The fog of AI hype will clear in 2018

AI and automation to move from theory to practice in 2018

Most of the focus of artificial intelligence (AI) and automation has so far been on the B2C sphere, with the exception of the fear-mongering surrounding the idea that robots are going to steal our jobs. In reality of course, AI and automation offer huge potential to businesses and besides, we’ve weathered similar revolutions before.

Looking ahead to next year, most of the industry professionals we’ve spoken to think 2018 will be a pivotal year in the business adoption of AI and automation; the year where the hype will give way to actual implementations. The following is a selection of key trends from leading experts in the artificial intelligence and automation industry, and offers insight into what we can expect in 2018, and which industries are most likely to be disrupted.


The fog of hype will clear

Artificial Intelligence hype will pass its peak. AI isn’t new but it’s grabbed the C-suite’s imagination and their budgets this year like never before. All too often though the projects have been box-ticking exercises in an AI arms race. 2018 will see the fog of hype clear and the C-suite demand a more demonstrable return on investment.

Heather Richards, CEO at Transversal


Excitement and fear gives way to realism

If 2017 was all about the excitement and fear of what artificial intelligence (AI) and automation might mean for us, I think that 2018 will be about realism. People will soon begin to realise that their jobs are not going to be replaced by bots overnight, and their expectations around what AI can help us with now and in the future will shift as well. 

In the business to business space, there are many repetitive processes that are currently carried out by humans or that require costly integration processes to automate. As a result, I expect to see platforms and products that facilitate low level automation becoming more normal in 2018. In practice, that could mean seamlessly entering a B2B lead from a website into a CRM system or following up on an invoice automatically based on the previous behaviour of a specific client. I think we’re more likely to see changes like these rather than a huge wave of new tools. In fact, I believe that the 'digital workforce' will gently arrive in many of the tools and systems we already use today. 

Jim Bowes, CEO and Founder of Manifesto


From theory to practice

2017 has been marked by a shift from discussions of AI in business to its real implementations. Of course, the flashy and visual applications of the technology are the ones drawing a lot of attention: self-driving cars out on the streets, AI winning in the game of Go, and so on. However, we believe that the type of AI that will be of most importance in the nearest years will be quite different from some of the most written about topics this past year. 

  • The “invisible AI” is to take the lead. We strongly believe that technology implementations that work behind the scenes to enhance and optimise already existing processes, as opposed to dramatic business model changes, are the one that will bring most economic effect in the nearest years. These types of AI have specific, measurable KPI’s, shorter implementation cycles and they already bring better results than traditional means could deliver. As an example, this “invisible AI” could help improve predicting demand, or product quality, or optimising specific process - from logistics scheduling to steel production, all while earning additional efficiencies on the go.  
  • “Offline" industries will significantly progress in AI use. Surely, we are more used to seeing AI helping us get better search engine results or e-commerce recommendations. But more traditional companies, such as those from the industrial sector, are in fact very well positioned to profit well from AI: they have enough data, plenty of routine and very complex processes, are very focused on optimisation due to high capital intensity of the industry and low margins. Moreover, their volume levels enable the potential to see major immediate effects from implementing AI: increasing throughput or cutting down costs by a mere percent for such industries results in millions of savings yearly. 
  • Practical adoption is now more important than the technology advancement. The core technologies are already here, many of which already existed for decades. It is not the development of new algorithms that will make businesses succeed, but the experience of their implementation to solve real business needs. It will require moving quickly with pilot projects, and learning from their results, rather than trying to build up the entire AI strategy from the start. Many adoption challenges are related not to technology itself, but to the adjacent business aspects: choice of correct metrics for AI to optimise, integration of black box models intro existing decision-making cycle, correct design of experiments to measure the effect from the model use and so on.  As controversial as it might seem, the companies that will focus on business and managerial aspects, will benefit the most out of the new technology.
  • Next year - next set of challenges. This practical demonstration of AI efficacy has triggered the adoption trend, that is to be advanced in the coming year. With more practical implementations, we will see more conversations arising about the next set of challenges related to AI use: complex AI-based optimisation instead of isolated projects, full automation instead of recommendations, redistribution of responsibility between man and machine, and so on.

Jane Zavalishina, CEO at Yandex Data Factory



Humans and AI are much stronger together

The companies that stand out are those that embrace AI and automation but keep the human involved. The combination of humans and AI is much stronger than each as standalone elements. There are already many players using analytics to gain additional insight into the environment. For instance, SIEM companies, vulnerability management companies, orchestration companies, and Next Gen Anti-virus companies are all using some form of AI and automation and are most likely to stand out from the competition thanks to it.

Tim Brown, VP Security Architecture at SolarWinds MSP



Machine-accelerated humans

There’s a lot of hype and confusion around artificial intelligence and its potential for many industries, including cybersecurity. Rather than unsupervised automation, the path forward in security will come from empowering humans with a combination of signatures, heuristics, behavioral analytics and machine learning.

