The promise of artificial intelligence (AI) has sent imaginations racing, even in the often technology-averse area of sales. In practice, salespeople have been leery of technology, from the telephone to CRM—but eventually, the pressure to succeed leads them to adopt technology or get left behind.

Artificial Intelligence AI is a little different. Because it represents a leap forward in not just capability but also in versatility, it’s something of a cipher—everyone projects their ideas of what it can do upon it. That can lead to dangerous thinking that positions AI as the secret to all that ails your sales organization. Treating an emerging technology as a panacea is a setup for disappointment, a swan dive into what Gartner calls the “Trough of Disillusionment.”

However, if its users have their thinking straight and understand what AI works best for, the rise of AI may be closer than anyone imagines. Will it make a big impact this year? Sort of—although its initial impact may be subtle.

We asked three of the most respected independent analysts what they thought 2017 held in store for AI. Here’s what they had to say.

Denis Pombriant, Beagle Research

It all relates to where you think we are in the long term economic cycle. Early in a cycle, it’s easy to sell because the market is a green field. No one has anything like what you have on offer, but they want something, so you take orders more than sell. Later on, as markets consolidate, you find it harder and harder to find new customers and any deal takes a long time because the people you sell to are neophytes.

For the most part we’re now in the later part of the cycle and it’s hard to sell. This makes it very important to be responsive to any opportunity. So, if you plan to be in a new or emerging market, your challenge will be handling the volume, otherwise your challenge will be finding good leads and pursuing them aggressively. I’d weight my efforts in the next year toward strategies that help find and close deals. The customer demographic will be relatively conservative and risk-averse people. So, AI and especially machine learning that can help identify and stack rank opportunities will be very valuable. Moreover, using these technologies to identify ripe opportunities in very large pools of territory data would be optimal because sales reps need to manage by exception. That means conservative buyers take a long time to decide, but you need to be tracking them closely because once they decide they expect sudden assistance from a vendor.

Esteban Kolsky, ThinkJar!

AI is overhyped and ill-understood (at best). I don’t think it will do anything for sales, or customer service, or anyone for at least 2-3 years if not longer. I cannot in good conscience contribute something to expanding the hype of a 50-year old technology that still is not understood by virtually all the people working with it in enterprise technology today. If you ask 10 people what AI is, you will get a) 10 puzzled looks, b) 12 different definitions, and c) virtually zero implementation or case studies (we are getting closer in use cases, but there’s a handful – at best – of lab-based implementations for them).

If you want to talk about better analytics, which is where we are and what we call AI these days, sure – let’s. Predictive is getting better, prescriptive is nowhere near a level of adoption or implementation that may generate a market, and the rest is a mess of offers and solutions with no structure. AI? Not there yet. Not even close.

But natural language processing, fuzzy logic, decision systems, automatic classification, machine learning, cognitive sciences (and those are the best known “pieces” of the real AI) don’t get used by virtually any vendor today – except for a few small, tiny startups that are starting to work in the next generation of AI.

Jim Dickie, CSO Insights

We’ve done a lot of interviews, and what we’re seeing is not people saying, “I need AI,” but rather “I need to bring more science to the art of sales.” They’re seeing a lot of no-decisions and longer sales cycles. We (sales) need to get our act together.

AI might help with that. Our research found that, when it comes to the sales process, 14 percent of companies have random processes, 48 percent have informal processes, and 31 percent have a formal process. But only 14 percent have a formal process that everyone is expected to follow, and which is integrated into technology. AI, by virtue of what it does, requires a formal process to work. So the emergence of AI is likely to drive more companies up the process hierarchy, which should make their sales processes more predictable.

However, AI is not one thing today but 50 things. We see a lot of vendors with products that offer pieces of AI for sales, but not “AI for sales.” We aren’t there yet. However, we always overestimate the impact of technology over the first two years and underestimate its impact over 10 years, and I have no doubt that, once integrated with other sales technology, AI will be life-changing for salespeople. Right now, salespeople spend 80 percent of their time trying to get the info they need and 20 percent of their time trying to act on it. When AI is mature, they’ll spend 5 percent of their time getting information and 95 percent of their time doing stuff with it.