There is no one moment when SaaS – Software as a Service – was conceived, because SaaS as a concept has a host of components; all of which have had to come together in the right context in order to produce value for any sector or vertical market. Different sectors have moved towards SaaS models at different speeds.

In technical terms, SaaS relies on cloud delivery at scale, a minimum degree of widely available connectivity, and enterprise-grade security. If any of these are weak, SaaS drops off the agenda.

Then there’s the business model and the attendant marketing around it – not only do SaaS businesses themselves have to sustain the transformation from a CapEx to an OpEx model, they also have to explain to their customers what’s going on, and how they can benefit from it. After all, the idea of renting rather than owning is challenging for some; and there are all sorts of (sometimes unresolved) questions about data security, ownership and power of redress. 

It’s therefore taken a decade or so for SaaS tools to become widely adopted, and in that time a set of characteristics of SaaS services has become clear:

• Upfront pricing on a subscription basis;
• Real-time, incremental service updates rather than monthly or even annual release;
• Exceptional ease of use and self-serve customisation for every user (particularly on mobile platforms);
• Device and use-case agnostic: learn once, use anywhere;
• Strength in numbers: easy integration with other SaaS tools;
• Information as the fundamental commodity of value – data in, more valuable data out.

It’s been worth reciting that summary of what makes a SaaS service, because Artificial Intelligence is indeed going to revolutionise these businesses, and it will do so by improving precisely those parameters.

Automation 

Let’s start with automation. AI allows more functions which may previously have had a manual component to be automated. This manifests itself in various ways – typical examples include using machine learning to automate aspects of customer service (especially self-serve), but we can also imagine using AI to improve, say, the onboarding process of a SaaS product, or  refamiliarisation after a major update, through automation. 

SaaS is spectacularly successful  – consumers and businesses alike appreciate the simplicity and 24/7 availability of online services and their predictable, balanced business model.

Given that the main user experience challenge of a SaaS product is remoteness (that’s the price users pay for the benefits of 24/7 access on their terms), AI can mitigate that remoteness and ensure that more customers get a more satisfying experience.

Continuing with automation, at a time when the doom-mongers are convinced that AI will replace jobs practically one-for-one, many analysts feel that AI will deliver most value when it is deployed in conjunction with human effort. This may go against everything that you think SaaS stands for, but perhaps the best way to view – and profit from – a market is to effectively triage interactions into those which can be handled automatically (the SaaS way) and those which require human intervention. 

Take, for example, the Lola app; created by Paul English, the co-founder of the Kayak travel website. Lola will deliver a combination of live travel agents and AI-augmented chat functionality. The purpose is to enhance the capabilities of travel consultants. With the help of AI, agents are productive to the tune of nearly twice as many journeys and make consistently better recommendations. There are still countless SaaS-only travel services, but Lola allows enough automation for human interventions to be viable, useful and deliver better experiences to the consumer. As a commercial proposition, AI-augmented human interactions drive SaaS interactions too.

Personalisation 

AI will also progressively support the personalisation component of SaaS services, too. Previously, user interfaces have become exponentially more complex as each iteration of a piece of software crams in more functions and menus onto the screen. 

But app developers, Clearbridge, say that the future of interactions is context-aware computing, driven by AI. With natural language processing and voice control; along with dynamic controls designed automatically from previous interactions to present just the configuration a customer expects when they need it, using SaaS tools will become child’s play. This reduces adoption hurdles and allows more services to become a part of each user’s daily software suite.

Shipping code even faster 

On the developer side, customer and competitive pressures both mean that the development cycle for SaaS products has been crunched down from several months to (in some cases) minutes. With tools like Docker, new code can be deployed in seconds – with confidence that it will scale perfectly to thousands of users. 

This near-real-time deployment means great competitive advantage, but bad code can be hugely and disproportionately costly. What if a SaaS service with 100,000 users crashes just for a day? Even before any legal remedy is considered, in terms of trashed reputation, the cost can be vast. AI is coming to the rescue, interpreting multiple code languages, assessing them for errors, learning with each new script analysed and improving the viability of each patch and update. 

Microsoft is not just using AI to test code, it is working on systems which will deploy AI to develop code itself – soon we will train AI to write code in the same way as we might train a dog to sit. 

Enhanced security 

Finally, of course, there is the security aspect of SaaS. These services are online by definition, and whilst scaled-up SaaS services undoubtedly deliver more cost-effective and holistic security than a localised on-premise solution ever can, the problem is that every user is their own weak link in the chain. We are all fallible, and – because life isn’t necessarily fair – we expect our SaaS services to take responsibility for our own mistakes (have you changed that password recently? Didn’t think so…). 

AI’s expertise in pattern matching, human factors assessment and machine learning mean that these techniques are rapidly becoming the most effective security patrol a SaaS provider can deploy: because AI is capable of learning to replicate the actions of a hacker or a delinquent user, rather than just a dumb perimeter device. Oracle, for example, has just added machine learning to its cloud-specific security regime.

SaaS is spectacularly successful  – consumers and businesses alike appreciate the simplicity and 24/7 availability of online services and their predictable, balanced business model. But those benefits also turn those businesses into commodities: customers can always choose to go elsewhere. AI, with its ability to integrate into any digital process, is set to be the essential driver of service, loyalty, stickiness and cutting edge functionality for the next decade of SaaS services.

Source: https://unbabel.com/blog/ai-changing-saas-landscape/