1、Outlook on Artificial Intelligence in the Enterprise 2018Introduction 3AI Application Today 4AI Potential 6 Requirements for Success 8 Conclusion 11 CONTENTS Outlook on Artificial Intelligence in the Enterprise 2018 Presented by Narrative Science in partnership with the National Business Research In
2、stituteAs AI moves deeper into mainstream enterprise, its important to realize this is just the beginning, not the end-point. Sectors that have pioneered the use of AI, like financial services and healthcare, will be joined in leveraging the power of artificial intelligence by sectors from product m
3、anufacturing to retail. Within those industries, AI will go from being a specialty innovation, to being involved in mainstream operating functions. At each step, AIs progression must be driven by reality, not hype: it should be deployed at the right time and deliver quantifiable business benefits an
4、d returns on investment. To reach that potential, AI must not be feared, but rather used as a humanizing force: putting a story, via intelligent language interfaces, and a face, via image recognition, in front of technology in the workplace. To better understand the evolving impact of AI in the ente
5、rprise, we recently surveyed leading business and technology executives from a variety of industries. The results, presented below, paint a picture of AI having a significant and imminent impact on everything from company strategy, to business operations, to job functions. The time has come for ever
6、y enterprise to come to grips with AIs impact and challenges, and begin to build a roadmap to harness its potential to transform businesses and industries. Outlook on Artificial Intelligence in the Enterprise 2018 3 Introduction Artificial intelligence has often felt like tomorrows story- flush with
7、 promise, but always just out of reach. But the truth backed by real-world deployment data and examples is that AI is here today, and quickly growing in its applications and impact on business and society. AI ADOPTION GREW OVER 60% IN THE LAST YEAR .In 2017, that number rose to 61%. In 2016, 38% of
8、respondents said they had implemented AI.Because artificial intelligence has been on the radar for so long often through headlines trumpeting its most sensational aspects its easy to miss that 2017 may very well represent a tipping point for AI as a mainstream business tool. Our survey data tells th
9、e tale: AI is here today, and its impact and application is growing 61% TECH INNOVATION IS DRIVING BUSINESS STRATEGY said their company has an innovation strategy to drive investments in new technologies like AI. 71% INVESTMENT IS ON THE RISE A majority of companies said they had budget dedicated to
10、 enabling innovation. 59% 62% .with most citing their budget had increased over the past year. AI IS HERE TODAY of respondents say they are implementing AI. That near-term focus, driven by allocated budget and in- production deployments, is all about taking advantage of AI in the present, not the fu
11、ture, to enhance productivity, create new products, personalize interactions, and automate rote tasks. AI has emerged as a practical tool with real-world applications, led by uses that should be familiar to anyone that has followed the technology, such as predictive analytics and machine learning (s
12、ee chart). Outlook on Artificial Intelligence in the Enterprise 2018 4 THE MOST WIDELY USED AI-POWERED SOLUTIONS PREDICTIVE ANALYTICS MACHINE LEARNING NATURAL LANGUAGE PROCESSING OR GENERATION VOICE RECOGNITION AND RESPONSE VIRTUAL PERSONAL ASSISTANTS / CHATBOTS DIAGNOSIS / RECOMMENDATION ENGINES OT
13、HER 24.51% 21.9% 14.38% 14.38% 11.44% 11.11% 2.29%Just as notable, however, is the growing use of AI to help humanize technology. Natural language processing/generation and voice recognition/response enable more natural speech-based communications, and personal assistants and chatbots put an even mo
14、re personal touch on people-machine interactions. BUSINESS INTELLIGENCE 90% of respondents working in business intelligence said they would be interested in incorporating AI to make their data analytics tools smarter. FINANCE 87% of respondents working in finance departments said they would be inter
15、ested in using AI to automatically generate insights and reporting in human-sounding language. COMPLIANCE/RISK 55% of corporate compliance risk officers said they use AI today or would be interested in using an AI-enabled technology to help create regulatory reporting. PRODUCT MANAGEMENT 68% of resp
16、ondents working in product management said they would be interested in using AI technologies to generate performance information (profit and loss, revenue, usage, churn, etc) about the products they manage. MARKETING/SALES 77% of respondents working in marketing and sales said they would be interest
17、ed in applying AI to automatically generate reporting in natural language. COMMUNICATIONS 43% of respondents said they send AI-powered communications to internal employees today; 35% said they send AI-powered interactions to customers. Outlook on Artificial Intelligence in the Enterprise 2018 5 Acco
