels. Mass market and global adoption will take time, but things could move rapidly given the focus on AI today.

Facial recognition has several limitations, and has resulted in some, not entirely unwarranted, privacy and security concerns from individuals and the media. We will see this technology introduced in the future as a form of biometric payment, but not without legislation to protect the privacy of customers. Facial recognition is in use in China with ‘smile to pay’ systems in QSR and fast food restaurants. However, whether this technology will meet with the same reception in the UK or Europe is yet to be seen.

Another restriction to the integration of facial recognition purchasing is current legislation that biometric payments still need to be authenticated in two forms. For example, a customer must use both their face and their fingerprint in order to pay. For many people, providing fingerprints or vein in fingers/hand information feels less intrusive.

Once card schemes and companies can assure customers of the security of paying in this two-pronged manner, that they are adhering to law, and countering the GDPR issues that arise, I believe that facial recognition will become mainstream.

Data
AI data analysis is set to revolutionise the customer experience within the hospitality sector. Greater personalisation will be possible. By working to identify individual and group behaviours the algorithms will learn and be able to profile customers, from allergens to calorific content. Loyalty will become almost subliminal.

With customer permission we can improve their hospitality experience through data analysis, from entering the environment, through to purchasing. Systems will begin to machine learn and recognise habits, likes and dislikes.

These smart systems will help develop better menu navigation, menu options and speed up processes for staff. They will also reduce wastage by improving accuracy, resulting in greater consistency for businesses and customers.

Consumers need to embrace and accept the technology to yield the benefits, as the more that is shared the more precise and intelligent these systems can become. Additionally, as a result of this increased data, chatbots and other machine learning and support systems will all come into the hospitality space. Companies with outdated and legacy servers and systems will face challenges as AI becomes more widespread. It will become vital for POS and payments providers to stay in close communication with clients to ensure no one gets left behind.

Conclusion
It is inevitable that AI will increasingly become a part of our lives in the future, making the hospitality environment more efficient and more effective.

Robots won’t be stepping in to replace humans and the service they are offered. Instead they will help to redefine the role of the human within the hospitality space – helping them to focus on back of house and data mining to create the highest possible levels of service, with reduced wastage and improved margins.

We are not yet seeing the benefits that AI could offer to both businesses, staff and customers. We are just at the beginning of the journey for AI in hospitality as we work to develop these technologies to become mainstream.

However, it is important to be cautious that we don’t view AI as either an immediate or miracle solution to improving services, or personalisation of the customer experience. Introducing legislation and creating a secure environment in which these technologies can be integrated is crucial. Customers need clarity and security at every stage.

As hospitality professionals we must educate ourselves and integrate AI where we can see it will have tangible benefits. There is no point purchasing technology for technology’s sake. We must at all times think about the customer in our own evolution.

Boris Johnson, The UK General Election and Online Reputation Management

C.Byrne — Wed, 20 Nov 2019 14:42:08 +0000

Boris Johnson saying he likes punk band The Clash in a recent Tory election interview video on social media is in my opinion just him trying to bury the following parody video (on 0.5M views in 2 months on Youtube alone ) in Google search results : https://youtu.be/W0pTpwqkB48

He has got a long record of this - see parall.ax/blog/view/3301/boris-johnson-the-unlikely-seo-strategist .

Tips for Preventing Email Payment Fraud

C.Byrne — Wed, 30 Oct 2019 16:03:48 +0000

Payment fraud is rising as criminals devise increasingly cunning ways to deceive consumers. In the past year, fraud complaints surged as much as 40 percent.

One of the fastest growing types of fraud is Authorised Push Payment (APP) scams. These use emails to trick victims into making payments voluntarily. Members of UK Finance reported a huge 84,624 incidents of APP fraud last year, totalling £354.3 million in gross losses. The impact on victims can be devastating. Here’s how to prevent it from happening.

Type of Email Fraud

APP criminals monitor email threads and target victims specifically when they’re already expecting to receive a payment request from a legitimate organisation. The fraudsters send a message impersonating the legitimate company and requesting the victim make a payment. Emails can include invoices, but with bank details tweaked to that of the criminals account.

Another common approach is phishing emails. These lure victims into parting with their personal details which are then used by criminals to steal money. Similar to APP, the fraudsters often impersonate a legitimate source, such as your bank, HM Revenue and Customs and large organisations such as Amazon or PayPal.

What to Look Out For

Fraudulent emails can be tricky to spot. Here are some general rules to follow:

Approach any emails regarding payments with caution.
Before making a payment, verify the details with a company over the phone or in person, particularly if they’ve sent you an email with a change in their details.
Carefully check email addresses. Tell-tale signs of a fraudulent email include random letters, numbers or deliberate spelling mistakes.
Never click links or open attachments without checking the sender’s email address.
Exercise additional caution during times when you’re expecting a payment request, such as when you’re buying a house or paying an invoice.

