Strategic Research Brief

Helium AI: Product Expansion
for Capital Raise Positioning

Deep research into Helium Business (team collaboration layer) and Helium Bees (autonomous agent framework), including market data, competitive landscape, and differentiation strategy.

Prepared: March 14, 2026 For: Neural Arc / he2.ai Classification: Internal

Two Products, One Strategic Moat

Helium AI is expanding from a multi-model AI platform for individuals into a team-native, agent-capable system. This brief covers two new product layers: Helium Business (shared knowledge, team prompts, collaborative AI workspace) and Helium Bees (autonomous agents with proactive and reactive capabilities under governed guardrails). Both products address two of the most urgent pain points in enterprise AI adoption: shadow AI data leakage and the lack of production-grade autonomous agent infrastructure.

Core Thesis for Investors

Helium AI solves the "shadow AI problem" (data leaking through unauthorized tools) and the "agent gap" (demand for autonomous execution without enterprise-grade guardrails) in a single, unified, multi-model platform. This dual-layer offering addresses two separate budget lines (team productivity + automation) and positions Helium for a significantly higher ACV than single-model competitors.

Shadow AI: The $4.6M Risk Enterprises Cannot Ignore

The data is clear. Employees are adopting AI tools faster than organizations can track, approve, or secure them. The result is a massive, uncontrolled data exfiltration surface that puts intellectual property, compliance, and competitive advantage at risk.

49%
of workers admit to using AI tools without employer approval
BlackFog 2026
77%
of employees have pasted company data into AI tools
LayerX 2025
$4.63M
average breach cost for organizations with high shadow AI usage
IBM 2025
82%
of employees using AI do so through personal accounts, not enterprise tools
LayerX 2025
86%
of workers now use AI tools at least weekly for work tasks
BlackFog 2026
68%
of employees use unauthorized AI tools at work
Gartner 2026
The Banning Paradox

Research consistently shows that roughly half of employees would continue using personal AI accounts even after an organizational ban. Prohibition drives shadow AI deeper underground rather than eliminating it. The only viable solution is to provide a secure, enterprise-grade alternative that matches the functionality employees already find on their own. This is the exact gap Helium Business fills.

Why Existing Solutions Fall Short

Most organizations face a fragmented landscape: employees use ChatGPT for writing, Claude for analysis, Gemini for search, Midjourney for images, and specialized tools for coding. Each tool requires a separate subscription, stores data independently, and operates outside IT visibility. A BlackFog survey found that 34% of employees use free versions of company-approved AI tools, raising serious questions about where sensitive corporate data is stored and processed.

The 63% of employees who believe it is acceptable to use AI tools without IT oversight do so because no company-approved option exists that offers comparable functionality. Enterprise solutions from Microsoft Copilot and Google Gemini for Workspace are locked to single ecosystems and single models. Multi-model aggregators like Poe and TypingMind serve individuals well but lack the team governance, shared knowledge, and compliance features that enterprise buyers require.

Helium Business: Team AI with a Shared Brain

Helium Business introduces team collaboration to the multi-model AI platform. With a minimum of five team member licenses at $80/user/month, organizations gain a shared knowledge base, shared favorite prompts, and collaborative AI workflows in a single secure environment. Individual users get knowledge base access starting at the Startup plan ($15/month with 500 MB), ensuring the knowledge layer is available across the entire pricing ladder.

Competitive Landscape: Multi-Model Team Platforms

Platform Multi-Model Team Workspace Shared Knowledge Base Shared Prompts Price Range
Helium Teams Yes (all major LLMs) Yes Yes Yes $80/user/mo (min. 5)
Poe (Quora) Yes Enterprise tier only No Bot sharing only $19.99/mo individual
TeamAI Yes Yes Partial Yes ~$25+/user/mo
Aymo AI Yes (45+ models) Yes No Partial From $4/mo
TypingMind Yes (BYOK) Team edition No Prompt library $249/mo (5 users)
MS Copilot No (GPT only) Yes SharePoint-integrated No $30/user/mo
ChatGPT Team No (GPT only) Yes No Custom GPTs $25/user/mo
Differentiation for Investors

No competitor offers the combination of multi-model access, a shared team knowledge base, and shared prompt libraries in a single product. Single-model vendors (OpenAI, Anthropic, Google) lock teams into one ecosystem. Multi-model aggregators (Poe, TypingMind) lack team governance and shared knowledge. At $80/user/month for the Teams plan, Helium commands premium pricing justified by unlimited prompts, full multimodal capabilities, and team-native features that no aggregator offers. The $15 Startup and $35 Pro plans create a smooth individual-to-team upgrade funnel.

Key Features to Emphasize in the Raise

Shared Knowledge Base

Teams upload documents, SOPs, and proprietary data into a centralized AI-accessible knowledge layer. Every team member's AI queries are grounded in company-specific context. This keeps data inside the organization, not scattered across personal ChatGPT accounts.

Shared Favorite Prompts

Standardize AI usage across the team. When one person crafts a high-performing prompt for a sales email, a QA checklist, or a code review, the entire team benefits. This creates institutional AI knowledge that compounds over time.

Model Flexibility

Different tasks need different models. Use Claude for long-form analysis. Use GPT for creative writing. Use Gemini for search-grounded research. Teams pick the best model per task rather than settling for one vendor.

Data Containment

All AI interactions flow through Helium's secured environment. Enterprises gain audit trails, usage analytics, and data governance controls. This directly addresses the shadow AI problem and reduces the $4.63M average breach risk tied to uncontrolled AI usage.

Market Sizing Signal

AI-powered collaboration tools are projected to boost team output by an average of 37%. At the same time, 72% of teams now use AI technologies for daily teamwork activities. The demand side is enormous. The supply side (secure, multi-model, team-native AI platforms) remains thin. Companies with strong AI governance save $287K annually compared to those suffering shadow AI chaos.

Helium Bees: Autonomous Agents with Guardrails

Helium Bees introduces autonomous AI agents that operate with both proactive and reactive capabilities, guided by user-defined guardrails. This positions Helium AI at the forefront of the agentic AI wave that defines 2026.

The Agentic AI Moment

In 2026, AI agents are projected to be embedded in 80% of enterprise workplace applications. The industry has shifted from reactive chatbots to proactive, goal-oriented agents that plan, execute, observe, and adapt. IDC, Gartner, and Deloitte all confirm this transition is underway at enterprise scale.

80%
of enterprise apps expected to embed AI agents by 2026
IDC 2026
40%
of enterprise apps will feature task-specific AI agents, up from 5% in 2025
Gartner 2026
46%+
CAGR growth in agentic AI adoption across enterprises
Industry Reports 2026

Proactive vs. Reactive: The Helium Bees Framework

Most existing AI tools are reactive. You ask, they answer. Helium Bees changes this dynamic by introducing agents that operate on two modes.

Reactive Mode

Agents respond to triggers, requests, and events. A team member asks a question. An email arrives. A deadline passes. The agent processes the input and delivers a contextual response using the team's shared knowledge base.

Proactive Mode

Agents monitor conditions and act before a human prompts them. Identify a sales lead going cold. Flag a compliance gap. Surface a knowledge base article that answers a recurring team question. Anticipate needs based on patterns, not commands.

Guardrails: The Non-Negotiable Layer

The concept you described as "guide servants" maps to what the industry calls guardrails. This is the governance architecture that makes autonomous agents safe for production. The research is unambiguous: agents without guardrails are risk amplifiers, not productivity tools. Leading enterprises demand "bounded autonomy" with clear operational limits, escalation paths, and audit trails.

Critical Investor Narrative

72% of enterprises have introduced multi-agent systems, but nearly half have experienced a Severity 1 incident involving hallucinated data or unauthorized autonomous behavior. Only 14% of organizations have production-ready agentic AI solutions. The gap between demand and safe deployment is where Helium Bees positions itself. Your guardrail-first approach is not a limitation to pitch around. It is the primary selling point.

Helium Bees Architecture Principles

  1. Bounded Autonomy: Every agent operates within user-defined constraints. Operational limits, spending caps, and scope restrictions are configured before deployment, not bolted on after incidents.
  2. Human-on-the-Loop: Agents execute independently within boundaries. Humans supervise outcomes, audit behavior, and intervene only when anomalies arise. This balances speed with control, the exact model enterprises demand in 2026.
  3. Escalation Paths: When an agent encounters an edge case or exceeds its confidence threshold, it escalates to a human team member. No unchecked decisions. No runaway execution loops.
  4. Audit Trails: Every agent action is logged. Every decision is traceable. This satisfies compliance requirements (SOC 2, GDPR, the EU AI Act enforced August 2026) and builds the institutional trust needed for expansion.
  5. Multi-Model Reasoning: Bees agents are not locked to a single LLM. They route tasks to the optimal model. This is a structural advantage over Salesforce Agentforce (Salesforce-locked), Microsoft Copilot agents (M365-locked), and Oracle AI agents (Fusion Cloud-locked).
  6. Shared Knowledge Grounding: Bees agents pull from the same shared knowledge base as Helium Business. This means agents produce answers and take actions grounded in your company's proprietary data, not generic training data.

Competitive Positioning: Agent Platforms

Platform Autonomy Level Multi-Model Guardrails Target Market Lock-in
Helium Bees Proactive + Reactive Yes Built-in, user-defined SMB to Mid-Market None
Salesforce Agentforce 3.0 Full autonomous No Self-healing workflows Enterprise (CRM) Salesforce ecosystem
MS Copilot Agents Background autonomous No Entra ID security Enterprise (M365) Microsoft ecosystem
Atlassian Rovo In-workflow agents No Project permissions Enterprise (Dev/PM) Atlassian ecosystem
Beam AI Self-learning Partial SOP-driven Enterprise Low
Zapier Central Task-based Partial Workflow rules SMB Low
Structural Advantage

Every major agent platform in 2026 is locked to a single ecosystem: Salesforce, Microsoft, Atlassian, Oracle. Helium Bees is model-agnostic and ecosystem-independent. For the SMB and mid-market segments that do not want to commit to a single vendor's stack, this is a decisive differentiator.

The Full Pricing Ladder

Helium AI uses a tiered pricing model designed to capture users at zero cost and graduate them through increasing value. The architecture creates a natural upgrade path from individual exploration to team deployment to autonomous agent adoption.

Individual Plans

Entry Point
Free
$0
100 lifetime prompts. No monthly renewal. Account pauses after limit.
  • 100 prompts total (lifetime)
  • Access to core AI models
  • Basic interface
  • No knowledge base
  • No integrations
For Individuals
Startup
$15 /month
500 prompts per month (~75 per week). Knowledge base access with 500 MB storage.
  • 500 prompts/month
  • Knowledge base (500 MB)
  • Google integration
  • Community support

Power User + Team + Agent Plans

Power User
Max
$100 /month
Everything unlimited. 5 GB storage. Full platform access for high-volume individual users.
  • Unlimited prompts
  • Unlimited image generation
  • Unlimited video generation
  • Unlimited slides
  • 5 GB AIM storage
  • All integrations
  • Priority support
Autonomous Agents
Helium Bees
$99 /user/month
Proactive and reactive AI agents with guardrails. Add-on for teams ready for automation.
  • Autonomous agent deployment
  • Proactive + reactive modes
  • User-defined guardrails
  • Human-on-the-loop oversight
  • Full audit trails
  • Knowledge base grounding

Pricing Strategy Analysis

The Upgrade Funnel

The free tier (100 lifetime prompts) functions as a trial, not a freemium product. It creates urgency. Users who hit the limit have already demonstrated intent and workflow dependency. The Startup plan at $15/month captures budget-conscious individual users. The Pro plan at $35/month is the core individual revenue driver, offering multimodal capabilities (images, video, slides, brand automation) that justify a 2.3x price jump over Startup. The Max at $100/month captures power users before they need team features.

Pricing Risk: Teams at $80 vs. Max at $100

The Teams plan at $80/user/month is $20 less than Max per user, despite including additional team features (shared knowledge base, shared prompts, admin controls). This inversion is strategic: the minimum 5-seat requirement means a $400/month floor, producing higher total contract value than a single Max subscription. Investors will ask why team users pay less per seat than individual Max users. The answer is volume commitment. Five guaranteed seats at $80 produce $400/month versus one seat at $100. The per-user discount rewards commitment, and the 5-seat minimum ensures the economics work. Consider clearly framing Teams as "Max features + team features at a volume discount" in marketing materials.

Helium Bees at $99: The Expansion Play

Bees at $99/user/month is priced as a premium add-on layered on top of Teams. For a 5-person team, full Helium deployment (Teams + Bees) would total $179/user/month, or $895/month for the team. This positions Helium in premium territory, justified by the combination of multi-model access, team governance, and autonomous agent capabilities. No competitor offers all three in a single platform. The $99 price point for Bees also signals seriousness to enterprise buyers. Agent platforms from Salesforce and Microsoft carry similar or higher per-user costs but lock customers into a single ecosystem.

Competitive Price Comparison

Product Price Prompts Multi-Model Images/Video Team Features Agents
Helium Pro $35/mo Unlimited Yes Yes No No
ChatGPT Plus $20/mo Usage caps No Yes (DALL-E) No No
Claude Pro $20/mo Usage caps No No No No
Poe Premium $19.99/mo Points-based Yes Yes No No
Helium Teams $80/user/mo Unlimited Yes Yes Yes No
ChatGPT Team $25/user/mo Higher caps No Yes (DALL-E) Basic No
MS Copilot $30/user/mo N/A No Limited Yes Add-on
Helium Bees $99/user/mo Agent tasks Yes N/A Integrated Yes

Revenue Scenario Modeling

$400
Minimum monthly contract for Teams (5 seats × $80)
Floor revenue per team
$895
Full stack per 5-person team (Teams $400 + Bees $495)
Max ARPU per team
$10,740
Annual contract value for a 5-person team on Teams + Bees
Full stack ACV

The Flywheel: How Business + Bees Compound

Helium Business and Helium Bees are not independent products. They form a strategic flywheel that increases retention, expands ARPU, and deepens the competitive moat over time.

The Flywheel Logic

Teams adopt Helium Business for multi-model AI access and shared knowledge. As usage grows, the knowledge base becomes richer with company-specific data. The richer the knowledge base, the more effective Helium Bees agents become, because they are grounded in proprietary context. The more effective the agents, the more workflows get automated. The more workflows get automated, the more indispensable Helium becomes. This creates compounding switching costs that no single-model competitor or standalone agent tool replicates.

Revenue Architecture

Individual Ladder (Free → $15 → $35 → $100)

The free tier creates trial users. Startup at $15 captures early adopters. Pro at $35 is the individual revenue engine with multimodal capabilities. Max at $100 serves power users. Each tier adds clear, tangible value that justifies the price jump.

Team + Agent Layer ($80 + $99/user)

Teams at $80/user (min. 5 seats, $400/mo floor) is the B2B land. Bees at $99/user is the B2B expand. A 5-person team on the full stack generates $10,740 ACV. This dual-layer pricing creates two distinct revenue lines from the same customer.

Positioning Language for the Raise

  1. "The Anti-Shadow-AI Platform." Helium eliminates data leakage by giving teams a secure, multi-model workspace that is better than the free tools employees are already using. If they prefer ChatGPT for writing and Claude for analysis, they get both. Inside one secure environment.
  2. "Agents that know your business." Unlike generic agent platforms, Helium Bees agents operate on top of a shared knowledge base built by your team. They do not hallucinate your company's processes. They reference your actual SOPs, product docs, and historical decisions.
  3. "Multi-model by design, not by accident." Single-vendor AI platforms (Copilot, Agentforce, Rovo) optimize for their own ecosystem. Helium optimizes for your team's output, regardless of which model performs best for any given task.
  4. "Guardrails are our feature, not our constraint." With 72% of enterprises reporting agent-related incidents, the market does not need more autonomous AI. It needs safe autonomous AI. Helium Bees ships with bounded autonomy, escalation paths, and full audit trails from day one.

Strategic Recommendations for the Raise

Product Priorities

  1. Ship Helium Business first. The shared knowledge base and shared prompts are the foundation for Bees. They also represent the lower-risk, easier-to-sell product. Land teams on Business, then upsell to Bees once the knowledge base has data.
  2. Position Bees as "coming soon" during the raise. The agentic AI narrative is extremely hot with investors. Show the architecture, the guardrail framework, and early demos. But do not over-commit on delivery timelines for full autonomy. The industry's biggest failure mode is shipping agent products before they are production-safe.
  3. Build the compliance story early. The EU AI Act enforces in August 2026. One in four compliance audits will include specific inquiries into AI governance. Helium's audit trail and data containment features are not nice-to-haves. They are table stakes for enterprise adoption in the coming quarters.
  4. Target shadow AI budgets. Companies waste an estimated $89K annually on unused enterprise AI licenses while employees use free tools. Helium's pitch should focus on consolidating scattered AI spend into one platform that teams will voluntarily adopt, reducing both cost and risk.
  5. Differentiate on model flexibility. The market is splitting into two camps: single-vendor platforms (Microsoft, Salesforce, Google) and model-agnostic platforms (Helium, Poe, TypingMind). For the investor pitch, position Helium firmly in the model-agnostic camp, but with enterprise governance that Poe and TypingMind lack.

Pricing Priorities

  1. Frame the free tier as a trial, not a product. 100 lifetime prompts is roughly 3 to 5 days of active use. This creates natural urgency without artificial time pressure. In pitch decks, show the free-to-paid conversion funnel as a core metric.
  2. Clarify the Teams vs. Max pricing logic publicly. Teams at $80/user is lower per-seat than Max at $100, but the 5-seat minimum means a $400/month floor. Spell this out in marketing. Describe Teams as "Max features, plus team collaboration, at a volume-committed rate." This eliminates confusion and reinforces that the discount rewards commitment, not a reduced feature set.
  3. Position Bees at $99 as a premium add-on, not a standalone. Bees derives its value from the shared knowledge base built by the Teams plan. Selling Bees without Teams would produce weaker agent output and weaker retention. Keep Bees tightly coupled to Teams in packaging and messaging.
  4. Track Pro-to-Max and Max-to-Teams conversion rates obsessively. The $35 to $100 jump and the $100-individual to $80/user-team transition are the two make-or-break points in the funnel. If Pro users are not upgrading, the multimodal features (images, video, slides) may not be compelling enough at that price. If Max users are not bringing in teams, the collaboration pitch needs sharpening.
  5. Prepare a "shadow AI savings calculator" for the sales motion. A team of 5 people each paying for ChatGPT Plus ($20), Claude Pro ($20), and Midjourney ($10) spends $250/month on fragmented, ungoverned AI tools. Helium Teams at $400/month replaces all three with a unified, secure, multi-model workspace. The $150/month premium buys data containment, shared knowledge, audit trails, and model flexibility. That is an easy ROI story.

Metrics to Track and Present

Leading Indicators

Knowledge base documents uploaded per team. Shared prompts created and reused. Model switching frequency per user. Agent tasks completed per week. Escalation-to-human ratio (lower = more trust in guardrails).

Business Outcomes

Shadow AI tool reduction (before/after). Time saved per team member per week. Number of workflows automated by Bees agents. Net revenue retention and expansion revenue from Business to Bees upsell.

Final Note

The strongest signal in this research is that the market is moving away from "more AI power" and toward "safer, governed AI at team scale." Helium's dual-layer product (Business for team governance + Bees for bounded autonomy) is architecturally aligned with this shift. The capital raise story is not about building another AI chatbot. It is about building the platform that makes AI safe and useful for teams. That narrative resonates with enterprise buyers and investors alike.