BAAF® Framework

The BAAF® Framework (Business AI Awareness Framework) is a methodological framework designed to guide the successful implementation of artificial intelligence in any business organization. It establishes the correct order of work: first understand the business model, then map the digital assets that support it, and on that foundation build situational awareness that perceives, comprehends, and anticipates through artificial intelligence agents. It is composed of three sequentially dependent abstraction levels: Business Foundation (L1), Digital Business Connectors (L2), and Situational Awareness (L3).

The BAAF® is applicable to B2B, B2C, and hybrid B2B2C organizations, regardless of their industry or size. Each level operates with independent pipelines, but there is a strategic alignment dependency between them: L2 pipelines must be aligned to the purpose defined in L1, and L3 pipelines depend on L2's integrity. AI agents operate at L3, and their goal is to be Coherent, Connected, and Aware. If the business operation does not align with the methodology defined in L1, agents will not be coherent with how the business should operate. If the digital connectors in L2 experience service losses or deficient integrations, the user experience degrades and agents lose their connection to the business. At L3, where AI agents are implemented, it is necessary to define the required maturity level (AI-Assisted, AI-Augmented, AI-Driven, or AI-Automated) and how the agent facilitates the business's situational awareness: what it perceives, what it comprehends, and what it anticipates. Without these definitions, agents lose situational awareness with the business. Once correctly implemented, all three levels operate together and continuously feed back into each other.

The BAAF® Framework simultaneously serves as:

• A diagnostic tool: by evaluating the state of each level, it reveals the gaps that prevent a successful AI implementation.
• A technical blueprint: the level map defines the architecture that the AI platform needs to operate aligned to the business.
• A reference framework for the implementation of artificial intelligence agents.

Most AI implementation initiatives fail, not only because of limitations in available technology, but because they are built on processes that no one has documented and on an incomplete understanding of the underlying business model.

Table 1: Common Errors in AI Implementations

The BAAF® Framework seeks to expose these gaps by establishing a logical and inevitable order: first understand the business (L1), then map and connect the digital assets (L2), and only then build situational awareness (L3).

The BAAF® Framework is composed of three levels. Each one fulfills a specific role within the framework, and their alignment determines whether AI agents will be coherent, connected, and aware.

Table 2: BAAF® Framework Levels

3.1 L1 — Business Foundation

The first level establishes the business model and the sales or marketing methodology that governs its operation. It is the starting point of the framework: every decision in L2 and L3 aligns to what is defined here.

B2C stands for Business-to-Consumer. It is the commercial model where a company sells products or services directly to the end consumer (individual people), rather than to other businesses. B2B stands for Business-to-Business. It is the commercial model where a company sells products or services to other companies or organizations, rather than to the end consumer. B2B2C stands for Business-to-Business-to-Consumer. It is a hybrid model where a company sells to another company, which in turn sells to the end consumer.

Table 3: Everyday B2C Examples

Table 4: Everyday B2B Examples

Table 5: B2C vs B2B — Key Differences

Table 6: Business Models and Applicable Methodologies

Understanding the difference between B2B and B2C is fundamental because it determines how AI agents are configured to facilitate the business's situational awareness. In B2B, the agent needs to understand an opportunity qualification process with multiple decision-makers and stages defined by methodologies like MEDDIC or MEDDPICC. In B2C, the agent must respond to consumer behavior: what attracts them, what problem they are trying to solve, what their value as a customer is, and where they are in the decision journey. In B2B2C models, both flows coexist and agents must identify which context they are operating in.

3.2 L2 — Digital Business Connectors

The second level identifies, maps, and connects all digital assets of the business. This is the most critical step of the framework: the digital connectors in L2 are the ones that provide the data necessary for L3 to perceive, comprehend, and anticipate.