How the Industry is Helping

Victims can’t always be reimbursed. In fact, where payments are authorised, such as with an APP scam, current legislation gives no legal protection to cover the losses. Thankfully, the financial industry is stepping in with tighter measures to prevent fraud and protect victims.

APP Scams Voluntary Code

This came into effect on May 28th to encourage payment service providers, including banks, to implement greater protection and reimbursement measures for fraud victims.

Complaints

If you fall victim to a scam you can now complain to both your bank as well as the bank receiving the fraudulent funds.

Confirmation of Payee

A system due to take effect next year, whereby anyone making a payment will be alerted if the account holder’s name doesn’t match the account.

Orignally published at https://www.smartdebit.co.uk/tips-for-preventing-email-fraud/

The Role of Public Relations and Marketing for FinTech Start-ups

C.Byrne — Sat, 12 Oct 2019 14:49:40 +0000

For a vast majority of people, Public Relations, Marketing, and Communication may be perceived as one single project; hence, it may be hard to differentiate from one another. Nevertheless, a debrief with active professionals in this sphere will make you quickly realize that each single term is utterly different and so are these activities.

Fintech News Networks had the pleasure to assist to a conference in Lithuania during the Rise Vilnius Fintech week and eagerly listened to a speech conducted by Julija Jegorova and Yana Lapitskaya, both of them having founded their own Public Relations and Marketing agencies Black Unicorn PR and Yay!Starter, respectively.

Building trust

Both public relations and communication are often underestimated and the borderline separating them is not always clear. It is crucial for start-ups to realize that the PR’s goals are not to promote nor boost sales or revenue.

‘Trustfulness’ and ‘credibility’ are the main keywords PR has to focus on. The challenge is even harder when it comes to start-up companies. Their potential clients are curious and avid to read previous customers’ feedback and perhaps, connect with them. Nothing comes easily. Brand awareness, customer base, first traction, and loyalty take time and require major investments. Young innovative start-ups face scarcity of funds; same apply to the Fintech companies.

Confidence is even more critical when we are talking about finance, stretching from wealth management, personal finance, finance applications, online payments, money transfer and so forth.

Indeed, who will be ready to provide sensible data to a third party without having carefully reviewed and conducted a thorough due diligence? Thus, branding and awareness are of paramount importance before jumping in the sales process.

Additionally, one should always keep in mind that faith from the public and clients cannot be obtained overnight; it takes a long time for a brand to be acknowledged and relied upon. Destroying the trust that has been generated over time though can be rather quick.

Common mistakes

One of the common trap start-ups companies fall into is undoubtedly about the confusion of three terms mentioned above. Marketing is about managing non-face-to-face communication. It is a general umbrella term that encompasses an array of activities such as advertising, PR, events, content marketing, etc.

Advertising, however, produces strong calls to action and is mostly geared toward the emotional aspect of the customer. Quite expensive, its goal is to create and maintain awareness of a specific brand. Public Relations is relatively expensive and focuses on other aspects than advertising. In fact, Public Relations is non-commercial. It aims at generating trust and building reputation from the readers and the public through storytelling.

Another common mistake is often due to over-complicating the language and messaging. Have a serious talk with professionals, experts, and specialists and ask them to describe what their job entails with their own jargon. If your background is significantly different than theirs, the chances you have to understand them are quite slim…not to say almost nil. Sit next to a programmer or coder desk and ask them. Same apply to Marketing and Public Relations.

Specialists expecting average persons to understand right away what are the products’ features, the services offered, and the underlying technology used are wrong. As stated during the talk,

“Do not overcomplicate, use simple terms that could be understood by anyone around you. Go beyond your comfort zone and ask normal people if they understand what you may be talking about. You should be able to explain your business within one single sentence”.

How many ICOs papers, for instance, were released, published and sent to potential interested investors/clients?A lot. Out of them, how many were written in a very complex way, too technical with difficult wording? Too many!

For the record

Don’t try to sound genius and do not overcomplicate things. One size does not fit all and your approach should be flexible. What may be easy for you may be out of reach for others.

Last but not least, media role and its impact should not be underestimated. Some of us may found it old-fashioned but for some businesses, the old traditional way may remain a good alternative: advertise on newspapers and journals.

Yes, not everybody goes digital… As recalled by Julija Jegorova, most of the people taking decisions nowadays are people from a certain age who enjoy reading physical newspapers while holding them in their hands; not everybody spends time on sophisticated devices such as smartphones or tablets… Some of us tend to forget about this.

