The “Growth Loop” – A Deliberate Go-to-Market Approach for AI Firms

Daniel Faggella

Daniel Faggella is Head of Research at Emerj. Called upon by the United Nations, World Bank, INTERPOL, and leading enterprises, Daniel is a globally sought-after expert on the competitive strategy implications of AI for business and government leaders.

The B2B AI Growth Loop

Most AI product firms are founded and grown in a similar way.

It usually goes something like this:

  • Founder(s) develop a hypothesis about how AI might improve a particular business result
  • Money is raised, an initial team is built, and a beta product begins development
  • Initial prospect conversations make it clear that the firm is wrong about it’s go-to-market hypothesis (almost no firm gets premonition right, per von Moltke’s famous quote), and they need to recalibrate their hypothesis to consider what they’ve learned about:
    • The motives of stakeholders
    • The current state of IT and data infrastructure of prospect firms
    • The workflows their product impacts
    • Where budgets do and do not exist within prospect firms

This usually leads to a phase where the startups has dozens – maybe hundreds – of haphazard conversations across industries, groping to find someone who might use their product. They tell prospects about their product (“hammer”), and hope the client will reveal some kind of problem that said product can help with (“nail”).

Startups in this phase sometimes spend years talking about their product as a set of abstract AI capabilities. Examples might include:

  • “Our hybrid AI approach allows you to improve your NLP-based search processes”
  • “Our computer vision solutions and digitize documents quickly”

This “hammer looking for a nail” phase is one that most companies could skip if they have a more deliberate process for finding product market fit.

We refer to this process as the Growth Loop:

Growth Loop (with Slide Details)
A screen shot from Emerj’s media kit for advertising services. Learn more about Emerj Media and the full media kit here.

Successful startups don’t “boil the ocean” by taking their AI capability-focused message into as many industries and prospective customer types as possible, hoping to see something stick.

The Growth Loop is an approach to achieving traction faster by deliberately testing target market messages, and iterating quickly to pivot from failure and double down on success.

This is the basis of our Emerj Media work with AI startups, and it’s a framework we encourage startup marketing leaders to apply well after their Emerj campaigns are over.

In the remainder of this article we’ll explore each of the four phases of the Growth Loop in greater depth:

1. Market Message

Developing a Core Marketing Message
Source: Developing the Market Message for an AI Product or Service – Finding a “Core Message” That Works

The first step in a deliberate go-to-market approach is to determine:

  • A prospective buyer profile (How)
  • A value proposition (What)

For firms who have already built much of their core product, the answers to these questions are bounded. For firms whose product is still relatively open-ended, the possibilities are much more broad.

The question to answer in either case is the following:

“We would have over 10x our current revenues if what group of prospects knew us for solving what problem?”

Inevitably, most firms will discover that there could be many potential answers to this question.

Most AI firms have “explore” market messages, not “exploit” market messages. Here is the crucial difference between the two:

  • An “exploit” message is one that is one that consistently generates new opportunities and sales.
  • An “explore” message is one that we think might have the potential to 10x our business, but we don’t have enough evidence to know.

The best way to limit the wide possibility-space of “explore”-able market messages is by looking at where you have the most current traction. The most important element of traction is sales, but other proxies can help to paint a richer picture in a company’s early days of go-to-market.

  • What kind of roles and titles are replying to your emails?
  • Which industry outreach efforts have lead to the most phone appointments and opportunities?
  • Of the current opportunities, what was the value proposition that “hooked” the prospect and made our product a possible fit for them?

Once these factors have been considered, list out all of the viable combinations of industry, title/role, and value proposition, and then order them from most likely to 10x your revenue to least likely to 10x your revenues. The end result might be look the following:

  • “Explore” Message 1: [Industry] Life Sciences, [Titles] Head of R&D, Pre-trial R&D, Post-trial R&D, Head/Director/VP of Compound Generation, [Value Proposition] Improve speed of drug discovery, find insights from previous experiments faster/make use of existing data stores
  • “Explore” Message 2: [Industry] Manufacturing, [Titles] Head of AI, Head of Digital Transformation / Innovation, [Value Proposition] Unlock the value of sensor data streams to reduce downtime and improve efficiencies
  • “Explore” Message 3: [Industry] Life Sciences, [Titles] Transformation / Innovation, Smart Health, [Value Proposition] Achieve AI pilot projects faster and prove the potential of ML to leadership
  • “Explore” Message 4: [Industry] Utilities, [Titles] Transformation / Innovation, Smart Health, [Value Proposition] Unlock the value of sensor data streams to reduce downtime and improve efficiencies

Note that it’s possible to test multiple value propositions or multiple different combinations of titles and roles in a single industry.

A marketing team that deliberately updates, orders, and works through their top market message combinations is much more likely to “strike oil” than a team that focuses on having a thousand haphazard conversations across a dozen industries to “see what sticks.”

Prioritizing and ordering allows a marketing team to develop targeted, specific collateral and campaigns, and quickly rule out markets that aren’t a fit – or double down on those that are.

