Get started

Guidance AI Systems Scoping & Pricing

Download

Guidance AI Systems » Guidance AI Systems Scoping & Pricing

AI Guidance AI Systems Development Process

Airlock’s Guidance AI Systems are expertly crafted to provide on-demand coaching and actionable insights, enhancing sales performance and customer engagement. These systems are designed to understand and integrate your business’s products, processes, and sales strategies, offering personalized support and immediate feedback. This enables a more efficient, real-time learning and sales environment, effectively transforming traditional sales and customer service into a more dynamic, interactive experience.

You can anticipate a one-time development fee, followed by a monthly recurring charge for monitoring and optimization. The final pricing will hinge on the specific scope of your project. Factors influencing this include the complexity of your AI, how many resources it needs to do the job you want, the number of integrations needed (such as CRM linkage), the desired sophistication of the AI’s abilities, and the frequency of AI knowledge updates.

Development fees range from $3997-$12,997 or higher, with monthly recurring charges of $397-$1297 or higher.

That’s a broad range! Most businesses can expect development of $3997-$5997 and monthly charges of $497-$797.

  • After an initial conversation with you we’ll build a prototype for free. We want you to be able to envision outcomes.
  • You’re in charge of scope and you’ll know what your investments will be before we get busy building your AI. No surprises.

Keep reading for how the entire process works, including what impacts pricing.

Conversational AI Agent development consists of several stages:

  • Pre-Production
    • Prototype
  • Production
    • Strategy and Scope
    • Design
    • Training
    • Development
    • Testing & Quality Assurance
    • Iterate & Adjust
    • Deployment & Integration
  • Post-Production
    • Monitoring and optimization

Pre-Production

Protype

Where we will, after an initial conversation with you, quickly create a visual flow and functional (though limited) AI to demonstrate basic capability, and allow us to talk through how things will work. The prototype is designed to let you make an informed decision about moving forward.

Required from you:

  • An initial conversation of around an hour

Deliverable: Functional AI Prototype

Timeline: 1-2 days

Investment: No Cost

Production

Strategy and scope

Where we identify what we want your Guidance System to do, when, why, and what are ideal outcomes.

  • Our guided process makes this easy, helping us quickly move from “what if we…” to “it will then…”
  • Often accomplishing the strategy involves integrating with additional tools, applications or platforms. Sending transcripts from the Guidance System to a training manager, for instance, or integrating with HR systems for employee onboarding.
  • We’ll define how broad your Guidance System’s understanding and capabilities should be. This will influence how much training it needs on your business. A few website pages and PDFS is often all it needs. For more granular, specific abilities, AI will need detailed, numerous resources.
  • The broader the strategy and scope, the higher the investment. The more specific the abilities required of AI, the higher the investment.
  • Where appropriate, we can suggest a phased approach to development.

Required from you:

  • Collaboration in generally one meeting of around an hour
  • Email collaboration as we share development
  • Sign-off

Deliverable: Scoping Document

Timeline: 3-5 days

Design

Where we build the logic and conversation flow of if the user does this, the AI should do that.

  • With strategy and scope understood, we can quickly map out a non-functional or minimally-functional logic flow that, together, we can visualize and understand.
  • The more complexity required of the logic flow, the higher the investment
  • At the end of this flow mapping, scope is set
  • 50% of initial invoice paid begin development

Required from you:

  • Collaboration in generally one meeting of around an hour
  • Email collaboration as we share development
  • Sign-off
  • 50% invoice payment

Deliverable: Visual Logic Flow

Timeline: 1-2 days

Training

Where we “teach” the AI your materials, brand, conversational tone and more.

  • The AI is only as good as what it knows. We begin by collecting all the necessary information that will populate the AI’s Knowledge Base. This includes FAQs, product details, pricing, company information, and any other relevant content that will help us satisfy the strategy. We train the AI with the collected data and knowledge base to ensure it understands user queries and provides accurate responses.
  • Beyond informational data, we’ll also train on the role it should play. If your Guidance System is a sales coach, for instance, we train the AI to know your sales process. If the AI will engage in role play, for instance as a customer, we’ll train the AI on common customer profiles.
  • Our easy submission process lets you simply upload what’s necessary as a result of our Strategic decisions together.
  • The more resources required in the AI’s Knowledge Base, the higher the investment.

Required from you:

  • Resources to train your AI, like website urls, product or services pdfs, brand guides, etc.
  • What processes or references the AI will require to act in its capacity to satisfy strategy.

  • Which resources we need will be established according to Strategy and Design steps

Deliverable: Internal Only

Timeline: 5-7 days (depending on your ability to submit resources)

Development

Where we combine Design and Training for a quality user experience with the AI behaving as expected.

