AI

Applied AI in Enterprise Sales

The intelligence era is upon us and the playing field in B2B software has tilted.

In the past three years, IT has evolved from the SaaS workflow applications that characterized the cloud computing era to those that help customers make intelligent decisions.

In this intelligence era, the source of competitive advantage is rapidly shifting towards unique data and self-learning algorithms.

The arrival of this next phase comes at the convergence of major trends that have been in the works for the past three decades: cheap compute, more data storage and advances in algorithms.

As with the previous evolution in B2B software, this brings a change in the expectations of investors as well as a new opportunity set that emphasizes the strategic position companies are able to build through intelligent software.

Residing at the intersection of intelligent software, enterprise and voice, there is an emerging category in B2B conversational sales intelligence that is changing how sales teams work by providing tools designed to take advantage of a valuable strategic assets for enterprises — sales call data.

Voice: An Untapped Data Asset

Today, there’s a multibillion dollar market made up of tools designed to optimize email — email deliverability, open rates, accessibility, etc., but the market for call optimization and analytics tools is much less saturated. In 2016 alone, 425 sales tech startups received more than $ 5bn in venture capital, but very few products are focused on the fact that 70% of sales conversations happen over the phone, compared to just 10–15% that happen over email.

Now, with the relatively inexpensive and declining cost of storage and computing, vendors are able to capture and unlock this new unit of voice data.

Sales conversations are a treasure trove of data because:

They are generated in volume: the average salesforce spends almost 750 hours per year in meetings with customers and prospects (1,100 meetings x average online meeting time of 43 minutes), and

They are significantly more content rich than the CRM records they stand to replace: the average B2B sales conversation contains 6,000 words per hour versus an average of 30 words in a typical CRM record.

Conversational intelligence can leverage this robust data to address significant pain points that currently exist in enterprise sales teams. Currently, organizations have very little insight into what their sales reps are saying, why one rep might outperform another by 600%, or how to provide effective feedback and improve conversion performance.

Progressing an individual sales professional’s performance comes down to effective coaching and training, a process that currently requires a supervisor to listen to hours of calls per rep, then use a checkbox system to subjectively try to estimate how to coach them.

It’s one of their most time consuming tasks, and it’s a process that’s not currently data-centric.

Importantly, the post-call analytics pain point is most heavily-emphasized in medium and large, enterprise organizations with bigger sales teams (sized around 50 to 150 reps), where less proximity to teams causes significant management friction.

Meanwhile, SMBs don’t have the luxury, (given lower ASPs and capital constraints) to hire large field sales organizations, so default to efficient inside sales. Intelligent sales software has the potential to address a significant multi-level problem.

Macro Themes: Why Now?

The macro-environment provides a landscape to pursue compelling investment opportunities in intelligent sales technology.

Increased budget for sales teams

Selling into functional areas of enterprises can be a daunting task. Through the noise of excuses typically provided for not signing up a software vendor, the most common refrain within functional groups is that they simply, “don’t have the budget.”

Companies that sell into the sales departments of companies, however, don’t hear as much pushback around budget. The fact is simple: sales departments drive revenue, which for most companies is the lifeblood of their business.

While other functional departments, such as HR, finance, operations, etc. are constantly looking to cut costs, sales departments are becoming more empowered to leverage technology to drive revenue especially as it delivers a clear ROI.

They therefore have budget to spend, which makes for the best type of customer. In addition, most corporations need to demonstrate organizational productivity, further accelerating the need to experiment with new tools. Not to say that sales teams are price-agnostic, but on a relative basis they are becoming much easier to sell into.

Sales tactics have become more data-centric

Today, more than ever, sales departments are being measured on effectiveness, and they have large appetite for products that can help generate ROI.

VPs of Sales are empowered to make software buying decisions, but must do so based upon concrete, data-driven strategies. Intelligent solutions that can drive tangible improvements in sales efficiency are a natural extension to sales teams’ tech stacks.

Sales departments have embraced SaaS

While still in its early days, SaaS has seen step functions in adoption, and sales teams within companies, across SMB and enterprise, appear to be the ones that have most overwhelmingly embraced the SaaS revolution.

This dynamic produces a more knowledgeable customer base and therefore a shorter sales cycle.

Because early software applications were easy to set up and required minimal on-premise integrations, salespeople started circumventing the IT organization, which historically had been the area that laid the red tape so many enterprise software companies met during the selling process. The inertia has continued over time, and sales teams are more empowered to maintain technology purchasing autonomy.

Competition is heated

With the aforementioned empowerment, also comes the double-edged sword of rapid change. Buying cycles can be shorter, but vendors can also be supplanted more quickly.

Often times sales departments will purchase software as a trial or proof of concept, and then convert to a broader deployment over time, or otherwise shut the software off altogether. Consistent R&D investment to keep pace with evolving customer requirements is critical.

Strong exit environment

Large horizontal players (e.g., Salesforce, Oracle, SAP, IBM, Marketo) remain active acquirers of vertical software companies.

Most of these sales tech sub-markets are large, and can support multiple companies with meaningful revenue streams. Additionally, sales teams are heterogeneous buyers and the data sets being captured are fragmented across customer success, inside and outside sales channels. As a result of these dynamics, the ecosystem allows for multiple winners, mitigating the downside in not selecting the winner.

Market Size

Consider a rough bottoms-up TAM estimation. There are 2mm inside and outside sales professionals in the U.S. alone.

Assuming a median price point of the prevailing products of $ 1,500 annually per user, there is an addressable market of $ 3.0bn. Accounting for service and customer success reps, there are an additional 5mm professionals in the U.S., or a potential $ 7.5bn revenue opportunity.

Addressing an even larger pool of all outbound professionals, where the phone calls are the primary business communication (e.g., marketing, partnerships, fundraising, recruiting), there are 20mm professionals in the U.S., equating to roughly a $ 30.0bn market.

Evaluation Framework for ML-Driven Companies

There is certainly much hype around startups self-identifying as “AI companies”, so it is important to develop evaluation criteria specific to startups that are enabling ML to create a competitive advantage in their respective market. Below is a framework for evaluating ML-driven companies that can be applied to the B2B conversational intelligence for sales space.

Strong go-to-market enabled by ML

An innovative ML solution is about unique technology unlocking entirely new opportunities that customers view as 10x better than its replacement, rather than just optimizing existing opportunities. By changing the risk-reward balance of the buyer and potentially reducing the cost of customer acquisition, ML-based products can be disruptive, but a technology innovation alone is not enough.

Proprietary access to data

Algorithms are off the shelf and available to everyone. Creating proprietary data through product usage or through key early partnerships is essential to creating sustainable competitive advantage.

Full stack products, not platforms

Platforms that serve hundreds of thousands of developers (e.g., Google, Amazon, Apple, IBM) are likely to win the ML-as-a-Service business.

They have more R&D, lower costs of infrastructure and far more marketing dollars than any startup. End-to-end applications offering revenue increases, cost reduction and / or increased productivity are much more viable paths for a startup to go-to-market. Designing and owning the product interface allows for instrumentation and gathering of proprietary data.

Experts in the field

Algorithms are increasingly commoditized, but those stock APIs are calibrated to be as broadly applicable as possible and generate baseline results.

To deliver an exceptional experience (e.g., 95% transcription accuracy), a startup will need domain talent in speech recognition, natural language processing and other core disciplines. Likewise, selling these products requires trust, respect and relationships within the industry.

Product Solutions in B2B Conversational Sales Intelligence

Somewhere between 2009 and 2010 a new and exciting technology broke into the forefront of ML research. One of the areas that made the largest strides was in automatic speech recognition (ASR), the task of automatically transcribing voice recordings into written words.

The accuracy of transcription engines surged from around 84% in 2012 to almost 90% in less than two years. These R&D achievements have propelled ML-driven intelligence into the market, where it is now demonstrating tangible results in helping businesses increase sales efficiency by analyzing massive amounts of conversational data.

The advanced technology provides a certain baseline to compete in the space; however, the primary axis for competition will be in go-to-market. The winning product solutions in the space will focus on how the product increases revenue, decreases cost or increases productivity (read: wins the buyer a promotion).

As observed in other enterprise software markets, market leading companies will be those that can layer data analytics on top of workflow automation to embed the product into a customer’s organization and create a sustainable competitive advantage.

Interoperability with existing tech stacks across CRM, HR management, etc. will allow multiple teams to orchestrate workflow and leverage new business intelligence in these new products. To that end, this emerging part of the sales value chain has the powerful potential to complement the full cycle of training, onboarding and development of sales teams.

Below is a breakdown of the current product features in conversational sales intelligence and their current value propositions that they seek to offer across members of the sales team, from VP of sales to account executive:

Peer- and self-coaching

Natural language processing and ML tools can deliver real-time sales coaching to deliver insights at scale and to conduct, through semantic (content) and sentiment (emotional) analysis. Individual sales reps and managers can review top performers’ winning sales strategies, increasing sales efficiency through identifying and socializing sales best practices.

Manage teams of sales reps

Reduce the time Managers and VPs of Sales spend reviewing the calls of sales reps. Supervisors can review calls and orchestrate real-time workflow as well as drive insights at scale through on-demand access to sales behavior.

Call recording and transcription alone don’t solve the whole problem. For example, if a VP of Sales wanted to provide feedback to her reps, she would have to shadow them or listen to their recorded calls to be analyzed at a later point in time, a cumbersome and time-consuming process that results in a fairly subjective analysis.

Reduce costs of onboarding new sales reps

Rather than relying on “trial by fire” to onboard new hires, sales team can build training programs using highlight reels to illustrate how top performers handle specific topics, manage the flow of the meeting, etc.

Market Landscape

Industry giants are announcing more AI plays — Salesforce acquired Metamind and launched Einstein, Microsoft rolled out Dynamics 365 with AI technologies Cortana, PowerBI and Azure ML.

Current voice solutions exist for enterprise sales like Salesforce’s Einstein or IBM’s Watson; however, these products primarily are deployed on the periphery of larger horizontal product suites and offer limited competition with startup vendors providing more powerful, heavyweight, point solutions.

Likewise, incumbent sales tech vendors, like InsideSales, SalesLoft, Persado and Verint are fundamentally sales enablement solutions which offer sales intelligence applications for top-of-the-funnel lead generation, and they typically reside in a different area of the value chain from the emerging class of advanced startups which are driven by voice.

Below are the emerging leaders in the category:

Gong.io

Gong.io, based in Tel Aviv with offices in San Francisco, came out of stealth with a $ 6mm Series A in June 2016, led by Norwest Ventures Partners and recently raised a Series A1 follow-on round of $ 20mm in July 2017. Gong uses NLP and speech recognition to make real-time suggestions to help steer the sales conversation and to provide training analytics.

Gong’s predictive analytics are primarily based on keywords to determine likely outcomes, but also is capable of detecting emotion in conversations and prescribing tactics to redirect the conversation. The software analyzes conversations from most audio sources, as well as web conferencing platforms (e.g., Cisco WebEx, GoToMeeting, Zoom), and then feeds the results to CRM systems.

Gong is initially focused on B2B sales in the mid-market, and 30% of Gong’s business today comes from outside sales and in other areas of CRM. The company has doubled revenues the last four quarters and claims its solution has contributed to a collective $ 1.0 billion in revenues among its customer base, which includes Act-On, SalesLoft, Sisense, Greenhouse and Zywave. Customer sentiment has been impressively positive, and it recently reported a 30-day NPS score of 76.

Gong was recognized by Gartner as a top conversional intelligence vendor in 2017 and by G2 Crowd for “Best Sales Coaching & Onboarding Software”. Gong.io currently counts 40+ employees and has rapidly has accelerated hiring of researchers and engineers in speech, NLP and related areas to improve user experience and data science. Following its recent funding, the company plans to add headcount in sales and marketing.

Amit Bendov (CEO and Co-Founder) has run software companies both as a CEO and in the C-suite executive. Prior to Gong, he served as the CEO of business analytics platform SiSense.

Eilon Reshef (CTO and Co-Founder) previously co-founded SaaS content management platform Webcollage, which was acquired by Answer.com in 2013. Together, they have assembled a world-class team with exceptional R&D backgrounds coming from SiSense, Panaya, Webcollage and ClickSoftware.

VoiceOps:

Voiceops, a Y-Combinator alum, breaks down core skills and highlights behavioral differences between the best and worst performers, scaling the winning behavior to the rest of the team. They are targeting PMs rather than individual sales users and have attracted initial customers, including Weebly, Advent, Livestream and Intermedia.

The company has validated its product with strong early product performance: one customer is experiencing a 5% lift in successful close attempts, which at a large organization like theirs equates to millions in revenue. Another customer has seen an 85% reduction in time spent preparing for coaching sessions and a 40% increase in close attempts, outcomes which lead the client to eliminate manual quality assurance from their sales process altogether. The company has analyzed more than 1.5mm statements and reports 99% transcription accuracy.

VoiceOps raised a seed round in April 2017, led by Accel investor Steve Loughlin, previously head of Salesforce Einstein and CEO of RelateIQ before selling to Salesforce in 2014. This was his first investment at Accel.

The team comprises engineers and data scientists from Harvard and Yale, and company founders Ethan Barhydt, Nate Becker, and Daria Evdokimova have cumulatively spent roughly a decade building sales analytics and customer support tools for Google, LinkedIn, Coinbase, Gusto and General Assembly. The current team size is less than 10 employees, but is actively expanding.

