Agentic AI for business support: a new approach to the 3Ms

Executive summary

For decades, researchers and practitioners have understood why entrepreneurs from underserved communities struggle to build sustainable businesses. The barriers are well-documented: limited access to capital (Money), restricted networks and customer bases (Markets), and gaps in business knowledge and coaching (Management). This framework, known as the 3Ms (Bates et al., 2007), has guided billions of dollars in philanthropic and public investment aimed at leveling the entrepreneurial playing field.

Yet the barriers persist. Despite an extensive ecosystem of small business development centers, community lenders, accelerators, and technical assistance programs, underserved entrepreneurs continue to face the same fundamental challenges their predecessors faced a generation ago.

This paper argues that agentic AI (artificial intelligence that can take autonomous action on behalf of users) offers a genuinely new approach to this persistent problem. Unlike previous interventions that struggled to scale due to cost, geography, and human resource constraints, agentic AI can deliver personalized business coaching to any entrepreneur, in any language, at any hour, at near-zero marginal cost.

This paper lays out the thesis and the open questions that come with it: Does AI-delivered business support actually improve outcomes? What's the right balance between human and AI assistance? And can this approach finally break down barriers that have proven stubbornly resistant to traditional interventions? We're answering those questions the practical way — in live deployments, alongside a small cohort of founding partner organizations who are shaping what gets built. The invitation at the end is for them.

The persistent challenge: why the 3Ms still matter

In their landmark 2007 study published in The ANNALS of the American Academy of Political and Social Science, researchers Timothy Bates, William Jackson, and James Johnson identified three fundamental barriers that explain why minority and underserved entrepreneurs struggle to build viable businesses. They called it the 3Ms framework:

The framework built on decades of entrepreneurship research, but its power lay in its simplicity. Bates and colleagues argued that these three barriers are interconnected and mutually reinforcing. Without capital, entrepreneurs can't access markets. Without market access, they can't generate the revenue to attract capital. Without management knowledge, they struggle to navigate either. Of the three, they considered management — "the skilled and capable entrepreneur, or the management team" — the most critical.

Why these barriers persist

The business support ecosystem has grown substantially since the 3Ms framework was first articulated. America's Small Business Development Centers serve hundreds of thousands of entrepreneurs annually. Community Development Financial Institutions (CDFIs) have deployed billions in capital to underserved markets. Corporate supplier diversity programs have opened doors to new customers. Accelerators and incubators have proliferated in cities large and small.

Yet the data tells a sobering story. According to the Federal Reserve's Small Business Credit Survey (2024), Black-owned firms are approved for financing at roughly half the rate of white-owned firms, even when controlling for creditworthiness. Immigrant entrepreneurs still struggle to access mainstream banking. Rural small businesses still face geographic isolation from markets and services. Women entrepreneurs still receive a fraction of venture capital despite starting businesses at record rates.

The barriers persist for structural reasons that traditional interventions struggle to address:

The equity dimension

These barriers don't affect all entrepreneurs equally. As researchers Candida Brush, Anne de Bruin, and Friederike Welter (2009) noted in their extension of the 3Ms framework, entrepreneurs face additional contextual barriers (including household responsibilities and the broader policy environment) that compound the core challenges. The entrepreneurs with the fewest resources (those who could benefit most from support) are precisely the ones least likely to access it.

For her, the 3Ms aren't just abstract barriers: they're the difference between a viable business and another statistic. As Nobel laureate Amartya Sen argued, development is fundamentally about expanding human capabilities and freedoms, not just economic metrics (Sen, 1999). This framing applies directly to entrepreneurship: the goal isn't just business survival, but enabling entrepreneurs to pursue opportunities that matter to them.

As Morris, Kuratko, and colleagues describe in their research on the "liability of poorness," entrepreneurs in poverty face compounding disadvantages: gaps in financial, business, and technological literacy; a scarcity mindset driven by constant resource constraints; and non-business distractions like housing instability and transportation challenges that compete for attention (Morris et al., 2022). Recent research confirms these dynamics persist, showing how entrepreneurs can move from poverty traps into "commodity traps" where the venture is undifferentiated, low-margin, and capacity-constrained, despite initial success (Morris et al., 2025).

Agentic AI: a primer

Artificial intelligence has evolved rapidly over the past several years, but most applications remain relatively passive. Traditional AI tools wait for users to ask questions and provide answers — think of a search engine or a customer service chatbot.

