Conviction Statement

One of the stickiest segments of the IT market has been ERP^[1]. This presents an opportunity.

ERP Vendors like Oracle and IT were known for their horrible implementation cycles and user experience.

Into this void entered SaaS products which could target department heads with functional specifications. The biggest leader of this has been Salesforce, which dominates the SFA, Service, and Marketing space; Atlassian for IT and development; and Hubspot for marketing.

Some could even be function and vertically segmented.

However, this created silos. Into this space entered RPA vendors, such as UI Path, Workato, and even Zapier in the SMB space. This empowered "citizen developers" who were in charge of operations to "stich together" the silos more readily without knowing how to code.

Concurrently, internal applications grew, and tools like Retool enabled developers to build more quickly from a low-code scaffolding. Businesses, especially those scaling, recognized that they needed specialized, bespoke tools to handle these workloads.

However, there appear to be a few potential seams in the market. Rippling, for example, pioneered the "compound start-up" by building a seamless internal data platform that enables them to build natively-integrated products around an entire set of functions. In this case, they have started with HR. Doing so allows them to restore control of a central data model, in this case, the Employee Data Model, and build workflows around that data model for different functional users.

This removes the need for integration and automation tools, giving a better UX for their tools.

The opportunity is to combine the trends across these:

• Citizen development - low or no code
• Rapidly built, bespoke applications
• Common shared data models to remove external workflow automation tools

The rise of LLM, as well as flexibility in speed and architecture of data backends, presents an opportunity to leverage these trends for a better experience for companies to streamline their operations.

The AI-Driven Business Systems product would allow for the following:

• Citizen development starting with text-based PRDs of clear workflows
• Ready to use internal applications complete with UIs, reporting, alerts, and workflows
• Easy extensibility upon a shared data model - no more RPA tools and better reporting

We are already seeing momentum of low-code automation, because it solves the problem of custom workflows with application silos.

Which brings us back to a long-term vision of what will become of the ERP market.

ERPs are essentially internal forms plus workflows plus a database. But because of the scale and the customization, these have been expensive to put together. These aren't likely to be disrupted anytime soon, but they illustrate what the next-decade could bring.

Fully customized, end-to-end business software that takes a fraction of the development costs of traditional ERP.

The consultants can be either much smaller boutique firms or internal teams that are able to think more cogently around their requirements, and it would introduce a product management discipline into internal tooling by abstracting away integrations with third party systems and data schemas.

This rest of the document was constructed using LLM to give a framework and has not been edited or battle-tested by a human yet.

Summary of Problems with the Conviction Statement

Here are some potential holes I see in the conviction statement around an AI-Driven Business Systems startup:

• Significant competition from major enterprise software vendors. Large companies like SAP, Oracle, Salesforce, Workday, etc. already provide integrated business systems, ERP, automation and low-code platforms with strong data models. They have huge R&D budgets, customer bases, and partnerships. A new entrant would face an uphill battle competing at enterprise scale.

• Difficulty expanding beyond initial targeted functions. While starting with HR data and tools may prove the concept, expanding into the full range of ERP capabilities from finance to supply chain is a massive, multi-year undertaking. Significant investment and execution risk.

• Reliance on emerging technologies like low-code and process automation. Although promising, these technologies are still developing. If they fail to gain mainstream enterprise traction, it undercuts a key element of the product and market opportunity.

• Potential saturation of the "citizen developer" and IT toolbox market. There are many vendors already providing platforms for business users to build custom tools, integrations and automations. The space is becoming crowded, and consolidation may limit opportunities for new startups.

• Barrier to transitioning existing enterprise tech stacks. Large companies have invested heavily in their existing software and data infrastructure. Convincing them to adopt an all-new AI-based platform would require very compelling value propositions around cost savings, efficiency gains and innovation. Not an easy sale.

• Scarcity of resources. Finding engineering and product talent with experience building enterprise-scale ERP, data platforms and business systems software may prove challenging. Also limited availability of large datasets to train AI systems. Resource constraints could slow progress.

• Long sales cycles and implementation times. The sales process for company-wide business systems often spans 6-18 months or more. Implementing and migrating to new platforms can take years. This poses high costs and risks that threaten the viability of a young startup.

While the vision is compelling, executing an end-to-end "AI-Driven Business Systems" product at enterprise scale seems incredibly difficult given the current limitations around data, AI, low-code tools—and the realities of large company decision making. That said, a successful product could be highly disruptive if the right approach is found.

Possible Personas

Operations managers

• Struggle with inefficient processes that reduce productivity and increase costs.
• Lack visibility into key metrics that would provide optimization insights.
• Rely on technical staff or costly agencies to build any solutions or automation.

