Artificial Intelligence Q1 2017 trends. What key lessons did we learn?


Artificial Intelligence Q1 2017 trends. What key lessons did we learn?

Artificial Intelligence (AI) Technologies Will Augment Your Enterprise Applications, Amplify Your Intelligence, And Unburden Your Employees

The recent publication by Forrester Research Artificial Intelligence Technologies, Q1 2017 (1) caught my attention last week.

We discussed the use cases of Artificial Intelligence in few our articles and blogs including our recent blog Top 5 Technology Areas that Impacting Retail in 2017 and Beyond. We believe that Artificial Intelligence achievements would play a crucial role in developing future technology solutions for retail and eCommerce.

The Forrester report evaluates the state of different solutions that are based on Artificial Intelligence and its business applications. The report examines current and prospective states of the 13 most important technologies.

The key takeaways of the publication are closely aligned with my experience in dealing with clients during earlier attempts to implement artificial intelligence technologies in eCommerce, marketing, CRM, and business intelligence.

Here are Forrester’s report key takeaways:

AI Technologies Will Augment And Enhance Human Work
AI technologies can fully replace humans in some scenarios. But most cases of AI will be where machine intelligence augment human work because they react faster and process more inputs than a human does.

I agree that this trend becomes highly visible in many consumer and business applications during last five years.

AI Systems Still Demand Considered Design, Knowledge Engineering, And Model Building
The goal of many AI systems is to have a functionally autonomous application. But an AI system requires significant human effort to design, engineer the knowledge that it represents, and build the models for taking inputs and executing actions.

Yes, we see that the teams that design and build self-driving cars, advanced robotics systems or modern virtual agents spent a lot of time on experimental design, model training, and testing to achieve desirable results.
For instance, Baidu (the Chinese equivalent of Google) has as many as 1,300 staff in its AI division and is targeting a launch of its own self-driving cars in 2020.

AI Technologies Demand New Skills, Not A New Team
AI technologies often require new skills to use such as familiarity with deep learning, text analytics, and emotional computing. However, building an entirely separate AI team is not the answer. You can build these intelligent systems with existing development and data science teams, albeit with deeper partnerships among them and novel new roles.

There are no doubts that ability to acquire new skills is the necessary condition for a success of any team working in the field of AI. As you can see from the graph below (2) the number of articles in artificial intelligence was growing almost exponentially in the last 50 years. The teams that are working in AI must keep up with all this new information and knowledge.

Articles-in-artificial-intelligence-trend graph

Will AI replace humans or help them?

When we talk about AI progress in the next 10 -20 years the emphasis is on how AI technologies can help humans rather than replacing human intelligence and creativity. These AI technologies can:

Amplify human intelligence.

Here are few use cases that are adopted by many businesses already:

  • AI powered scoring tools, combined with a training chatbot make salespeople smarter.
  • Customer support associates routinely use automated process guidance and knowledge search.
  • Software developers may benefit from analysis of best practices and pitfalls of development decisions.

Liberate employees from banal or onerous tasks.

Many responsibilities at enterprises require little cognitive effort on the part of humans, but they have been historically out of reach of machine intelligence.

The systems powered by artificial intelligence algorithms may process audio signals, images, and video and drive automated decisions based on the outcome.   These technologies combining with Natural Language Processing (NLP) may automate many routine customer services or business process tasks.

Enable robotic processes for self-healing and self-correcting systems.

AI technologies can also be extremely valuable in scenarios where there will never be a direct interaction with a human being beyond the setup and deployment.

I always think about NASA and automated space operations when we talk about the self-healing and self-correcting systems. However, such systems today have much more close to the earth applications in automated IT operations, manufacturing processes, infrastructure or business process automation. An example is HIRO™ technology developed by Arago (3).

AI technologies mimic humans’ abilities

Authors of the report arranged all 13 AI technologies in a way how they mimic humans’ abilities to sense, think, and act while constantly learning.