Agentic AI represents something different. These systems can take autonomous action on behalf of users, remember context across interactions, and adapt their behavior based on each user's specific situation. Rather than simply answering questions, agentic AI can guide users through complex processes, proactively surface relevant information, and coordinate multiple steps toward a goal.

The distinction matters for business support. An entrepreneur doesn't just need answers: she needs guidance. She needs someone (or something) that understands her specific situation, remembers what she's working on, and can help her navigate from where she is to where she wants to be.

How agentic AI differs from generative AI

The distinction is fundamental: generative AI creates content in response to prompts; agentic AI takes autonomous action to achieve goals.

Why agentic AI suits business support

The characteristics that make agentic AI powerful are precisely what's been missing from scaled business support:

AI in adjacent fields

Business support isn't the first field to explore AI coaching. Mental health platforms like Woebot have demonstrated that AI can provide meaningful therapeutic support. Educational applications like Khan Academy's Khanmigo are personalizing learning at scale. Financial planning tools are helping consumers navigate complex decisions.

These applications share a common insight: AI works best not as a replacement for human expertise, but as a way to extend that expertise to people who would otherwise go without. Business support faces the same fundamental constraint: there aren't enough advisors to serve every entrepreneur who needs help. AI can fill the gap.

Addressing the 3Ms through agentic AI

How might agentic AI actually address the three fundamental barriers identified in the 3Ms framework? This section explores the mechanisms: both the traditional approaches and how AI might augment or transform them.

Access to management knowledge

The traditional model

Technical assistance programs have historically delivered management knowledge through workshops, one-on-one advising, and cohort-based programs. These approaches work for entrepreneurs who can access them. But the constraints are significant: programs have limited capacity, advisors have limited hours, and entrepreneurs have limited time to travel to where services are offered.

The agentic AI model

Agentic AI can deliver management knowledge on-demand, personalized to each entrepreneur's specific situation. Rather than attending a general workshop on cash flow management, an entrepreneur can ask questions about her cash flow: with an AI that understands her business model, her industry, and her specific challenges.

This isn't about replacing human advisors. It's about ensuring that every entrepreneur has access to baseline support, regardless of whether they can get on an advisor's calendar.

What Ena does

Ena gives every founder in a program an always-on concierge under the organization's own brand: voice or text, in any of 32 languages, at any hour, grounded in the organization's actual programs, partners, and knowledge base. The AI remembers previous conversations, building context over time — and every conversation flows into the organizational memory that briefs the founder's coach before the next session, so the humans walk in prepared rather than starting from scratch.

Early users have used Ena for everything from pricing strategy to handling difficult customer conversations to understanding business structures. Kelly Evans, VP of Entrepreneur and Economic Development at Chicago Urban League, has validated the approach with her network. Tony Wilkins, a well-established angel investor and startup figure in Chicago, has found value even as an experienced entrepreneur. And Dr. Alex DeNoble, Professor Emeritus at San Diego State University, praised Ena for "her" business acumen and high quality advice.

Access to markets

The traditional model

Connecting entrepreneurs to customers and business opportunities traditionally happens through networking events, supplier diversity programs, industry associations, and matchmaking initiatives. These approaches depend on getting the right people in the same room — physically or virtually — and hoping connections form.

The challenges are obvious. Networking events favor extroverts and those who already have professional networks. Supplier diversity programs work for businesses pursuing corporate contracts but not for those serving local markets. And none of these approaches scales: each connection requires coordination and follow-up.

The agentic AI model

Agentic AI can serve as an intelligent matchmaker within entrepreneurial ecosystems. Rather than waiting for entrepreneurs to find the right programs or connections, AI can proactively surface opportunities based on each entrepreneur's profile and needs.

More importantly, AI can connect entrepreneurs with each other. Many of the services small businesses need (accounting, legal, marketing, printing) are provided by other small businesses. AI can facilitate these B2B connections at scale.

What Ena does

Ena is built to route referrals intelligently. When an entrepreneur needs an accountant, Ena can connect them with small business accountants in the network. When they need legal help, Ena can route them to attorneys who serve small businesses. This creates a network effect: every entrepreneur who uses Ena becomes both a potential customer and a potential service provider for other entrepreneurs in the ecosystem.

The AI doesn't just connect entrepreneurs to programs: it connects them to each other.

Access to money (capital)

The traditional model

Helping entrepreneurs access capital has traditionally meant connecting them to CDFIs, preparing them for loan applications, and providing financial literacy education. These services are valuable but labor-intensive. Loan readiness programs might work with an entrepreneur for months before they're prepared to apply.