We believe AI is just one tool for driving efficiencies and addressing staffing needs — it is not going to replace human intelligence any time soon. It's more accurate to say that AI will play a role in creating machine-accelerated humans: an army of commanders, hunters and responders who use a hierarchy of expert systems to help unearth and prioritise critical threats from the billions of records and petabytes of information gathered on the network daily. With this approach, the 10,000 to 100,000 observations most security analysts are used to experiencing can be narrowed down to a highly consolidated handful of the most critical and relevant threats with which they can reason.

Gene Stevens, co-founder and CTO of ProtectWise


Regulation will need to become part of the agenda

The question of how powerful AI will be regulated is perhaps one of the most pressing of our modern age. These machines will “think” through complicated algorithms and perform thousands or millions of actions per second. In short, no human ever stands a chance of keeping up, and oftentimes even in retrospect cannot understand what happened. And that is exactly the reality that widespread AI will eventually bring to any part of life that is touched by technology — and that is most of it.

Our concept and relative perspective of time must change, and quickly. The financial markets are a good place to examine to see how we deal with these timeframes, as they are already familiar with autonomous algorithms and understanding time and causality.  Time precision is already a fundamental aspect of governance, accountability, and auditability in finance, and microsecond or less vision is necessary in everything from re-creating markets following a “flash crash” event to detecting and protecting against errant algorithms or market manipulation by cyber-attackers.  We'll need a similar resolution to understand how or to intervene if powerful AI decides to go rogue or autonomous (think Skynet).

David Murray, CMO, Chief Business Development Officer at Corvil


AI becomes AIR, becomes IA

AI becomes AIR. The gratuitous use of ‘AI’ as a largely meaningless buzzword in hyperbolic sales pitches will reach fever pitch. AI will be the air that B2B breathes, seen everywhere and in everything. Perceived as table stakes and the oxygen that fuels B2B, businesses will continue to creatively reinterpret what they do, what they sell, and how they sell it through the lens of AI. Whether it’s ABM, CRM or IoT, AI will be prefixed to tired B2B TLAs to give them a new lease of life. But as we reach peak AI-hype, clients will want to look beyond the jargon and see evidence that AI can improve business performance. And clients will increasingly refer to this chart, probably the most insightful and useful business graphic on AI produced to date…

Dr. Paul Marsden (@marsattacks) is a chartered business psychologist working with SYZYGY, a WPP digital agency group


AI as small business champion

In 2018, AI will be the small business champion, as vendors spot gaps in the market to help these businesses market themselves and grow. AI is already disrupting the world of poorly managed and rarely renewed DIY websites and helping sole traders and micro SMEs to get online in seconds - with business owners across the construction, accountancy and plumbing industries in particular using a combination of Do It For Me (DIFM) service and AI to build their websites.

Ultimately, those who understand and promote AI as a way to drive efficiencies and upskill staff, will get ahead in the year to come.

Sebastian Lewis, CEO at Mettrr



Opportunities for cyber criminals and security professionals alike

Cybersecurity is an arms race and the weaker party will resort to asymmetric means to achieve its goals. Just as organisations are adopting machine learning and AI to improve their cybersecurity posture, so are the threat actors. Attackers will increasingly use machine learning to speed up the process of finding vulnerabilities in commercial products, with the end result being that attackers will use ever more new exploits without signalling that AI was involved in their creation.

AI will also increase the number of qualified cybersecurity professionals as it lowers the barriers of entry into the profession and allows less trained individuals to still be effective on the front lines of the cybersecurity battle. In addition, AI will allow existing cybersecurity professionals to move up-market by leveraging AI to find more complex attack scenarios before they do significant damage.

Oliver Tavakoli, Chief Technology Officer at Vectra



Business voice assistants finally emerge from the lab

We foresee that in 2018 the first generation of business voice assistants will be put through their paces – at least in trials – by leading suppliers and early adopter enterprises.

Scott Hoffpauir, CTO and Co-Founder of BroadSoft


The rise of the ‘meta meeting’

It’s been proven that face-to-face interactions improve relationships – both in business and personal life. In 2018 and beyond, we’ll start to see artificial intelligence play a larger role in the process of relationship building among colleagues, customers and partners. Facial recognition technology will be built into remote collaboration tools to read visual cues. Why does this matter? It will enable a meeting host to pivot their conversation if needed, and inform more effective post-meeting follow-up. For instance, a salesperson will know the likelihood that they closed a deal or an advertising executive will be informed that the idea they are presenting to clients is falling flat. This “meta meeting,” which focuses more on the feeling of the meeting (body language, tone, etc.) rather than the actual conversation, will gather insights and learnings that help the meeting host facilitate better human connections and drive positive results.

Jim Somers, VP of marketing at LogMeIn