Examples of AI in sales outreach
Can AI start high quality conversations in sales outreach? Early signs are mixed. In an industry starved for better ways to get customers talking, artificial intelligence (AI) may offer promise. On the marketing side, AI is already helping B2B marketers discriminate between buyers and visitors, convert more leads and send better leads to sales.
This article examines the potential for AI in cold and warm sales outreach -- based on actual examples I've researched. Indeed, we are seeing true, machine learning AI already initiating and engaging warm leads in conversations.
But what about cold sales outreach?
So-called AI apps are using email and text messaging in various contexts:
- Outreach to cold prospects -- to re-start discussions for hand-off to sales
- Immediate response to inbound “hand raiser” leads -- to filter warm leads to sales
- Nurturing of warm leads using email and web chat
What I've found is disappointing: Most sales outreach and engagement examples aren't truly intelligent. Machine learning isn't involved in most software apps. But this doesn't mean they're without value.
The promise of AI in sales outreach
The promise of artificial intelligence in sales outreach is simple: "Reach -- and engage -- more people, faster, at scale, with less effort." The promise to optimize messages also exists with AI.
In its simplest form, AI has the ability to see which outbound email messages outperform others during testing. Just like Facebook, Google or LinkedIn ad algorithms do. The rest is a matter of doing more of what (we observe) works, less of what doesn't. Plus, fresh iterative testing can also be included.
Ultimately the goal is to initiate more conversations with more efficient outreach (sales prospecting) efforts. Namely, messages that feel like they come from humans. The objective is to:
- Start more conversations with more contextualized messages,
- Advance those conversations forward for hand-off to a human.
Worth noting, AI is not being applied with starting, advancing and closing customers in B2B or B2C settings.
Apps like Tact.ai are supporting field sales reps to, for example, know who to visit, why and when using bot-based research tools and very basic machine learning.
AI is also being applied in the realm of warm leads -- email and SMS text interactions with customers. More on that shortly.
The reality of AI in practice
Most sales engagement (outreach) software tools billing themselves as AI aren't truly intelligent. Most are not using machine learning.
Are such AI tools reaching out and actually initiating conversations with customers? Are bots doing sales outreach in place of, or to enhance, human outreach?
The answer is yes but we're in early days. Humans (and anti-spam tools) are quick to pick up on patterns being used in most sales outreach. Here is one example:
Notice how the above highlights show variables (or "tokens") where automation is enhancing the customization of the message.
This attempt at relevance is a good example of how (what's being called) AI in sales is manifesting. It is a combination of automated web scraping (the article name above, for example).
How AI in sales works
Currently, AI technology tools are provided with human-created message templates or "formulas." Once these are loaded in the machines get to work.
This, so far, involves templates where a bot researches customer-side data and fills-in-the-blank. So far these are easy to discover observations which, often, leverage geo-location. These variables (or tokens) are then injected into messages. Example:
The first sentence can, and is, being automatically researched and placed within the message... automated. The goal is to create a more personalized, "human-like" message which doesn't seem automated. These examples come from a company called Kalendar AI.
To a large degree, "I know ___ about you, Mr./Ms. prospect" is becoming an established spam pattern. In short it's often seen as a cheap attempt to seem relevant. Can it be effective?
Yes. It depends on a handful of factors in our members' experience.
Here's another example:
Likewise, bulk emails can be sent out by the tens of thousands using such fill-in-the-blank observations based on locality, as seen in this example. But these are coming at a price.
While these kinds of messages are working for some sellers they can also backfire. Conceptually they rely on shallow observations which some customers already detect as low-quality ways to strike up a selfish, sales-focused conversation.
"I remember vividly once early in my career a prospect asking me, 'Oh, so do you know that restaurant on the corner next to XYZ place (that I called out earlier on the phone)?' When I lied and said I did, he replied 'Well, there isn't one, and I know you're lying' and hung up on me.
From that point on, I don't lie to prospects about fake things in their city just to try to be relevant."
Early success examples
Automated sales assistants are helping large and small organizations heat-up warm leads with impressive return on investment.