The challenges compound for entrepreneurs who lack banking relationships or have damaged credit. They may not know what capital products exist, which ones they might qualify for, or how to position their business for approval.

The agentic AI model

Agentic AI can provide ongoing financial coaching that meets entrepreneurs where they are. Rather than a one-time loan readiness program, AI can work with entrepreneurs continuously: explaining credit concepts, helping them understand their financials, and preparing them for capital conversations over time.

AI can also help entrepreneurs understand the landscape of capital options. Many underserved entrepreneurs don't know that CDFIs exist, or that they might qualify for grants and microloans. AI can surface options that entrepreneurs wouldn't otherwise discover.

What Ena does

Ena provides guidance on credit, capital readiness, and financial operations. When entrepreneurs have questions about improving their credit score, understanding their business financials, or preparing for a loan application, Ena can help. When they're ready to pursue capital, Ena can connect them with responsible lenders: banks and CDFIs that serve small businesses without predatory terms.

The key word is responsible. The AI connects entrepreneurs with vetted capital providers, not predatory lenders who target vulnerable businesses.

What we're still learning

Agentic AI for business support is young territory, and we'd rather name the open questions than pretend they're settled. These are the questions we're working through in live deployments — with real founders, real advisors, and the organizations that serve them:

The advantage of an operating layer is that the evidence builds itself. Every deployment shows us which topics founders actually raise, where they stall between sessions, and when they choose the AI versus asking for a human — evidence, not anecdotes. All of it flows back into how the platform gets built, and into the impact reporting our partner organizations owe their boards and funders.

An invitation to founding partners

When this paper was first published, its closing invitation was to research partners. The platform has moved past that stage. The invitation now is to the organizations the thesis was always about.

We're selecting a small cohort of founding partner organizations — accelerators and incubators, BSOs and small business centers, economic development organizations, chambers and associations, and the consulting firms and coaches who serve founders — to deploy Ena OS under their own brand and shape what it becomes.

What founding partners get

• A roadmap your program actually shapes — founding partners set build priorities
• A direct line to the founding team, not a support queue
• Founding-partner pricing, locked in as the platform grows
• A phased rollout that fits your cohort calendar, not ours

What we ask in return

Honest feedback, and a case study when the results are real. That's the exchange: you get a platform molded around how your program actually works, and the field gets documented evidence of what AI-supported entrepreneur support can do.

If that sounds like your organization, start the conversation — you'll hear back from the founding team, not a sales rep.

Conclusion: an opportunity to shape the future

The 3Ms framework, first articulated by Bates, Jackson, and Johnson in 2007, has guided our understanding of entrepreneurial barriers for nearly two decades. The barriers it identifies — Money, Markets, and Management — remain as relevant today as when the framework was first articulated. But our tools for addressing those barriers have largely failed to scale.

Agentic AI offers something genuinely new: the possibility of personalized, always-available business support that doesn't depend on advisor capacity, program funding, or geographic proximity. For the first time, we might be able to ensure that every entrepreneur (regardless of language, location, or resources) has access to quality guidance.

Ena OS is the working test of this possibility. The platform is live, real founders are using it, and what began as a coaching hotline has grown into an operating layer for the organizations that serve them. Early signals are promising — and we hold ourselves to reporting what the deployments actually show, not what we hope they show.

The business support sector stands at an inflection point. AI is advancing rapidly. The question isn't whether AI will transform how entrepreneurs receive support; it's whether the sector will shape that transformation or simply react to it.

We have an opportunity to be intentional about this. To build AI systems that genuinely serve underserved entrepreneurs rather than creating new disparities. To ensure that this technological shift advances equity rather than undermining it. And to keep the humans — the coaches, advisors, and program teams founders trust — at the center of it, better equipped than they've ever been.

That work happens with practitioners, not around them. It requires ESO leaders willing to test new approaches alongside proven ones, and to shape the tools rather than inherit them.

The invitation is open.

About the author

Jason William Johnson, Ph.D., is the founder and CEO of Ena Intelligence. His career has focused on the intersection of entrepreneurship support, technology, and economic equity. For more than 15 years he led entrepreneurship support programs — coaching founders, running cohorts, and managing teams of advisors — across local and multi-regional technical assistance initiatives serving underserved entrepreneurs. He holds a PhD in Organizational Leadership and brings over 16 years of experience in economic development. Reach him at jason@enaintelligence.com.