HelpHound for the medical profession - Google AI

We asked Google AI 'How does HelpHound help doctors with online reviews?' A simple enough question. Here is Gemini's answer (the full page first - you can replicate this search for yourself)...

Click on image to enlarge

And the text element only (for ease of reading)...

We rest our case.

A seismic shift in the reviews landscape - CMA to use AI to examine online reviews

This post is directed at those businesses that have so far escaped sanction by the Competition and Markets Authority (the government body responsible for enforcing the law regarding online reviews in the UK). If you are a HelpHound client, you do not need to read any further, but you may like to do so to reassure yourself that your business is in full compliance with current legislation.

First, we repeat the rationale behind compliance with the CMA's regulations...

We understand why businesses cherry-pick (not so much those that 'gate*'); it's exactly why we introduced moderation: to give businesses the confidence to allow** all their customers to write a review. 

Of course, the business above could simply replace the invitation to write a review to its own website with one to write a review directly to Google, but that would mean bypassing moderation and increasing the risk of inaccurate and potentially misleading reviews finding their way to Google.

The CMA is on the case of businesses that cherry-pick happy customers to write reviews (that was confirmed to us in no uncertain terms in a conversation with a senior CMA staffer). Until now, that meant a pretty exhaustive and in-depth investigation, and businesses, particularly SMEs, quite understandably reckoned that their chances of appearing in the CMA's crosshairs were remote.

Not any more. As of this month, the CMA will be employing its own bespoke AI to track down offending businesses. 

Now, we don't know exactly how the CMA's AI will work, but if we put ourselves in the CMA enforcement team's position for a minute, we know exactly what we would be programming its AI to look for...

• Few reviews relative to the number of customers the business has, or onboards over a given timescale 
• An erratic pattern of reviews; many one month and then few - or none - the next
• Sudden spikes in positive reviews - especially following a negative review
• Use of more than one review platform. Not always an indicator, but any sign that the business is inviting reviews to one platform (Trustpilot, say) and then only inviting those that post a 5* review there to copy it to another, more visible, site (Google is the obvious choice). Called 'gating*' by the regulators and Google
• Businesses that show only 5* reviews on their own websites
• Businesses that show unattributed reviews on their websites
• Businesses that are proactive in inviting customers to post reviews, but cannot prove that all their customers have had the opportunity to post a review**
• Businesses that use mechanisms that control the timing of the review (commonly immediately post-purchase)
• Businesses that reward customers for 5* reviews

None of the above, taken individually, are concrete proof of any contravention of the CMA rules. But they will be enough for the regulators to move on to the next step: formal investigation. This may/will involve the sequestration of computer hardware and email records. And we seriously doubt any business wants to experience that at first hand.

And the list doesn't end there. The great thing about AI is that, once programmed, it will find just about every business that fits the regulators' target criteria. All the CMA will need to do is issue sanction notices to those businesses. Exactly as HMRC did on 10 July this year, for businesses not in compliance with the Anti-Money-Laundering legislation. The onus will always be on the business to prove its innocence. And ignorance of the CMA regulations will be no defence.

And finally...

The most obvious question from businesses that have cherry-picked historically is, 'If we become compliant today, will the CMA sanction us?' We cannot give you a definitive answer to that question, although any sensible person would imagine that the CMA's priorities would lie in the direction of currently non-compliant businesses. What we can say for sure is that being able to say 'We ensured we became 100% compliant as soon as we were made aware of the relevant CMA regulations' is a far better response than 'We are currently ensuring that all our review management will be in compliance with the CMA regulations in future'.

Further reading:

• AI and reviews - an important update

HelpHound - AI and reviews - an important update

There have been two seismic shifts in the world of reviews this year...

• Google's Introduction of AI search (Gemini) 

and...

• The CMA's warning that it is introducing AI to establish which businesses are contravening its core regulations relating to reviews (see below)

Regulation 

We asked Google's AI a simple question: 'Is there anything suspicious about [XYZ]'s reviews?' This is what it returned:

This was the very first business we asked Google's Gemini AI to interrogate - it was the first law firm 'near me' in search, so no cherry-picking for our example. You can be sure that if we can do this using proprietary AI, then the CMA's bespoke system will work even more effectively - and in numbers; this, again from a Google AI search...

