AWS and its rivals have created amazing AI tools for businesses – so why not make them easier to access?

Every business ought to be using AI. From the kiosk selling phone cases on the Euston Road to Procter & Gamble, there’s guaranteed to be some application of AI that can help a company speed up its workflows, improve its offerings and/or grow its business.

To some, this may sound delusional. The term “AI” has been slapped onto so many products and services in the past few years that it’s become meaningless at best. At worst, it’s synonymous with laughable errors, misleading “hallucinations” and eerie images with six-fingered hands and weird malformed lettering. Why would you want that anywhere near your business?

But while our social media feeds are flooded with eye-catching nonsense, the big online service providers have quietly been adding a whole spread of serious, functional AI tools to their portfolios.

Amazon Q virtual assistant

For example, if you’re on AWS you can now make use of the Amazon Q virtual assistant – a chatbot similar to ChatGPT and other AI assistants, but which can mine data specifically from your own servers, and from linked external sources like Exchange, Salesforce and Slack.

Because you can control where Amazon Q gets its information from, you can be confident that its insights are based on pertinent facts and figures; you don’t have to worry about random interpolations from Bleak House showing up in the middle of your quarterly reports. The walled architecture also means there’s no chance of your private data popping up in someone else’s feed.

Amazon Q can generate code too – and you can build further smarts into its creations using a host of AWS AI libraries and functions.

AWS AI tools… continued

And there’s more. Amazon Transcribe performs AI-assisted speech-to-text conversion for audio recordings and live calls; Amazon Polly does the reverse, turning text into natural-sounding synthesised speech.

Amazon Rekognition offers real-time analysis of image files and video feeds, while a whole bevy of other tools, with stylish names such as Comprehend, Kendra, Lex and Textract, each perform their own clever tricks with tensors.

And if you still can’t find the function you want, Amazon Bedrock offers a unified API for a whole range of third-party AI engines, including Anthropic, Meta AI and Mistral. Honestly it’s hard to think of something you can’t do.

Related reading: Can you build a generative AI app with Amazon Bedrock?

Great AWS AI tools, why the slow adoption?

Yet even with all this potential on the table, the general adoption of AI remains, shall we say, tentative. Reputational issues doubtless play a part, but some of it has to come down to marketing – too much about eye-catching image generators, too little about genuinely useful business tools.

I’m willing to bet that even clued-in IT types didn’t know about half of the hosted AI features I mentioned above, and if you go to the AWS website to try to discover them, you’d be forgiven for imagining the company was trying to drive you away.

Yes, sure, there’s a web page – a slightly confusing set of pages, actually – detailing Amazon’s various AI tools, and plenty more besides. But that’s just the problem: it’s a daunting list of technologies, each with its own brand and mode of usage, not organised by what they can do for your business, but lumped together simply because they’re all built on AI.

That might make sense to computer scientists, but if you’re looking for business solutions it’s hard to know where to start.

Business first, AI second

The site also does little to address the business concerns that AI brings into the conversation. Because managers have heard the horror stories: they need to know exactly how data shared with an AI is kept private and secure, and how far they can trust AI-generated output.

They may also have questions about regulatory compliance, or ethical considerations such as social bias and the knotty question of where the AI’s enormous energy requirements are coming from.

I don’t doubt that the answers to such questions can be found in the small print when you sign up. But how many potential customers will get to that point? We need to see the key reassurances right there on the page in bold type.

Worse, if you want to talk through the whole thorny subject with an expert, the handy “connect with a specialist” button brings up a contact form with no fewer than 11 mandatory fields.

Entering into any sort of dialogue requires you to give up your business email and phone number, along with your job title and industry; there isn’t a box explicitly saying “please add me to your database of leads and pester me with the hard sell from now until eternity”, but the inference is hard to avoid. It really feels like the goal is to discourage casual enquiries.

Wider industry problem

I say all this not because I think Amazon needs my business advice. Take a look at AWS’ annual revenue, then compare it to mine, and it’s pretty clear which one of us has a better overall handle on the industry. And to be clear, I don’t mean to single out AWS: hop over to IBM and Microsoft and you’ll find they’re no better at communicating what, specifically, these new computing modalities can actually do for you.

Then again, there’s an argument to be made that setting AI services apart, as something specialised and exclusive, is itself smart marketing. While the technology has earnt its share of mainstream scorn, it still carries a whiff of magic. That would quickly evaporate if AI services were just another item on the menu of hosted services, alongside virtual servers, CRM platforms, app engines and so forth.

But the truth is, the glamour needs to go. Like it or not, AI is going to be an ubiquitous fact of everyday life, probably in years rather than decades. So we need to get it into perspective: it’s not as foolish as its reputation might suggest, but nor is it some sort of hyper-advanced technology accessible only to maths geniuses.

AI is a tool that, if deployed sensibly and practically, can benefit everyone – the businesses that can work more intelligently and efficiently, the consumers who reap the rewards, and, yes, the tech giants like AWS who’ll be servicing the ever-growing market for AI services.

It’s clear to me that AWS has the tools. Now it’s a matter of putting them into business and IT professionals’ hands.

AWS in the news

• AWS launches $230 million initiative for generative AI startups
• AWS Malaysia goes online as part of $6 billion investment
• AWS and IBM each partner with Nvidia in AI push