Business analysts
- Spend more time gathering requirements and managing solutions rather than high-value analysis. 
- Use tools that require advanced technical skills, limiting their own ability to prototype and configure solutions. 
- Locked into generic off-the-shelf platforms that may not fully meet unique needs.

Department heads 
- Need process automation, reporting and insights to manage teams, budgets, and outcomes at scale. 
- Lack technical backgrounds to build solutions themselves or fully evaluate options.
- Forced to compete for technical resources or budget for expensive custom development.

Possible Feature Requirements

The following is just a hypothesis. I'm not certain these are truly necessary but want to strawman these ideas.

At a minimum, an AI-based low-code automation platform should have the following key features:

• A visual process mapping interface.
  The ability for non-technical users to map business workflows with an intuitive drag and drop canvas. Connecting various steps, logic, data, and actions into an automated process flow.
• Pre-built workflow components and automations.
  Common elements like forms, alerts, queries, file ingestion, etc. that users can quickly drop into their workflows without engineering everything from scratch.
• Data integration capabilities.
  The ability to connect with commonly used data sources, APIs, and applications to pull data into workflows and push data back out. Supporting real-time monitoring and automation.
• Logic and decisioning features.
  Options like conditional logic, variables, if/then rules, and formulas for building smart decision paths into workflows. Ability to route processes dynamically based on parameters.
• Reporting and analytics.
  Providing insights into how automated workflows and processes are performing. Visualizing KPIs, outputs, cycle times, and optimization opportunities. Ability to report insights back to workflow logic.
• Management and administrative functionality.
  Ways to manage user access control, monitor resource usage, schedule/trigger automations, handle errors, perform audits, backup data, and generally govern the platform.
• Deployment and security options.
  Ability to securely run the platform in both public cloud and private on-premises environments. Options to configure data encryption, redundancy, and compliance based on customer needs.

The biggest technical risks include:

• AI and ML components not yet ready for production or at the level required for truly optimized workflow automation and decisions at scale. May require more human configuration and management than marketed.
• Immature automation technology leads to unstable or insecure processes and data handling. Vulnerabilities that could disrupt customer operations or lead to non-compliance.
• Lack of ecosystem and interoperability with other key platforms and systems. Unable to integrate and exchange data with tools customers already rely on. Significant migration needs.
• Scarcity of expertise in key technical areas like process mapping standards, real-time data flows, automated decisioning, natural language parsing, and workflow analytics. Resource constraints for ongoing engineering and product development.
• Scalability challenges where the platform cannot keep up with real volumes of customer activity, data integration needs, reporting requirements, and concurrent automations. Performance issues and downtime.
• Inability to operationalize and make the platform simple/accessible enough for non-technical business users as marketed. Still demanding professional developer skills to truly configure and benefit from.

Ideal Initial Workflows

Attributes

• Internal facing - Focusing on automating internal processes first reduces security and compliance risks. Once proven internally, options to open up external access become possible.
• Non-mission critical - Starting with areas that improve efficiency and insights but won't disrupt core operations if there are issues. Once the platform matures, it can then be applied to more critical workflows.
• High-value but accessible - Workflows that provide significant benefits across an organization but don't require advanced technical skills to automate and optimize. Quick wins.

Examples by Vertical

Some examples of good starter workloads by industry include:
• Healthcare:
› Patient intake and onboarding
› Resource scheduling (staff, facilities, equipment)
› Patient communications and outreach
• Higher Education: 
› Student admissions and enrollment
› Course scheduling and registration
› Faculty onboarding and compliance training
• Nonprofit: 
› Volunteer recruitment and management
› Donor outreach and stewardship
› Event planning and logistics
• Professional Services: 
› New client onboarding
› Timesheet and expense reporting
› Recruiting and interview scheduling 
• Retail/E-commerce:
› Supplier and inventory management
› Customer service request routing
› Loyalty and promotions management

Possible Go To Market Strategies

Industry segments

• Healthcare - Large market with many workflows around patient management, billing, resource scheduling, etc. that require optimization and insight. Compliance needs also drive demand for auditability and control.
• Education (higher ed and K-12) - Schools and universities have diverse operational needs but limited budgets and IT resources. Platform could be tailored for admissions, enrollment management, student services, etc.  
• Nonprofits - Organizations focused on missions over revenue need accessible, cost-effective tools for managing fundraising, volunteers, programs, and events. Verticalized platform addressing their unique needs could gain strong adoption.
• SMBs (retail, professional services, etc.) Platform tailored for the workflows of specific trades and optimized for limited technical sophistication could gain grassroots momentum.

End users

• Operations managers - Responsible for key processes, budgets, resources and outcomes. Look for solutions to optimize productivity, reduce costs, and gain data-driven insights. Early adopters.  
• Business analysts - Technical enough to use more complex tools but prefer to spend time on high-value analysis vs. tedious solution configuration. Serve as internal champions to deploy and scale the platform within organizations.  
• Department heads - Drive demand for solutions that provide process efficiency, reporting, and KPI tracking across teams and budgets but from a non-technical business user perspective.

Engage mostly with pre-built solutions and configured use cases on the platform vs. building net new applications themselves. Rely on operations managers and business analysts to develop and manage more complex implementations.

Growth Channels

1. Channel partners - Working with solution providers, consulting firms, and system integrators that already serve your target mid-market organizations. Provide incentives for partners to build solutions and workflows on your platform that they deploy for clients. Co-sell with partners on new deals. This leverages their existing customer base and expertise while fueling partner-referred revenue that often has higher retention rates.

2. Product-led growth - Using a freemium model, free community edition, or low-cost simple workflow tools to drive broad initial adoption and trial. Then upsell based on usage, workflows created, team members added or other product qualifiers. This efficient self-service acquisition model works especially well for platform models where value expands over time through broader usage and more advanced functionality.

3. Content and education - Creating a lot of valuable and thoughtful content on workflow optimization, automation best practices, productivity, and business transformation. Educate the market on key opportunities and use cases, then position your platform as the solution. Convert readers into trial users or leads for direct sales. Sponsor key industry events to raise brand visibility.

4. Vertical market focus - Developing specific vertical solutions, workflows, and the ability to quantify benefits for target industries like healthcare, education, finance or retail. Speaking the language of specific verticals and tackling their biggest operational challenges helps efficiently penetrate these segments through tailored marketing, sales and channel strategies.

5. Referrals - Focusing on delivering exceptional value and customer experiences so that existing customers refer others. Referrals are the most efficient source of new leads since the customer is already sold on the value and ready to buy. They simply need help determining the best level of access or functionality. Reward existing customers for referrals that turn into revenue.

6. Less efficient options for this type of solution include:

   1. Paid advertising - Expensive, better suited to targeting specific keywords or audience segments for strategic campaigns. Not ideal for broad customer acquisition.
   2. Direct outbound sales - Cold calling, emailing and other forms of outbound selling tend to be inefficient especially for a complex platform sale. Inbound interest and channel partnerships are better models.
   3. Enterprise sales - Long sales cycles requiring significant investment to land major accounts. Should only pursue for largest opportunities with the halo effect of brand credibility. Not scalable for mid-market.

An integrated strategy leveraging channel partners, product-led growth, content marketing, vertical focus and customer referrals will be most efficient for broad customer acquisition. Then enterprise and strategic paid advertising can accelerate growth over time. But a strong self-service funnel and network of advocating partners is key.

Competitors

Rippling

Rippling is a fast-growing HR and business systems software startup.

• Rippling targets small to midsize businesses with an all-in-one solution for HR, payroll, benefits and IT management. They aim to simplify hiring and onboarding using an automated cloud platform.

• Rippling has raised over $245M in funding, allowing them to scale rapidly. They claim over 1,600 customers and 300,000 employees paid to date. The company is not yet profitable but focused on growth.

• Rippling's key vulnerability is their reliance on just HR and adjacent offerings. They do not provide a complete ERP solution spanning finance, supply chain, inventory, and other core business domains as larger vendors do. May be limited to SMB at current scope.

• An opportunity exists for an AI-driven integrated business systems suite to compete with Rippling. Additional capabilities around intelligent workflows, predictive analytics, benchmarking and optimization across a full set of back-office functionality could provide significant added value for customers.

• Rippling will likely face increasing competition from major enterprise software vendors adding similar easy-to-adopt HR and payroll solutions tailored to SMB. These larger competitors can bundle additional applications into their suites.

• Rippling may struggle to expand into midmarket and large enterprise if they rely only on their current HR and basic IT management tools. More sophisticated ERP functionality and compliance will be required. Could be an acquisition target for a bigger provider seeking Rippling's modern UX and audience.

To summarize, Rippling has found success targeting SMB by simplifying HR tasks like employee onboarding using an automated cloud platform.
However, their focus on primarily HR and adjacent areas also poses a key vulnerability relative to broader ERP suites, especially as competition increases. An opportunity exists for AI-based business systems software to provide integrated HR, financials, and business optimization at scale.

Rippling's future likely depends on continued fast-paced innovation to stay ahead of larger competitors, potential expansion into more robust ERP offerings, targeting higher midmarket clients—or possibly being acquired. Their position today seems ripe for disruption by a solution able to deliver a modern, AI-optimized experience across all critical back-office functions, not just HR.