18、rding to our survey, these AI capabilities are having an impact on a growing number of business departments and functions:These are some of the most crucial tasks any business can undertake. Getting day-to-day work done. Pleasing customers. Anticipating and responding to business challenges, both ta
19、ctically, in the near-term, and strategically, in the long run. Taking full advantage of artificial intelligence in these scenarios can be a tricky balance. Concerns about the potential negative impacts of AI from privacy to de-personalization to job replacement demand that AI be deployed transparen
20、tly and in a clearly-defined manner. Workers and customers both want to understand where AI is being deployed and how AI is being used. At the same time, for AI to be most effective, the AI must also be seamless, bordering AI is ultimately about people helping people better interact with technology,
21、 automate mundane tasks, and generate analysis. So its crucial to understand what individuals think about AI: What are their expectations? What role do they see it playing in their everyday work? What roles and tasks do they believe it can address? How do they prefer to interact with it? To reach fu
22、ll potential, AI must be a transparent and ultimately humanizing force, bridging the people-machine worker-technology gap Survey respondents anticipate an array of high-level benefits from artificial intelligence, including: 1. Identifying business opportunities 2. Automating repetitive tasks 3. Imp
23、roving workforce productivity 4. Competing with peers Outlook on Artificial Intelligence in the Enterprise 2018 6 Survey respondents anticipate an array of high-level benefits from artificial intelligence, including: THE TOP 4 BENEFITS ENTERPRISES SEE FROM USING AI SOLUTIONSPERSONALIZED COMMUNICATIO
24、N: Nearly a quarter of survey respondents said they believe that AI-based natural language generation will help them produce more personalized and relevant information at scale for customers and prospects. on invisible. Interactions with AI must be natural and human improving rather than replacing p
25、ersonal interactions and day-to-day work tasks. Improving interactions calls for an AI that is not just a computational or processing engine but a vehicle for more human-like people-machine interactions. We see that today with consumers interacting with their Amazon Alexa personal assistants, posing
26、 natural language questions and receiving correct and personalized responses. The results from a query posted to Alexa or other emerging digital personal assistants are not only effective they can border on the magical. Technologies such as natural language generation (NLG) are especially important
27、in the enterprise, helping workers automate and improve rote tasks and enhanced crucial business processes. Advanced NLG utilizes artificial intelligence to better understand what people want to communicate, highlight what is most important and impactful, and then deliver the results in natural lang
28、uage. Its full power is realized not as a standalone solution, but as one tool integrated into other AI-driven analytics applications, producing narratives capable of explaining insights that raw data and even data visualizations cannot do alone. DATA ANALYSISof survey respondents said they are inte
29、rested in using AI-powered technology that provides natural language insights into data. In particular, they anticipate AI-based natural language generation to yield more consistent accurate data reporting; save time and free employees to focus on more high value work. 95% From our survey results, r
30、espondents believe AI and applications like natural language generation can help them better perform many important everyday work tasks, including: IMPROVED ACCURACYof survey respondents said they spend significant time reviewing their teams work and correcting inaccuracies, critical processes that
31、could be improved and automated using AI. 50% PRODUCT IMPROVEMENTof survey respondents said they have considered integrating AI-powered technologies to help differentiate their products. 33% 24% Outlook on Artificial Intelligence in the Enterprise 2018 7Outlook on Artificial Intelligence in the Ente
32、rprise 2018 8 Realistic adoption timelines and quantifiable ROI is critical for enterprise AI success Weve seen the power AI has to transform business processes and help employees get their work done. But the deployment of artificial intelligence is clearly not without challenges as well. Skepticism
33、 about the power and impact of AI can largely be traced to past missteps: not understanding what problems AI can address, overstating its potential impact; mistiming its readiness; and failing to measure the return on AI investments. It is critical for enterprises considering AI deployments to avoid
