The Robotic Gold Rush
The worlds most prolific crypto-currency expert isn’t very smart, yet overcomes below-average intelligence with speed, specialization, work-ethic, and an uncanny ability to learn. Borrowing a lesson from Wall Street, he doesn’t use his own money. He doesn’t even trade. His time-honored formula for getting rich in a gold rush: don’t search for gold — sell picks and shovels to miners. He’s also mercenarial: anyone with .028 Bitcoins a month (today about $300) can hire him. Although his best stuff costs 3 times that.
Meet Haasbot who, along with his brethren bots, collectively accounts for the bulk of global cryptocurrency trading: Forbes estimated 80%. Bots also dominate equity trading — the percentage is harder to gauge given Wall Street’s penchant for secrecy and the relatively closed environment of traditional exchanges.
As noted, bots aren’t particularly smart — they’re the computer equivalent of manual labor. Remember Judge Smails in Caddysack: “the world needs ditch diggers, too.” Factory robots — neither smart, innovative nor charming — best humans with speed, cost-effectiveness, precision and persistence. The same qualities Haasbots bring to crypto trading.
Factory robots and Haasbots perform Robotic Process Automation (RPA), a slice of the unfolding-in-front-of-our-eyes Fourth Industrial Revolution (4IR). Whatever 2.0 (or 3.0), pivot, game changer, seismic shift — they’re all trite and understate what’s happening, so Salesforce.com went with 4IR: “a new chapter in human development, enabled by extraordinary technology advances commensurate with those of the first, second and third industrial revolutions.” ¹
RPA bots aren’t new — they’ve been around for decades, earning a lousy reputation by serving as henchmen for shady programmers. Cisco categorizes bots as malware along with viruses, worms, and Trojans. Bots don’t deserve this bad rap — they’ve just been following orders (unlike humans, bots have no choice).
What is new is machine learning. Now bots teach themselves. Thankfully, as of now, they’re still following human orders.
IBM’s Adam Cutler says todays’ “smartest” AI has an IQ of 47. An Imbecile (by traditional classifications) — who not long ago was an Idiot, and will soon progress to a Moron. So machines aren’t yet scary smart. But, without humans holding them back, they learn scary fast. When converted to machine-learning, Google Translate improved about as much overnight as the old version had in a decade. It translates 150 Billion words daily — as accurately as a human, who can translate, on average, 450.
We’re still a ways away from C3P0, HAL 9000, or even Siri making doctors appointments. Who knows how far away — Elon Musk says todays “double exponential change” will consistently beat predictions. For now, AI is micro-specialized — essentially computerized idiot savants. So trading cryptos requires an entire bot team:
- Trade Bot
- Accumulation Bot
- Order Bot
- Ping Pong Bot
- Scalper Bot
- Trend Lines Bot
- Intelli Bot Alice
- Mad Hatter Bot
- Inter-Exchange Arbitrage Bot
- Market Making Bot
- Flash Crash Bot
- Zone Recovery Bot
Haasbots can execute a pump and dump crypto trade. T-3's ‘Trump and Dump” bot expands on that, analyzing our erratic Chief Executives Twitter feed for public company mentions with negative sentiment — then shorting their stock.
Haasonline Software is well-known in the crypto space (cryptocurrencies are another 4IR gold rush). Household names dominate AI — Google, IBM, Facebook, Apple, and Amazon get the attention. Yet the most useful AI capability I’ve seen comes from elsewhere in Silicon Valley.
By the scale of the FAANGs Salesforce.com is a nit — under $10 Billion in revenue, “only” 150,000 customers. But they’re the center of a $250+ billion ecosystem, and growing like mad. Introduced last year, their Einstein AI is changing the lives of millions of salespeople — and hundreds of millions of Salesforce’s customers customers.
Since AI is expected to decimate jobs, announcing “we’re implementing AI to help with your job” is akin to saying “I’m from the government and I’m here to help.” Yet Einstein does help, at least in his RPA role, serving as a salespersons best friend: the secretary.
Secretaries, eg administrative assistants, are the unsung heroes of sales. Comparing low and high-performing sales teams, McKinsey found the best ones devote over twice the staff to sales operations and admin. Quoting their study: “Sales operations and administrative support are sometimes a victim of overly aspirational cost-saving efforts. Yet they are invaluable because they enable frontline and pre-sales employees to spend more time with customers and to focus on sales and growth.” ² Yet this support is disappearing.
Lacking help, it’s not surprising one study found that salespeople spend just a third of their time selling³. Because survey-based studies are susceptible to the Margaret Mead factor — “what people say, what people do, and what they say they do are entirely different things,” Pace Productivity instruments salespeople with stopwatch-like gadgets. Based on precise time monitoring, Pace’s 2017 results show that a salesperson that spends 33% of their time selling should be thrilled:
Selling time is lower today than twenty-five years ago — pre-CRM. In 1991, StorageTek, a major mainframe-era peripherals vendor, made news with an ambitious salesforce automation project. Six months after implementation, they credited the system with increasing their reps selling-time percentage from a third to 45% — double todays average.