CenturyLink is achieving a 20-to-1 ROI and generating hundreds of leads per month it was not previously able to create. All thanks to an AI tool contacting 30,000 prospects per month.
“This is a huge volume, and the biggest benefit is we can do it quickly,” says Katie Cindric Federhar, Manager of Marketing Operations at CenturyLink.
Ms. Federhar says they’re able to have real (although AI-powered) conversation with each lead… a personal email thread that can cut through the clutter and get customers’ attention. AI, in this case, is purely being used to “warm-up” relationships for human sales reps to pursue.
The company started by reaching out to 300 accounts that did not have a sales rep. As a result reps (who were initially opposed to using an AI approach) clamored for more AI-driven leads.
AI wasn’t threatening sales -- it was powering better quality conversations.
Early disasters
At the same time juvenile forms of AI is powering an increasing number of email outreach blunders. I'm referring to what amounts to 1970's mail merge of data tokens aimed at hyper-personalizing emails.
For example, here is the opening line to one I recently received:
Aside from the comical (perhaps effective?!) subject line, the “HTTP Error 403” indicates the tactical nature of this email campaign: It’s being run by an unintelligent bot.
Do the pros overweigh the cons?
Yes, AI allows you to send "personalized" messages at scale. But at what cost?
With AI becoming more and more common, SPAM has become ever more prevalent. As a result, buyers are getting smarter. They're becoming increasingly familiar with AI technology and are able to identify copywriting patterns which:
- reference general information (likely grabbed from their website, LinkedIn profile, etc.) pretending as though they've done actual research
- have no relevancy to their problem; a generalized message
- educate -- providing unsolicited information
Don't get me wrong. Integrating AI does start conversations for some. But let's face reality: If you're sending emails to thousands of people, someone is bound to take the bait. But it depends on your target market, average deal size, B2B or B2C sector, etc.
As CEO of SocialRep, Chris Kenton mentions, the problem is "Businesses have learned to look at their markets as fish to be shot in a barrel as efficiently as possible, with no regard to those that swim away."
Exactly. Businesses have become so comfortable with the simplicity and scalability of AI that they've treasured the small percentage that engage whilst ignoring the majority they piss off.
Not considering negative metrics can hurt you in the long run.
AI can damage your reputation
You could send email messages at scale using AI. But who does it serve if you're messages aren't delivered?
Chris Kenton puts it best, "To the extent automation helps replace humans in the process of reaching and engaging customers, it's a problem, because it causes customers to recoil and put up barriers (DNC lists, can-spam laws, etc.)"
Customers are on the defensive. No longer are they taking AI and spam lightly. Moreover, buyers have put systems in place to keep spammers out of their inboxes.
To make matters worst, numerous AI companies (like all sales engagement or marketing automation providers) are developing bad email sender reputations. Meaning, integrating these technologies may lower your chances of being delivered. Beware.
How much sincerity must be faked to earn trust?
At some point AI in sales outreach becomes trickery. Depending on your philosophy attempts to "appear like a human being who knows X or Y factoid" about the prospect is trickery.
Some of our members send messages using AI technology. The difference however, is the incorporation of effective communication skills. Buyers are PEOPLE who desire HUMAN interaction. On the contrary, selling takes time. Additionally, we (sellers) don't have the time to individually send messages to leads.
But does that mean we can't send meaningful provocations at scale? Heck no.
Don't rely solely on AI
Regardless of the negatives, AI poses a lot of opportunities in sales. But it all depends on how you use it and the AI technology itself.
If you're relying on AI to formulate and send your messages at scale, it may hurt in the long run.
However, incorporating an effective communication tactic in the messages you send can yield better results. Moreover, consider sending messages at a lower scale. Remember, a proper sales automation tool should help you to "personalize" messages faster, not send faster.
How serious are you?
Sending messages at a higher scale won't matter if the message itself doesn't provoke curiosity.
And you can't provoke curiosity by blending in and sounding like the typical spammer.
Consider developing your message sending confidence and communication skills by joining our community of conversation enablers.