What does the CMA mean by 'suspicious activity'? It is simple, really: any activity that tilts the review gathering and posting process in favour of the business (and, therefore, by definition, against the interests of the consumer). At the risk of repeating ourselves, here are the most obvious examples...

1. Cherry-picking. Every business that invites reviews is either tempted down this road or actively does it. Why is it illegal? Because the CMA has said so! Seriously though. Take a minute to think about it; hand-picking 'happy' customers to invite to write a review must work in the business's favour.
2. Gating. The act of pre-qualifying the above-mentioned customers. There are many mechanisms - the obvious one being an e-questionnaire. Not only against UK and EU law but expressly against Google's terms of service (risking deletion of the entirety of a business's Google reviews and red-flagging).

This issue with both of the above, post-AI, is that the regulators have gone, overnight, from relying on whistleblowers and exhaustive, time-consuming and expensive follow-up investigations to a click of a mouse and bingo! - a comprehensive list of those flouting the law, along with comprehensive proof*.

We frequently hear businesses dismiss regulation (and at least half the businesses we meet for the first time are in breach of one or more of the CMA's core regulations). It will either be a case of 'Why would they - in this case, the CMA - bother?' or 'If they bothered, what is the likelihood of [our business] being sanctioned?'*

But now the CMA are introducing sophisticated AI, it is time to stop 'parking on the double yellow lines.' It's not as if parking legally, in the context of reviews, is going to cost businesses either - professional review management has proved itself to be a profit centre time and again (in fact, so comprehensively that we now guarantee results or your money back).

*Again, there are those who doubt us. The key here is that a) a sanction by the CMA will not be viewed as a victimless crime by consumers, the press and social media - you can probably write the headines yourself - and b) every business in breach will have a positive treasure trove of evidence for the CMA's AI to pick up - from the reviews themselves (and the posting pattern), to the emails/texts inviting them (or not inviting them, as the case may be).

4.9 is the new 4.5

We all remember the days, not long past, when we would use a business provided it scored over 4.0. No longer. To feature in a competitive search, and to attract custom wherever it may appear in search, a business now needs three things more than ever before...

• Great SEO

• A Google score of 4.8 or better

• A feed of its reviews on its own website - that, in turn, feeds through to Google searches

In the case of the latter, the feed can be of Google reviews or its own reviews. The advantages of the latter being...

• that the individual reviews themselves can be moderated, with factually inaccurate or potentially misleading (and damaging) reviews being resolved pre-publication anywhere (on the business's own website and/or on Google). We estimate, based on over a decade's experience and the statistics behind it, that our moderation adds between 0.2 and 0.5 to a business's overall Google score. Of course, it is not only the score that is important;  almost, sometimes more than the score, is the elimination or correction of a potentially* damaging review at the top of the business's Google feed (now almost certainly referenced by any AI search as well).

• that the business owns its own reviews, rather than donating the data they contain exclusively to Google

*We struck out 'potentially' as all negative reviews are damaging. The business may know that the reviewer is being unreasonable or even factually inaccurate, but the reader does not, and almost everyone searching for a high-value business will take the trouble to read any 1* reviews a business may have. Moderation is a vital safety net for any service or professional business.

SEO

You will have spotted the HelpHound client. But do you know where that '4.9 from 129 reviews' and its accompanying 5 gold stars come from? Most consumers will bet their last penny that they are from Google, but they are not. They are taken directly from the client's own reviews, moderated by HelpHound, and hosted on their website for all to see...

Why host your own reviews on your own website? There are many reasons, not least of which is the marketing boost they provide. Ammunition for advertising and PR, your website, your social media and print materials. But it's not entirely accidental that HelpHound clients tend to a) score higher - by definition, a moderated system must outscore the equivalent unmoderated system, even if it only enables the business to avoid having a single inaccurrate or misleading review a year (we have clients who recieve one a month - and they are invariably resolved before publication), and b) feature more prominently in all the vital searches - local/map etc.

Artificial Intelligence

Our article from the day of Google's launch of Gemini AI for search

Two here: first, the boon that AI provides for the basic day-to-day management of reviews, helping draft responses is a major one for our clients. Second: the warning from the CMA that it is going to be using sophisticated AI to identify businesses that flout its core regulations (see above). 

