AI Machine Learning and Automation Will Impact Your Business in 2018

7 Ways AI, Machine Learning and Automation Will Impact Your Business in 2018

AI is taking over tasks from human workers and creating new value for businesses by doing tasks human workers could never do. This isn’t something that will happen – it’s been going on for some time. But this year it’s reaching a critical mass as basic automation measures have become widespread, while more advanced AI implementation allows early adopters to pull ahead.

Here are seven areas where AI, automation, and machine learning will change your day-to-day and your quarterly plans for the year ahead.

1. Data Visualization

Data visualization lets humans make decisions based on emergent patterns in large, complex data sets, by surfacing those patterns in a way the brain is well equipped to understand.

Most of us react better to the idea of seeing the bigger picture than to reading a massive amount of text. Infographics aren’t just popular amongst marketers because they get shared on Twitter – they make it far easier to transmit complicated or new ideas fast.

(Here’s an interactive infographic that explains why infographics are so effective if you want to get all meta in a new tab.)

Our brains process visual information faster and more holistically than they do text – which is why it’s easier to understand a picture of a pair of scissors and a dotted line than a paragraph of instructions on opening a package.

When it comes to processing complex data, visuals with annotations are king – that’s why we represent terrain with maps, using a code of lines and colors. Imagine navigating your way across town with nothing but written instructions and you’ll see what I mean.

Sisense was designed to solve this problem. It’s designed for businesses that are struggling to derive accurate insights from their data – which, at last count, is basically everyone.

sisense-homepage

Source

Providing intuitive dashboards and visualizations lets Sisense deliver data in ways that human users can just grasp and act on.

That’s going to be increasingly important as the amount of data that floods into businesses expands exponentially, while our ability to analyze it and make sense of 1,000-field spreadsheets stays exactly the same.

More forward-looking is the marriage of machine learning and AR. Salesforce is experimenting with using the Oculus Rift augmented reality platform to display the results of its machine learning component, delivering the value of big data in visual form so that people can more quickly understand it and make decisions based on it.

salesforce oculus rift video

2. KPI Tracking

Salespeople increasingly live and die – or at least, get paid or get fired – on the strength of their KPIs. Right now, many salespeople spend several hours per day on logging and adding data that are used to both assign KPIs, and to measure performance against them.

AI assistance, initially in the form of bot assistants built into or operating inside larger sales productivity ecosystems, will increasingly take over the logging and recording functions.

This is a classic example of a simple fix with available technology. We’re not talking superintelligent computers – just shaving minutes off a rep’s day until they add up to hours and dollars. The technology to do this already exists and is already being used.

Behind the scenes, a much more impressive and transformational change will be taking place.

AI bots are doing something humans can already do. They’re just doing it faster, more accurately and for nearly nothing. It’s radical in its effects, but not in and of itself.

But AI is capable of more. It can do jobs human beings can’t do at all, or that would have been available to only a handful of multibillion-dollar organizations.

Expect to see AI detecting and designing individual KPIs tailored to sales reps’ specific strengths and weaknesses.

Will this mean the end of grinding on the phones?

Nope. (We have a different plan to deal with that.)

But it will mean that KPIs are better designed to bring sales reps’ efforts and the company’s goals into focus on the same target. Which is what they’re supposed to do in the first place.

Both these capabilities are likely to first emerge as a service, then become integrated into larger sales tools before becoming standard features of all but the smallest sales productivity suites.

My favorite example of this is the movie Moneyball.

For all the goofy humor, Moneyball is really a movie about statistical analysis, though for some reason they didn’t market it that way.

The idea is that the methods everyone else was using to spot great ball players had huge blind spots. Using complex data analysis, Brad Pitt and his friends were able to identify ideal players that were statistically guaranteed to achieve the hits and runs required to win baseball seasons (and sign them cheap).

Using automated analysis of actual rep performances, measured against performance indicators, should increase both the number of home runs and the amount of unshowy solid play that actually hits goals and makes businesses grow.

Without machine learning to identify the relevant data points and automation to log the actions, this would be a full-time job for a room full of guys with slide rules.

With those advances, it can constantly happen in the background, generating personalized KPIs that reflect required improvements, strengths, weaknesses and real contributions to company-wide goals.

3. Buyer Behavior Insights

The more you know about your buyers, the better you can sell to them. That’s true in prospecting, true in closing, true in every stage of the marketing and sales process.

We’ve seen how AI will let sales organizations spot the trends in sales rep activity across sales cycles and different clients, allowing sales operations to design working practices that are more successful.

Outside the organization, the same types of tools will be used to better understand the buyer journey.