Despite CRM, the tyranny of tiny tasks continues to increase.
While salespeople grumble about the overhead of CRM, its not the problem. Email is.
McKinsey found that email consumes 28% of an average workers’ time. Since CRM and email are different IT silos, CRM doesn’t reduce this time — it increases it. Salespeople are “human APIs” connecting the two systems. Before email, salespeople dictated or pecked-out rough letter drafts. Admins took over to edit, print, mail, copy and file. The transition to email was a double whammy for reps. They took on the entire burden of written correspondance, and this transfer of labor cracked the door for the aforementioned overzealous cost-cutting.
Managing email isn’t hard — just repetitive and time consuming. In a good CRM implementation, “filing” an email only takes a few extra seconds. Every. Single. Time. That’s the tyranny of tiny tasks. Plus context switching means an simple email alert burns a minute or more.⁴
So sales operations is the perfect candidate for RPA.
Einstein devours these tasks. It continually monitors and “files” emails with the company and opportunity they relate to. It checks calendars to do the same with appointments. Einstein notices new names in emails — say you’re close to winning a deal and counsel appears. It sorts out the particulars and asks permission to create a new contact. In concert with other software, Einstein also manages schedules like a secretary, “negotiating” meeting times even as calendars fill up. You send one email, and the rest get done without lifting a finger — even rescheduling.
Salesforce.com figures all these little things will add up, on average, to a 6 hour weekly savings per rep. That’s a lot! Especially since half of all reps don’t make quota.
Einstein’s “intelligence” comes from:
- Discovery — analyzing data to discern potentially useful information
- Prediction — extrapolating likely outcomes base on discovery
- Recommendation — suggesting actions to achieve outcomes
- Automation — completing actions with minimal or no human intervention
None of this is new. Combining them, and adding machine learning, is. And that’s where things start to get spooky.
As Einstein scours emails, it notes insufficiently active deals — defining sufficient on its’ own using historical data. Salespeople and executives both appreciate early warning of slipping deals. Sales managers cheer the absence of vapid deals that show up on forecasts each month like the full moon. But the “big brother” aspect is unnerving. John Williams, the CIO in charge of StorageTek’s innovative sales force automation, hid the project from executive management for 6 months because: “They [sales] needed to use it without the fear of thinking management was checking up on them.”
Yet that’s exactly what Salesforce’s CEO Mark Benioff is doing with Einstein:
“this is a capability that I use with my staff meeting, when I do my forecast and I do my analysis of the quarter, which happens every Monday….We have our top 20 or 30 executives around the table……..And then I ask one other executive their opinion, and that executive is Einstein……I will literally turn to Einstein in the meeting and say, ‘OK, Einstein, you’ve heard all of this, now what do you think?’ And Einstein will give me the over and under on the quarter and show me where we’re strong and where we’re weak, and sometimes it will point out a specific executive, which it has done in the last three quarters, and said that this executive is somebody who needs specific attention during the quarter…….”
He went on to add:
“For a CEO, typically the way it works is, of course, you have various people, mostly politicians and bureaucrats, in your staff meeting who are telling you what they want to tell you to kind of get you to believe what they want you to believe. Einstein comes without bias.”
While his customers get their feet wet, Benioff has progressed to power-user: humiliating executives in front of peers, and putting them on the hot seat. How long before Einstein makes RIF decisions?
In contrast to Benioff’s anti-social application of AI RPA, the other emerging AI category is “social bots.” This isn’t new either — 20 years ago I implemented a customer-service chat bot. I had to “teach” it everything, and it wasn’t very good. A few years ago chat bots re-emerged. They still weren’t very good. Actually, they were awful, and so disappeared — until a recent renaissance.
Gartner excitedly predicted that 85% of consumer interaction would be with automated agents by 2020. Thats not happening. Siri, Google Assistant and Alexa are amazing at answering questions, but human conversation is incredibly complex. As a case in point, here’s a recent “conversation” with Microsoft’s “Mo” Messenger bot (I’m in blue):
Mo admirably tries to constrain the conversation without moving to strictly multiple choice questions as the most successful bots are doing. But it fails — and sounds like a complete idiot.
So chat bots aren’t mimicking (intelligent) humans yet. But they will. Germany already passed a law requiring bot conversations to be labelled as such, getting ahead of the ball. In B2B, bots like Drift now get positive user reviews and deliver positive ROI (we’ll be doing a Drift test case in an upcoming post).
Like guns or nuclear weapons, the AI genie is out of the bottle. In The Coming AI Wars, the Huffington Post said: “every job a human can do will be augmented by…and possibly replaced by AI.” AI is just one piece of 4IR. It’s like the Web in 1995, being played at 10x speed.
Consider this blog a 4IR treasure map: a guide to the emerging 4IR marketing and sales stack, and to transforming people and process to leverage it.
Next post: why coffee isn’t just for closers anymore.
¹ World Economic Forum https://toplink.weforum.org/knowledge/insight/a1Gb0000001RIhBEAW/explore/summary
² In the Best Sales Teams, About Half of the People Are in Support Roles
³ 2015 Insidesales.com