Digital connectors include:

• Internal and external APIs
• Operational and analytical databases
• Communication channels: WhatsApp, Email, Voice (Amazon Connect)
• Forms and data capture points
• CRM, ERP, and management systems
• Third-party services: payments, logistics, healthcare, etc.
• Real-time events and webhooks
• MCP links (Model Context Protocol): a standard protocol that enables AI agents to connect directly with tools, services, and external data sources

Table 7: Examples of Digital Connectors by Business Type

The L2 analysis produces the business digital asset map — a deliverable that is a prerequisite for any AI agent implementation. This map enables the AI Deployment Team — developers, product owner, and commercial team — to have clarity on the available digital assets, the scope of agent deployment, and the level of situational awareness required. At AI-Driven and AI-Automated maturity levels, agents must also have operational visibility over connectors: what data is available, what channels they can use, and how to respond to the loss of availability of a digital connector.

3.3 L3 — Situational Awareness

The third level implements the concept of Situational Awareness, originally taken from aviation and formalized by Mica Endsley.

The concept of Situational Awareness was developed in the context of military aviation, where pilots must make critical decisions in real time with changing information and multiple variables. In a business environment, AI agents face an analogous challenge: they must interpret business signals, comprehend their meaning in context, and anticipate the next state to act or recommend actions before it is too late.

Table 8: Situational Awareness Levels (Endsley)

Table 9: Examples of Situational Awareness Applied to Business

Relationship Between AI Maturity Level and Situational Awareness

Not all agents need the same scope of situational awareness. The chosen AI maturity level determines how far the agent goes in each SA dimension:

Table 10: Situational Awareness Scope by Maturity Level

Choosing the right maturity level is not just a technical decision. Section 4 elaborates on what state the organization must have at each framework level to support these maturity levels and what it implies in terms of risk tolerance.

The Role of Prompt Engineering in Situational Awareness

Prompt engineering is the mechanism through which the agent's situational awareness is configured. Through the prompt, it is defined what data the agent must perceive, what business rules it applies to comprehend context, and under what conditions it must anticipate and act. A prompt misaligned with business logic (L1) or available data (L2) produces an agent that operates without real situational awareness.

The BAAF® Framework includes an AI maturity axis as a diagnostic tool. This axis enables the AI Deployment Team — developers, product owner, and commercial team — to evaluate the current state of each framework level and determine which agent level is viable for the organization.

Table 11: AI Maturity Axis — Required State by Level

Business Foundation State by Maturity Level

• Defined: The business model and methodology are documented but not necessarily executed consistently.
• Executed: The methodology is applied consistently in business operations.
• Optimized: The methodology is continuously reviewed and adjusted based on results.
• AI-optimized: AI itself suggests methodology adjustments based on data.

Digital Connectors State by Maturity Level

• Partial: Some connectors are integrated, others operate in isolation.
• Connected: All critical connectors are integrated and operate with stability.
• Real-time: Connectors transmit data in real time with active monitoring.
• Self-healing: Connectors detect and recover from failures automatically without human intervention.

Situational Awareness State by Maturity Level

• Reactive: The agent responds only when requested.
• Dashboards: The agent presents organized information for the human to make decisions.
• Predictive: The agent anticipates situations and recommends actions.
• Autonomous: The agent perceives, comprehends, anticipates, and executes without intervention.

An organization does not need to reach AI-Automated. The ideal level depends on its business model, resources, and risk tolerance. Each maturity level implies greater agent autonomy and less human intervention. The greater the autonomy, the greater the exposure to unsupervised errors: incomplete data in L2, gaps in business logic in L1, or scenarios not covered in the agent's prompt. Risk tolerance defines how much autonomy the organization is willing to accept, considering that no level of the framework is free from gaps.

Table 12: Examples of Risk Tolerance by Maturity Level

The BAAF® Framework is not a static bottom-up flow model. The levels feed back into each other continuously, creating an improvement cycle that evolves with the business. This mechanism is what keeps coherence, connection, and awareness aligned over time: what L3 detects can reveal gaps in L1 or L2, and changes in L1 or L2 can expand or limit the scope of situational awareness in L3.

Table 13: Feedback L3 → L1 → L2

Table 14: Other Feedback Directions

The framework is industry-agnostic. The section below illustrates how the AI Deployment Team would apply the framework in three typical scenarios, including the recommended AI maturity level for each case:

Table 15: Scenario 1 — Dental Clinic (B2C)

Table 16: Scenario 2 — Vehicle Dealership (B2B + B2C)

Table 17: Scenario 3 — Enterprise Contact Center (B2B)