  • AI technology is implemented, integrated, and configured to align with your specific requirements as identified in Strategy and mapped out in Design. This includes setting up natural language processing (NLP), machine learning models, and other AI components.
  • We’ll align AI behaviors to the logic flow we’ve agreed upon. For instance, “Ask two questions, then provide insight, then ask a followup question.”
  • The more complexity required of development, such as integrations or a large Knowledge Base, the higher the investment

Required from you:

  • Login or API credentials for any integrations
  • Patience and support as we build your Guidance AI System
  • Up to 2 check-ins for feedback as necessary
  • Sign-off

Deliverable: Internal Only

Timeline: 2-3 weeks (sometimes longer for very large projects)

Testing and Quality Assurance

Where together we identify and rectify any knowledge or functional issues.

  • Any outside integrations will be tested at this time too.

Required from you:

  • Patience and support as we build your Guidance AI System
  • A meeting of around 30 minutes as we share your Guidance AI System
  • Resources like time and people to test your Guidance AI System

Deliverable: Guidance AI System ready for testing

Timeline: 3-5 days (sometimes longer if integrations are complex)

Iterate and Adjust (Up to two rounds)

Based on our combined experiences and feedback, we iterate the Guidance AI System for improved performance, up to two rounds.

  • Any required additions to the Knowledge Base to accommodate improved performance are implemented
  • Iterations are performed in-scope. If any feedback requires new scope, we’ll be happy to discuss.

Required from you:

  • Patience and support as we iterate your Guidance AI System
  • Email collaboration as we share development
  • Sign-off

Deliverable: Guidance AI System ready for deployment

Timeline: 3-5 days (sometimes longer if iteration is complex or scope shifts)

Deployment and Integration

Where we prepare the AI to live on your website or other digital destination

  • Typically the Guidance AI System will live on a website. If you have bespoke needs, like on an intranet or inside an app, we’ll establish this understanding early in the Strategy discussion.
  • Launch and monitoring, where we closely monitor its performance during initial stages.

Required from you:

  • Admin credentials to your website
  • Final 50% of invoice paid

Deliverable: Guidance AI System live on your site

Timeline: 1-2 days

Overall Timeline: 3-6 weeks (sometimes longer, depending on complexity)
Overall Investment to go from ideation to installation:
  • Starting at $3997, up to $12,997 or higher
  • Complexity, as indicated in all earlier steps, impacts pricing
  • Most businesses are between $3997 and $5997. We’ll establish this together early in the process. No surprises.

Post-Production

Monitoring and Optimization

Where the AI is continually fine-tuned to learn from real-world experiences, new data, technology or language model updates and more. With our ongoing support, your AI remains efficient, effective, and future-ready.

  • Conversation transcripts are reviewed each day, each week or each month to identify new opportunities for training
  • New training is conducted on the AI
  • Analytics that suggest an opportunity outside of AI is shared accordingly with you. For instance, “We see that a lot of people are struggling with this part of your process – maybe an opportunity there for you.”
  • The more frequent the review cadence and more complex the AI, the higher the monthly investment

Timeline: Daily, Weekly or Monthly

Monthly Investment:
  • Starting at $397, up to $1297 or higher
  • Complexity and cadence, as indicated, impact pricing
  • Most businesses are between $497 and $797/month. We’ll establish this together early in the process. No surprises.
  • No long-term commitments. You can stop anytime with 60 days notice.
  • Preferential pricing for semi-annual or annual payments.

Questions? Answers.

How does the Guidance AI System personalize its interactions with different users? Does it adapt based on user behavior, location, or other factors?

Every interaction is unique. If Tony is engaging with your Guidance System at the exact same time as Steve is, they’re both having unique, original experiences. Personalization can be straight forward. If the AI obtains a name, for instance, it’ll remember and use it throughout the chat. Personalization can also be achieved through thoughtful questions where the AI picks up on individual needs and issues. While the AI is trained to “speak for itself,” we can be very specific on what it says and when. At this point, ask a question about this thing. Then share an insight. Then ask another question. And so on. Throughout this process, the AI is adept at taking the conversation’s essence, mulling it over with what it’s learned from its training materials, and then delivering a fitting response.

How does the Guidance AI System integrate with existing systems, like HR systems or spreadsheets? Is there a seamless handoff if a human needs to intervene?

Your AI can seamlessly integrate with a broad array of processes, including many HR, CRM and sales systems. Beyond that, stakeholders can be alerted of new conversational activities through email or text, and we can even append these details to a Google doc for effortless export.

As conversations unfold in real-time, our AI is skilled at discerning user intent and key events. This enables your reports to highlight elements the AI identifies as urgent needs, specific interests, potential timelines, or any other pivotal aspects integral to your sales journey.

Data & Privacy: How do you ensure data privacy and GDPR compliance? Does the AI store conversations or personal details of the visitors?

At present, our services are offered exclusively within the United States. The platforms and systems we utilize, including but not limited to ChatGPT, Claude, Amazon Web Services (AWS), and Google Cloud Platform (GCP), each have distinct privacy and security policies. We ensure these policies are made available to you before formalizing any engagement.

It’s important to highlight that:

  • Conversations facilitated by our system are not stored or repurposed for training beyond your specific use case and conversational system. We do not directly use any data for further training of any LLM models in general.
  • Both encryption-at-rest and encryption-in-transit are rigorously implemented.
    However, it’s crucial to understand that we cannot be held liable for potential data or security breaches that might occur due to vulnerabilities in our software vendors’ systems. Additionally, any communications produced by the AI, based on the materials you’ve provided, are not under our direct liability.
Can the AI operate in multiple languages?