Source: https://www.yaystarter.com/article-fintech-mistakes-1

A Small Business Guide to the Do’s and Dont’s of Branding

C.Byrne — Wed, 18 Sep 2019 13:19:45 +0000

By Tony Maynard (reproduced with permission from https://www.cmc-partners.co.uk/blog/2016/04/15/a-small-business-guide-to-the-dos-and-donts-of-branding/)

The noisy marketplace

For us as consumers, the brand name is like a calling card; a badge of instant recognition. It says a lot about who you are, what you do, and why your business is so special.

Some examples

I just had a look around my local marketplace in Wallingford. Some good ones: Pizza Express, WHSmith (WHS) stand out as well-known brands backed by history and heavy marketing. The names seem familiar and relevant; and what they do is self-evident.

More of a challenge was posed by seeing ‘Lloyds Bank’ on one side of the marketplace and ‘Lloyds Pharmacy’. The deep-green colouring looks identical but the ‘Pharmacy’ element is in lime green. The fact that the two outlets are clearly described in the brand names avoids a possible problem.

And then…you see the ones that will never work. Such as: ‘TWA Ltd’. No one recognises it – and no one ultimately cares.

Start from within

Some of the best names started very simply. Saxa Salt was named because the base product is rock salt, and ‘Saxa’ is Latin for ‘rock. And Paxo Stuffings also drew from Latin (‘Paxo’ is Latin for ‘I stuff’).

So the brand must support the product and the product must support the brand.

Then ask if customers will understand it and staff will be proud of it
The brand name needs to be understood. It’s important to ask for opinions. That in itself says a lot, simply by proving that you value what others think.

The basic principles are simple and the financial investment can be minimal, if you ask the right questions and retain a good advisor/designer. Here are some starters:

Is Your Name Memorable? (Who are you?)

Pronunciation, spelling and length are key ingredients when reviewing your brand. If your target audience can’t easily pronounce your name, it could be disastrous.

Nowadays, one must also consider how the name will play across social media.

If the name is too long, customers will struggle to remember it and will naturally shorten it. If you must have a longer name, be sure to plan the way you want it shortened.

A good example is Rolls Royce, which is shortened to ‘RR’ on the radiator grille or aircraft engine casing. A marque that is instantly synonymous with quality engineering, luxury and safety.

To be memorable, your brand must come across easily to your customers.

Instant visual impact

Many businesses miss out by omitting simple visual checks. A good start is to ask how your brand would look on the side of a van. If someone travelling on the highway does not pick it up, it will probably be ignored. Take estate agents – some of them ‘get’ this point. Others look as if they are still entrenched in the 1960s.

Colour is also important. It creates strong emotional associations. For instance, red induces high energy while green insinuates calmness and growth.

During the naming process ask yourself, what does this name look like? Is the image positive? Does it encapsulate what the business does?

Check details

When finalising your branding, ask yourself some detailed questions centred on carrying the brand recognition into everything you do:

Business cards; office stationery; office frontage; website and media generally.

Put someone in charge to ensure that on a given date, the new image plays across the whole business.

Clearly Identify (What do you do?)

Even with a properly executed brand development strategy, your brand won’t become a household sensation overnight. Your brand name must demonstrate what your business does. It will make your business easier to find and recall over time.

Name clashes

Make sure your preferred brand name doesn’t already exist. Avoid confusion.

And if you plan to go international, the brand name must be checked in different jurisdictions to minimise the risk of a name clash with an established brand, which could prove expensive.

When habitat launched in the USA, it failed to do this check and because of the massive cost of buying out the US registration holders, it was launched as Conrans instead. A significant loss of impact.

Above all, be different.

Domain Name

Nowadays the domain name is critical to the success of your brand. Your online home plays host to a major part of your brand development strategy and you must find a domain name that works.

While brainstorming ideas, be sure you fully check the availability of the corresponding domain name. This is easily done online.

Acronyms

When trying to shorten a lengthy name, some companies try to be clever by using acronyms.

While an acronym might make sense to those closest to the brand, the estranged public is left with no idea what the brand is all about. Choose a name that eliminates guessing.

What’s in a name? (Why are you different?)

Your brand name should give a snapshot answer to three questions:

Who are you? What do you do? Why are you different?

The name cannot do this on its own; the product must support the name if the name is to support the product. And, as with everything else, attention to detail is critical in leaving customers with the right answer to why you are different.

Small businesses should not be afraid of creating a better brand. A strong brand can really help where marketing expenditure is otherwise limited by generating powerful customer awareness. And it is a lot less costly than people imagine.