Building models of how people buy are notoriously complex and difficult. People tend to buy in unpredictable ways and for their own reasons. Real buyer behavior insights that allow sales to sell the way their buyers want to buy? That’s gold dust.

biggest problems new SDRs make

Source

Neural networks that self-teach offer a way to build an understanding of buyer behavior based on real actions and natural similarities, instead of starting with a model and trying to push buyers into it somehow.

neural network buckets

Source

At the level of sales ops and the C suite, predicting sales revenue based on analysis by neural networks has similar advantages.

Expect to see this technology offered by well-funded startups in 2018, and to see its growth mirrored in bigger companies that can afford to throw R&D dollars at it. Either way, by the end of 2018 sales operations and revenue planning software based on neural nets will be a feature of highly technology-oriented sales organizations.

4. More Productivity

The math is brutally simple: there are 24 hours in a day. And how ever much you might hustle and love the grind, you still have to eat and sleep. The most fanatically committed salesperson can only do so much in a day.

So the worst possible plan for sales departments and businesses to adopt is ‘work harder’ or ‘work more.’ Most sales people’s income is directly tied to performance, and they care about success for its own sake too. Salespeople know this: we’re already pretty close to peak grind here.

Improvements in productivity have to come from somewhere else.

Some come from better training, better organization, better allocation of the scarce resources we already have to work with.

But many, many more will come from automation.

Let’s take this point to the wider economy. (I swear this will be brief.)

Using data from the International Federation of Robotics, researchers George Graetz from Uppsala University and Guy Michaels of the London School of Economics can show what’s changed in productivity with the advent of automated factories.

Across 17 countries and 14 industries, average labor productivity rose .36 percentage points. GDP rose .37 percentage points. In other words, almost all the GDP increase came from per-worker-hour productivity increases.

And while .36 doesn’t sound like much, it’s the same increase in productivity we got from the steam engine in the years 1850 to 1910. We call that the industrial revolution. This is another one, happening right now, powered by increases in per-worker-hour productivity as a consequence of automation.

What’s that going to look like in sales?

At every level, the benefits of automation will be felt before even those of AI and machine learning.

Batched and automated internal and external communications are a characteristic first entry of automation into the sales environment. Many salespeople are already familiar with auto dialers, automated voicemail and some degree of CRM data entry and retrieval automation.

The next stage will be automated communication between levels of the organization and automated communication with leads. Seventh Sense pulls data from HubSpot, Marketo and company email systems to behaviorally profile leads and deliver the right email at the right time automatically. This kind of functionality will likely start out as an as-a-service tool before being subsumed into CRM.

Inside organizations, we can already see tools that aspire to be sales-wide or company-wide communications layers sitting on top of extant productivity systems like CRM and marketing automation tools, and letting users quickly share information from them with other departments.

Slack is moving in this direction, and its internal ecosystem of bots is automating many workaday interactions at the same time. Tel Aviv-based startup Alto uses AI as a communication layer with CRM, allowing conversational interactions between CRMs and users.

5. Transformed Social Selling

Remember when social selling was hot?

At first, salespeople flinched from the idea. To them, it sounded too much like the enthusiastic marketers they’d heard talking about ‘engagement,’ ‘interaction,’ and ‘virality.’

But then, social selling took off as salespeople realized it really did augment sales.

It’s not just that target demographics are on social – we know that they use social as a source for the content they use to make purchase decisions. So social selling gave sales people an opportunity to intercept leads closer to the beginning of the buyer’s journey.

social symbiosis infographic

Source

And like most sales techniques that actually prove their worth, before too long, ‘social selling’ became simply ‘selling.’

The trouble with social selling is that it takes a lot of time. We’re talking about a lot of touches. Even LinkedIn soaks up time and effort, and it’s such an amorphous environment that most sales professionals aren’t sure whether their work on social is paying off; their managers are even less convinced.

the state of social selling infographic

Source

The typical salesperson doesn’t know how to use social to sell, doesn’t feel confident in their ability to wing it, doesn’t work at a firm that teaches them how to do it and has a manager who doesn’t believe it adds anything to the bottom line.

So we have a disconnect. From the average sales person’s point of view, ‘social works – but not for me.’

Sales has access to increasingly powerful lead nurturing automation tools, like BuzzBuilder Pro, which automates common prospecting tasks like follow-up emails and social contacts.

buzzbuilder homepage

Source

Taking things a step further is Nudge.ai, which analyses and assesses social interactions for likely value as well as automating repeatable or scheduled actions.

nudge landing page

Source

Probably Nudge’s biggest selling point is automating introductions through your extended network, taking a lot of basic legwork out of social selling.