Presently our capabilities are limited to US English.

Culture & Tone Adherence: How is the AI programmed to truly reflect a company's culture and communication style? What's the process to fine-tune this?

Just like with our knowledge base, any materials you provide on company culture will help train your AI. Any materials you possess that shed light on your organizational ethos can be used to fine-tune your AI. At its core, the AI is trained to resonate with your target audience. This could involve understanding their demographics, professions, or any other nuances central to your sales engagements. Guidance on tone like “casual yet professional” is taken to heart, and if you have specific jargon, acronyms, or unique brand expressions, those are integrated seamlessly. If you have a brand style or brand voice guide, we’ll implement it. Additionally, we offer a dedicated ‘Brand Style’ service to fine-tune this communicative approach further.

Should the AI face questions outside its primary function, it’s trained to not make guesses or venture beyond its knowledge base. It’s trained to not speculate, opine or engage. Its training is for responses to stay strictly within the boundaries of its training. Hallucination is a term to describe AI’s tendency to sometimes make up information or stray from its training. We have safeguards in place to, if not eliminate this phenomenon, minimize it significantly.

Before your Guidance System is introduced, it undergoes testing, with your team playing a crucial role, to ensure it communicates in a manner consistent with your standards and strategies.

Feedback Loop: If the AI doesn't know an answer to a question, can it be trained to know it next time?

Yes. When the AI encounters a question it can’t answer, our system is primed for assessment and improvement. This often involves refining the AI’s prompt configurations, supplementing its training materials, or reimagining parts of its conversational strategy. This is part of the crucial value of ongoing Monitoring and Optimization.

Customization & Branding: How customizable is the AI's interface? Can we adjust its appearance to match our branding?

At this time, icons and colors can be customized. Structural aspects like dimensions or placement on the page aren’t customizable – for now, the AI conversational system will “live” inside a button on the lower right of a website page.

Continuous Improvement: How often is the AI updated or improved? Do you provide regular updates based on new findings or technologies?

Generative AI platforms, including ChatGPT, and the associated systems we utilize undergo regular updates, although the timeline is beyond our direct control. Rest assured, when significant updates arise that might influence the design or efficiency of your AI conversational system, we’ll promptly inform you and offer relevant recommendations or guidance.

Training & Setup: How much involvement is required from our end during the initial setup and training phase? Will we need to provide specific scripts or conversation flows, or is it more hands-off?

We have a simple process to retrieve essential knowledge assets from you when we start building your Guidance AI System, which are the materials we’ll use to train and teach your AI about who you are and what you do. Beyond that, you’ll have checkpoints to review and test your AI and make suggestions and refinements. But, you’re never responsible for holding our hands or micro-managing. The entire development process is fun, interesting and easy.

ROI Measurement: Do you provide any analytics or reporting tools to help track conversions, lead quality, and other relevant metrics?

Analytics we capture include:

  • Interactions: Total number of engagements users have had with your AI.
  • Recognition rate: The percentage of messages understood by your AI, i.e. did not trigger the fallback intent.
  • Unique Users: Unique users that have engaged with your AI.
  • Unique Sessions: Unique user sessions with your AI.
  • Top Intents: The most popular queries users ask your AI.

You can also compare different time periods to see how your AI is performing with each iteration of your assistant. You can then cross reference these analytics with your internal ROI measurement tools to determine how it’s working. We also regularly use this data to improve your AI.

An important feature of Guidance AI Systems is the ability to share AI conversations for human intervention. For instance, Tony’s sales role-play scenario with the AI can be sent for Natasha, his manager, to review. Natasha can then work directly with Tony on issues she identifies from his experience with the AI. Additionally, we’ll keep a human-in-the-loop on the process, alerting Natasha if we see trends that she might want to address more generally.

What might I expect to pay a dedicated lead-setter or support agent to monitor my website and answer questions in real-time?

For hiring a dedicated sales coach or an employee onboarding agent, you can generally expect to pay between $4,000 to $6,000 per month. This cost covers the services of an experienced professional focused on enhancing sales skills or efficiently managing the onboarding process for new employees, ensuring they are well-prepared and integrated into your team.

Limitations: Are there any limitations or scenarios where the AI might not be the best fit for lead generation? For instance, industries or target audiences that may not respond well to AI interactions?

While scoping your project, if we identify areas where AI might be problematic—especially concerning sensitive topics—we’ll recommend a human-centric approach. Though our platforms adhere to HIPAA and SOC-2 compliance standards, there are niches, such as within healthcare or finance, where human interaction is irreplaceable. That said, the value of an AI Expert is undeniable in most scenarios. For example, while we won’t design a medical AI for diagnostics, we can craft one to help pharmaceutical representatives fine-tune their sales approach. Simply – we don’t think AI is yet an appropriate tool for medical, financial, or sensitive advice or processing. We won’t take on projects where we lack confidence in our ability to have a positive impact on you or your users.