34、 such mistakes in the future. Heres how we found our survey respondents dealing or in some cases not dealing with these issues: TIMELINE FOR ADOPTION With all of the hype and uncertainty that AI brings, it can be challenging for businesses to decide when to dive in. We found this to be true in our s
35、urvey, with 42% of respondents citing they are uncertain as to when they would deploy AI. In some ways this is understandable, as many enterprises are still working to understand how AI can impact their business. We expect that uncertainty to steadily decrease as AI continues to prove its worth. The
36、 largest number of respondents said they are uncertain as to when they would deploy AI.Outlook on Artificial Intelligence in the Enterprise 2018 9 AI EFFECTIVENESS The best driver for increased AI adoption is success in early deployments, and here we see some early challenges. Asked if their organiz
37、ation was effective at using AI, respondents “strongly agreed” or “agreed” for the following use cases: Early market challenges exist for AI deployments. DECISION- MAKING EXTERNAL COMMUNICATIONS INTERNAL REPORTING 44% 35% 48% AI ROI For AI to be successful, return on investment must be tracked and u
38、ltimately realized. According to our survey, while 69% of respondents track the ROI of AI, 31% are either not tracking ROI or are not seeing returns on their current AI investments. Those numbers must improve as AI projects move from pilot phases, to production deployments expected to drive business
39、 results. The good news is that companies seem to be taking a broad view of the impact of AI, an approach that will help them more quickly quantify AI benefits. According to our survey, enterprises are tracking AI ROI via a range of metrics, including time savings (26%), operational efficiency (21%)
40、, level of accuracy (21%), employee productivity (19%) and new product revenue (14%). A broad view of AI impact will help companies quickly quantify benefits.Ultimately, enterprises must judge the timeline for deployment and the return on investment of AI based not on pie-in-the-sky projects but rea
41、l-world impact on business departments and processes. That said, it wont always be a straight line to success. For instance, as we saw earlier, risk/compliance officers have been one of the earliest groups to deploy AI to help with their regulatory reporting. Yet at the same time, 39% of respondents
42、 said they wouldnt use AI for regulatory reporting, with 21% of this group saying they are concerned that they cant audit or review AI-driven regulatory reporting. That lack of human-level comfort with an AI process worry about whats happening inside an AI black-box is something enterprises will hav
43、e to address on a department-by-department, employee-by-employee basis. Outlook on Artificial Intelligence in the Enterprise 2018 10Conclusion But go beyond those headlines, and the real impact of AI begins to come into focus. More and more organizations are deploying AI-powered technologies, with g
44、oals such as improving worker productivity and enhancing the customer experience that are not only laudable, but achievable. A focus on realistic deployment timeframes and accurately measuring the effectiveness and ROI of AI is critical to keeping the current momentum around the technology moving fo
45、rward.Based on our survey findings, enterprises would do well to frame AI adoption as a step toward humanizing technology. Critics often fear the dehumanizing effects of new technologies, replacing rich human communications with cold, sterile machine interactions. With AI, the impact can be just the
46、 opposite. Technologies like voice recognition and natural language generation can improve how we interact with technology and machines, making it both more natural and more powerful, delivering interactions and yielding insights above and beyond what has been possible up to now. That is the promise
47、 of AI and one that is increasingly within the reach of a growing number of organizations. It has been an interesting year in the development of artificial intelligence. It is difficult to read the business press without running across some mention of AI and its power to yield new insights and gener
48、ate real value-as well as the challenges of taming its impacts on business and society. More and more organizations are DEPLOYING AI technology. The promise of AI is becoming more natural and more POWERFUL. AI adoption should be framed as a step toward HUMANIZING technology. Outlook on Artificial In
49、telligence in the Enterprise 2018 11Survey Methodology narrative-science NarrativeScience narrativesci CORPORATE HEADQUARTERS 303 East Upper Wacker Drive Suite 1500 Chicago IL 60601 312.477.0590 NEW YORK OFFICE 120 East 23rd Street 5th Floor New York NY 10010 646.248.6378 SEATTLE OFFICE 111 South Jackson Street 4th Floor Seattle WA 98104 WASHINGTON D.C. OFFICE 1133 15th Street NW 12th Floor Washington D.C. 20005 National Business Research Institute deployed the survey online from August 24th to October 9th,