These tools represent the simplest and most easily implemented of the changes we’ll see this year, though. Underlying them is the promise of semantic analytics. Semantic analytics tool Cognizant says:

Source

The goal of semantic analytics in sales is to identify high-value, likely sales as early as possible based on what they say across social as well as other channels. As this becomes available, expect it to roll out as both SaaS products and in-house builds.

Salesforce is heavily invested in this field already, with Salesforce Analytics spanning ‘the entire spectrum of analytics — from basic reports and dashboards; to advanced analytics in the Einstein Analytics apps and platform; to AI-powered predictive analytics through Einstein Discovery.’

Last year Salesforce announced a partnership with collaborative data company Alation, aiming to ‘help our Einstein Analytics users work smarter and faster.’

‘With Alation’s AI-powered data catalogs,’ Salesforce promised on its blog, ‘Einstein Analytics users will be able to find the right data set and right data elements to pull into a report or analysis faster, and save time normally spent searching for the data they need.’

But other players including Cogito Analytics are either offering Salesforce APIs, standalone services or integrations with popular CRMs.

These tools will work from both directions at once, both facilitating reps’ social efforts directly by automating introductions, outreach, and follow-up, and using a broader view to identify and prioritize real leads early.

6. Sharper Inbound

We don’t talk much about inbound on this blog. But it’s still important for a lot of organizations.

Inbound sales reps are under different pressures than outbound. Response time is critical for inbound success – leads generated online require response times of 5 minutes or less and the majority of these sales go to the first responder.

Dealing with large volumes of activity on a tight timeline is a tailor-made role for automation. That’s why online lead generation ‘first response’ is increasingly automated. We all know this isn’t really Dan McGaw, hanging out in his website’s backend in case I happen by:

effin amazing landing page

Source

I’m pretty sure that’s not Charles either:

wine dot come home page

Source

They’re automated chatbots that will work to gather information about me while they answer my questions, both nurturing and qualifying me as a lead. They’re providing all-important rapid lead response.

So we’re familiar with the idea that some of inbound can be automated.

But we also know that many of these so-called bots are really nothing more than ‘mechanical Turks.’

mechanical turks

Source

Before it was an Amazon product, a mechanical Turk was a fake AI – it looked like an automaton but inside was a person pulling the strings.

As soon as you get past the first couple of questions, most chat solutions are like this. A real rep has to be plugged into the back of them to deliver anything more than simple either-or queries and responses, and they mostly serve as a buffer for rep-lead interactions. What they don’t do is take a significant amount of FTE expenses off the company or work off a rep’s desk.

Conversica takes things a step further. Using automation backed by AI, it gives its users ‘always-on’ automated lead response that automatically qualifies leads, nurtures them and even works to stay in touch afterward.

conversica homepage

Source

In 2018 we can expect to see this kind of inbound automation become the norm, and the requirement for real reps get moved further along the sales cycle.

7. Sharper Outbound!

Back on our home turf. Outbound will be majorly affected by advancements in AI, automation and machine learning in the year ahead.

In fact, I’d go so far as to say, the division between highly-successful outbound teams – say, the top 20% of performers – and the rest of the field will largely be on the basis of how they adopt automation and AI assistance.

The challenges facing outbound teams overwhelmingly have to do with large numbers of prospects, large amounts of research and the relationship between the two.

If you have a whole bunch of prospects to call, you can’t do much research. If you do a whole bunch of research, you can’t call many prospects. And so, outbound teams have a foot on each stool, and the gulf is widening all the time.

With SDRs making 100+ calls a day, even the most carefully-designed sales process with personalized KPIs, clear role differentiation and solid, data-backed lists and goals, is going to struggle.

But they don’t have much choice: those outbound SDRs could be researching leads in depth – but they’d never keep up with the volume, or handle the masses of data required to even know for sure they were researching the right prospects.

Up until now, the scattergun approach – for all its flaws – has been the only game in town.

But it doesn’t have to be that way anymore.

We go into this whole idea in loads of detail in this blog post. Suffice to say; there can be huge changes in the efficacy of outbound prospecting – like, 2X or 3X changes – when you leverage the power of AI in ways that are already well-tested and understood.

Conclusion

Across every department of modern business, radical change is already underway. We’ve seen the tip of the iceberg in bots and the automation of basic administrative tasks. But in 2018, we’ll see early adopters pull away from the pack, powered by actionable insights gained from machine learning and supported by efficiency gains acquired from smart automation.

At the same time, easy-to-access tools that augment human work, or allow access to the fruits of low-level machine learning, will become available as standalone business tools, probably sold on the SaaS model. Simultaneously, enterprise providers will be playing catch up by buying promising startups or integrating this functionality into their own offerings from scratch.

How are you using AI and automation already – and where do you see it going for your industry this year?