Getting great reviews

Convincing? You bet. We have purposely chosen a client in one of the most sensitive areas to show that a) people using extremely 'private' services will write reviews. If patients of a women's health clinic will write a review, albeit in less volume, so will clients of a lawyer or a financial adviser. 

More and more evidence points to three factors that override all others;

1. Warn the individual - face to face if at all possible - that it is now standard practice for your business to invite all its customers to write a review, and stress how helpful that is a) to other people in their position and b) for you personally
2. Invite the review by email, not SMS. This has been shown to produce better quality reviews, consistently. Fewer 'They were great' and more of the kind you see above.
3. Follow up with a one-minute phone call. Businesses that do so - and we have plenty of evidence of this - often convert over fifty per cent of the invitations they send out.

Results

These bear repetition, in the context of our fees and all of the above. Detailed numbers lifted from an industry forum (in this case estate agency) and collated by someone totally unconnected with HelpHound. Again, we repeat: these are guaranteed, and have been for over 18 months now. 

To summarise

• Don't wait any longer to become compliant with the CMA regulations; the CMA will not warn anyone before they act. The benefits of compliance (see 'Results' above) so far outweigh the cost, from the point of view of both time and £. And sanction by the CMA will hurt - both the fine and the attendant publicity

• Get a compliant reviews feed embedded into your website

• Aim to achieve a score of 4.9 for your business as soon as possible

• Get the SEO kicker of reviews working for your business, again, as soon as possible

Google's Gemini AI goes live

Remember the date. 31 July 2025. The day the world of search took a giant leap. 'A giant leap where?', you may well ask. The honest answer, in spite of all the noise, is 'Only time will tell'. And the main thing only time will tell is just how many web searches, and of what kind, will be conducted using AI going forwards (at this point we should apologise to the tech-sophisticated minority - we're going to focus on Google here, since it is currently 93 percent of the UK search market and that dominance currently shows no sign of waning).

What has changed?

Have you seen this yet?

Or this?

We know you are keen to see what happens next, so let's look at the search above in 'conventional' Google search...

...and now using 'AI mode'...

What you - and therefore your potential customers - will note, pretty well straight away, is that the second - the AI search - contains far more qualitative information: Winkworth heads the first search by dint of excellent web design and SEO, plus the schema pulling through its own reviews. For the AI search, Google Gemini has scraped information from over 80 different sites (if you conduct a similar search, you will see the ticker counting as it searches) and is providing the searcher with far more information and a proper response to the 'best' query. 

What should concern any business that hasn't got its review management house in order is that the AI search doesn't stop there...

...making the 'best' look even better.

Here's a specific search for a medical practitioner...

And here's another search, which anyone considering such treatment might be expected to make...

This is the result of over 120 separate searches by Google Gemini - almost certainly including the clinic's Google reviews and the reviews hosted on its own website and on HelpHound servers.

No need to highlight any of this as it is all relevant!

Conclusion

All the while Google allows people to opt in/out of Gemini, there will be some that stay with conventional 'ten blue link' search (although that is also headed by AI results), but then we remember people saying that Ask Jeeves was here to stay. There is no doubt that 'long string searches' for high-value services will gravitate towards AI; the only question is 'How long will it take?'

Last of all - but very important

Regular readers will know we never relent when it comes to compliance with the CMA regulations. But we also know that many reading this article will be content to continue to flout those regulations (which have the force of law), the key rules being those prohibiting cherry-picking (businesses must allow all of their customers, clients or patients to write a review if they invite any to do so) and gating (pre-qualifying customer opinions before inviting selected customers to write a review. 

Until now, we would probably have agreed that the risk of sanction was pretty low - even despite this conversation with the CMA - but no longer. The CMA is introducing AI, with special reference to online reviews...

...AI will identify, in seconds, businesses that flout the CMA regulations by looking for patterns of review writing and the content of the reviews themselves. To take one obvious example: a business looks great on a review site but far less good on Google. Another: the ratio of reviews to customers - the business conducts dozens of transactions but gets few reviews. 

The key is to look at the opportunity cost of compliance: a compliant business - and all HelpHound client businesses are compliant by definition of having the 'Write a review' button on their websites - can relax in the knowledge that their reviews are bullet-proof: from criticism by competitors ('they only have great reviews because they cherry-pick') and the CMA. Our clients don't pay for our review management because they have cash to spare, they do so because they know that our professional review management feeds straight through to the bottom line - guaranteed.