As a general rule, we don’t recommend that marketing teams focus on more than 2-3 “explore” markets at any time.

2. Thought Leadership and Positioning

Gravity Assets - AI Marketing and Sales - Emerj
Source: Best Practices of Growth – B2B Lead Generation Blueprint for AI Products and Services

Once we’ve determined who we’re pursuing, and what value proposition we’re leading with, we can determine what kind of content we need to lead with.

For enterprise AI, there is no way to simply list your benefits in bullet points, list your pricing tiers, and put it out into the market to take orders.

AI is an educative sale.

The AI firms that win are the ones that “turn on the light bulb” of their prospects, and show them how AI could help them to:

  • Solve a persistent, important problem, or
  • Take advantage of an opportunities they haven’t been able to previously capture

This can be done through a variety of media, from podcasts to white papers to research to webinars and beyond.

For “explore” messages, AI firms should think about creating educational but relatively affordable PDF assets or webinars in order to win new leads, start new conversations and test the

For “exploit” messages (i.e. those messages that have consistently proven to attract new opportunities and sales) require more substantial content investments, which might include ongoing commitments to blog and social media content creation around said topics. We generally do not advise vendor clients to create ongoing blog or media content unless they have a proven “exploit” message – as ongoing content creation is a substantial effort and expense.

Regardless of “explore” or “exploit” message maturity, any AI go-to-market effort requires content assets.

3. Awareness and Lead Generation

Once we know who we’re marketing to, what value proposition we’re presenting, and what content assets we’ll be using to attract and engage them, it’s time to get in front of as many of them as possible and drive appointments, demos, and (hopefully) sales. At bear minimum, through enough appointments we will learn enough to discover whether this market is worth investing in into the future.

The two channels that we see most commonly among successful, high-growth AI startups are:

  • Outbound email and LinkedIn campaigns – There is no substitute for direct engagement from well trained salespeople. Nearly all the successful AI startups I’ve observed have demand generation teams capable of identifying ideal target customers and pursuing them directly. There are many great resources on this topic, so I’ll refrain from elaborating too much more on this topic.
  • Content and email campaigns on industry publications – This is our focus at Emerj Media, particularly in the industries where we have a sizable audience, including: Financial services, life sciences, heavy industry, retail/eCommerce.

Other channels might include:

  • LinkedIn or Twitter PPC advertising
  • Conference or event sponsorship
  • Public relations (usually with the help of an outside firm)

The key for all “explore” messages is make sure that all marketing investments can be held accountable. The purpose of a prioritized list of “explore” messages is to quickly and efficiently determine which of them have the potential to 10x our business, which of them need to be refined/altered/retried, and which of them need to be eliminated.

For this reason we generally advise companies to only invest in PR firms when they have found a strong “exploit” message. We’ll discuss accountability in greater depth in section below.

4. Learn and Capitalize

Holding Campaigns Accountable

Content (thought leadership) and campaigns (outbound email, PPC advertising, sponsored content, etc) are only combined when they can be tested.

Testing budgets and criteria of success will vary based on the nature of the campaign and the war chest of the marketing team, but they should almost always involve the following elements:

Financial Budget

This is the most self-explanatory. Whether it’s $5,000 or $50,000, an “explore” campaign must have a bounded budget. When the resources are spent, the team should consider this as a checkpoint to decide whether or not to continue.

Factors to consider:

  • Will your budget include the time of your marketing staff focused on this campaign, or just the money spent on content and assets with outside firms?
  • How much of your budget will you spend on upfront content assets (PDFs, webinars, etc) compared to the spend on reaching new audiences (email, PPC, etc)?

Criteria of Success

B2B sales is always challenging because of the time it often takes to close deals. This can often be vastly more complicated with AI solutions as compared to enterprise SaaS sales.

Success always has to do with winning new business – but this can vary drastically depending on the sales cycle of the product.

For some more surface-level AI solutions – or solutions selling to other startups (with vastly more AI fluency and faster sales cycles than enterprise) – the success of a three-month push into a market might even include closed deals. For a three-month test of a market message to the enterprise, or for a complex, integrated sale, success might be measured by number of sales opportunities generated.

Example thresholds:

  • “Explore” Message 1: [Industry] Life Sciences, [Titles] Head of R&D, Pre-trial R&D, Post-trial R&D, Head/Director/VP of Compound Generation, [Value Proposition] Improve speed of drug discovery, find insights from previous experiments faster/make use of existing data stores
    • Success Target: Six qualified sales opportunities in the pipeline by [date]
  • “Explore” Message 2: [Industry] Manufacturing, [Titles] Head of AI, Head of Digital Transformation / Innovation, [Value Proposition] Unlock the value of sensor data streams to reduce downtime and improve efficiencies
    • Success Target: Two sales of at least $15,000 in contract value by [date]

On a quarterly basis, marketing leaders should determine (a) which “explore” messages to test, and (b) what thresholds or success targets to set for each market message campaign. At end-of-quarter, leadership should pull together numbers and determine where to iterate and where to double down.