  • Image and video analysis
  • Biometrics
  • Speech recognition
  • Text analytics and Natural Language Processing (NLP)


  • Machine learning platforms
  • Deep learning platforms
  • Semantic technology
  • AI-optimized hardware
  • Swarm intelligence


  • Natural language generation
  • Decision management

Sense, think, and act

  • Robotic process automation
  • Virtual agents

Artificial Intelligence Ecosystem Phases

The Forrester report plots all 13 technologies positions in Ecosystem phases in one chart. I find this chart very illustrative to understand the current and predicted state of all technologies.

Artificial-intelligence ecosystem phases

Ecosystem Phases definitions(4):

Phase 1: Creation in the labs. In the Creation phase, the technology is not yet commercially available or is only available as an alpha version. Cutting-edge IT shops and emerging technology teams want to understand how long it will take to get their hands on beta versions. The technology is not ready for production implementation except in specialized cases, and its potential for business value-add is uncertain.

Phase 2: Survival in the market. During the Survival phase, the very first commercial and open source products hit the market; initial production environment deployments take place; and the ecosystem expands to include suppliers, customers, and enablers like systems integrators. Leading-edge customers and end users begin to share their experiences with the product. There are many vendors competing to earn plaudits and early customer wins.

Phase 3: Growth as adoption takes off. In the Growth phase, the ecosystem either reaches a level of diversity and resilience that sustains the technology’s existence or the ecosystem lacks momentum and the technology slowly slumps into decline and obsolescence. Widespread implementations produce piles of evidence, which allows potential customers to make better-informed decisions. The technology’s value proposition has crystallized and stabilized. Vendor consolidation begins. New entrants must acquire an existing player to enter the market.

Phase 4: Equilibrium from the installed base. During the Equilibrium phase, which can last for several years — or even decades for hardware — the ecosystem is large and resilient. The benefits and limitations of the technology are documented and well known. At the end of this phase, installed customer numbers fall as some firms switch to new substitute technologies. The market is highly consolidated, customer numbers flatten, and revenues level off or decline.

Phase 5: Decline into obsolescence. In the Decline phase, forcing factors — such as new regulations, changes in the business environment, a shrinking talent pool, or the launch of disruptive competing technologies — destabilize the ecosystem. Customers, vendors, and complementary organizations like service providers bail out. Some firms continue to run the technology, but vendors stop supporting it.The

Forester’s report also provides the profile card for each of the technologies with the following elements:

  • Definition Usage scenarios
  • Vendors
  • Estimated cost to implement
  • Ecosystem phase
  • Business value-add, adjusted for uncertainty
  • Time to reach next phase
  • Trajectory (known or prospective)

I recommend downloading the Artificial Intelligence Technologies, Q1 2017 report for all business executives who want to get a sense how artificial intelligence technologies provide benefits today or may help or create a competitive threat in nearest future.

Future of Artificial Intelligence

We imagine that the evolution of artificial intelligence technologies has broader scope and time horizon than it is outlined in Forrester’s report.

I want to conclude the blog with the quote said by Stephen Hawking in a video appearance at the 2017 Global Mobile Internet Conference Beijing:

“I believe there is no real difference between what can be achieved by a biological brain and what can be achieved by a computer …Humans, who are limited by slow biological evolution, couldn’t compete and could be superseded by AI”

You may also find interesting an infographic published  by Futurism:


  1. TechRadar™: Artificial Intelligence Technologies, Q1 2017,
  2. Global Research on Artificial Intelligence from 1990–2014: Spatially-Explicit Bibliometric Analysis,
  3. Hiro™ Overview,
  4. The Forrester Techradar™ Methodology Guide,
  5. Stephen Hawking issues familiar warning to China against AI,
Igor Nesmyanovich, Ph.D., CISSP

Igor Nesmyanovich, Data Scientist and CEO of Eradium. Igor began his career as a space scientist, and for more than two decades applied the unique art of science to the emerging digital world. Igor is a Certified Information Systems Security Professional since 2004. Today Igor’s focus is on helping clients to accelerate business growth with technology innovations at the intersection of marketing, science, and technology.