Chorus.ai

Similar to VoiceOps, Chorus is optimized for transcribing calls; however, the company’s unique edge in speech recognition, NLP and AI technology enables teams to immediately discover moments that impact selling outcomes and use those insights for real-time sales coaching and for replicating sales best practices across organizations.

Chorus unites the entire sales team’s conversations into a single dashboard view, as well as a “playlist” feature for curating call moments, and then integrates the data with CRM systems, freeing reps from having to take copious notes, and scaling a manager’s ability to coach their team to move deals forward without having to monitor or review every call.

The company raised a $ 6.3mm seed round in October 2016 and graduated in quick succession with a $ 16.0mm Series A in February 2017, with follow on participation from Emergence Capital as well as new investor Redpoint Ventures.

The team at Chorus has already attracted high-profile pilot customers in the mid-size sales category, including Qualtrics, Marketo, Cisco, Talkable and Dynamic Signal, which in 2016 alone, used the software to analyze more than 500,000 sales conversations. This proprietary data acquisition has been valuable in further augmenting Chorus’s ML engine, which reports just two weeks of product usage for mid-size customers to achieve positive results from its learning models.

The company plans to double down on the real-time advantage of its platform. For instance, if a customer on the phone references a competitor, Chorus could flash an informational aid on screen with known differentiators and past successful pitches to give the sales rep a smarter “ace card”.

Cogito

Based in Boston, Cogito spun out of the Human Dynamics Lab at MIT and raised a $ 16.0mm Series B in November 2016, led by OpenView. Cogito’s software evaluates hundreds of behavioral signals through voice to provide real-time coaching and customer experience analytics.

Agents are guided to speak with more empathy, confidence, professionalism and efficiency, depending on the emotion detected through callers’ speech, while early signs of customer frustration and intent to purchase help improve service and close deals. Humana, for instance, reported that Cogito improved customer satisfaction by 28%, improved employee engagement by 63% and increased the resolution of first-time callers by 6%.

The company has attracted an impressive customer base which includes organizations with large call center workforces like Humana, Zurich and Blue Cross Blue Shield, which have deployed the software across workforces ranging from the several hundred to several thousand.

Cogito will continue to focus on healthcare, insurance and financial services customers as well as expand its use cases beyond customer service into other inside sales functions. Hiring trends are also accelerating, as the company counts 75 employees and has already outpaced its 2016 goal of doubling its 40-person workforce by Q4 2017.

Other Market Entrants

Other startups are gaining customer traction and have raised seed funding in the past 18 months. Startups TalkIQ, Oto.ai, Qurious.io and Tethr are building competitive products, particularly in more advanced features like real-time call analytics.

What Will Determine the Winner(s)?

The winning software solutions in the conversational sales intelligence space will outperform in the following areas.

What is the data moat?

Novel algorithms are increasingly making their way into the public domain, so, the key differentiator for startups and ultimately long-term competitive advantage is access to proprietary data sets.

Because the data sets are far more fragmented and because the application of ML in the enterprise world is emerging, B2B startups have the opportunity to accumulate proprietary training data as a competitive advantage.

Building best-in-class models that enable conversational analysis at a fraction of what it would cost now, minimizes customer churn and makes it virtually impossible for a new entrant to get to that level of accuracy at the same costs.

Can the product rapidly penetrate and integrate with user workflow?

ML that can go the “extra mile” and disappear into the product can reach adoption at scale throughout an enterprise customer.

Companies that can build long-term trust in the product through empathetic product design that creates simple UI and non-obstructive, real-time coaching will achieve quick CAC payback cycles, an accelerated path to profitability and, ultimately, valuable long-term customers (LTV / CAC).

Can the startup validate ROI to multiple constituents of a sales department?

Sales organizations are nuanced and comprise stakeholders with varying incentives across AEs, Sales Managers and VPs of Sales.

Sticky products that demonstrate effective sales outcomes can win approval from multiple layers of an organization, leading to opportunities for sales expansion and negative net churn metrics, and, consequently, a defensible market position.

Is there a path to operating leverage at scale?

Many of the early-leaders have selected the mid-market entry point, though sales intelligence has the ability to impact businesses of all sizes, particularly the whitespace in SMBs, which don’t have the luxury (given lower ASPs and capital constraints) to hire large field sales organizations, so default to efficient inside sales.

Startup vendors who can demonstrate an early path to operating leverage will be best-positioned to expand their customer aperture and capture valuable shares of the overall market.

Let’s block ads! (Why?)

Machine Learnings – Medium

Comments
To Top