Iterating or Doubling Down

Once a campaign has run for enough time, or accrued enough evidence as to being on or off target, leadership must decide (a) how to double down on successful campaigns, and (b) how to change course with unsuccessful campaigns. In each case, changes can be made at the level of target audience, content, or traffic source.

We’ve found it easier to help marketing leaders deliberately adjust their campaigns by providing a simple “menu” of next step campaign options. Below, we’ll explore these in depth.

Actions to take when an “explore” message campaign is achieving or exceeding success targets:


    • Expand Titles: A market message that is working well with Heads/VPs of Sales might be tested with a broader set of related titles, such as CMOs, CROs, Directors of Demand Generation, etc.
    • Expand Industries or Sub-Industries: A market message that is working well in US retail banking might be extended to larger credit unions, or to banking leaders in Western Europe. The same message might also be tested in industries with similar problem sets or workflows – such as insurance or wealth management (if the product is applicable in these areas).


    • More Robust Demand Generation Assets: Develop more rich research-backed PDF assets, or a series of related webinars or explainer videos. Higher levels of content investment are warranted only when a message is winning.
    • Long-Term Content Commitments: When a truly “exploit”-able market message is found (i.e. when we’ve found a market and value proposition that is consistency delivering opportunities and sales, and has the ability to at least 10x the business), it often behooves a firm to commit to an ongoing calendar of blog and social media content.


    • Double-Down Investment in Working Traffic Sources: If a PPC channel is working, running more ads and creative variations may be warranted. If a publishing channel is working, testing more email inclusions and more sponsored content may be warranted.
    • Traffic Source Expansion: Expanding from one PPC source (i.e. LinkedIn) to another (i.e. Twitter, Facebook look-alike audiences, etc). Expanding from one B2B publication to another. If event sponsorship and speaking is working – trying more events with similar audiences may be warranted.

Actions to take when an “explore” message campaign is failing to achieve success targets:


    • Change Titles or Industry: If the feedback from the market (preferably from early sales calls) is that a different role or title, it may be worth exploring those new titles. For example, a sales team may find that the CDOs aren’t resonating the a market message, but Heads/VPs of Operations are responding more favorably.


    • Change Value Proposition: If banking fraud leaders aren’t responding as much to an efficiency-oriented value proposition, but they seem to be concerned adamantly about compliance-related issues, it might be warranted to create PDF or webinar assets related to how your solution can reduce regulatory risk.


    • Change Creative or Messaging: If the value proposition still seems viable but clicks seem low, changing the images or text used in email or PPC ads may be enough to change a losing ad campaign into a winning one – or at least collect more feedback from live sales calls to make an educated decision about the campaign’s potential.
    • Change Traffic Source: Some publications or PPC channels may simply not be a fit – and if click prices or lead quality is weak, it’s possible to salvage a campaign through deliberate experimentation in other traffic sources.

Generally, campaigns that are not hitting their targets fall into one of the following categories:

  1. Not enough clicks or engagement from the campaign to know if the message is resonating with the prospects.
  2. The message or value proposition is not a fit for the target prospects, or the prospects are without power/budget.

Never confuse a category 1 campaign failure with a category 2 campaign failure.

If a campaign simply isn’t getting enough response (clicks, opt-ins, appointments) to gather feedback, it may be a weak message, or it may simply need adjusting to win appointments. If a campaign has resulted in real conversations with viable prospects, and it is clear that the value proposition is not a fit for those prospects – this is a much more firm signal that the market message itself needs to be adjusted or abandoned, and that optimizing lead channels (clicks, opt-ins) is not the core issue.

It’s important to note that salespeople are the most important source of feedback about early traction from “explore” messages, and sales leadership should be involved in the decision to prioritize market messages, and to iterate or double down on them.

Concluding Thoughts

The key difference between AI firms who find market traction and those who do not lies in how deliberately they approach their “explore” phase. Even successful AI firms often fight for years to arrive at a truly “exploit”-able market message. To put it simply, smart marketing leaders at AI firms:

  • Openly embrace the challenging, hands-on process of finding product market fit for enterprise AI products
  • Deliberately select “explore”-able market messages (specific target audiences matched with specific value propositions)
  • Set explicit success criteria for “explore”-able market messages
  • Iterate or drop failing campaigns, and double-down on winning campaigns
  • Maintain consistent feedback loops with sales leadership about which prospects or industries are showing promise, and about which campaigns and value propositions are garnering the best response

Emerj Media for Marketing Leaders

Emerj Media is the arm of focused on publishing, and connecting AI vendor firms with enterprise leaders at scale. If you’re interested in refining your market message, developing powerful thought leadership content, or opening new channels to reach AI-focused executives – learn more on our Emerj Media page.

Stay Ahead of the AI Curve

Discover the critical AI trends and applications that separate winners from losers in the future of business.

Sign up for the 'AI Advantage' newsletter: