Spend Matters, a Hackett Group Division, is pleased to announce the 2025-2026 roster for our ‘Future 5’ procurement solution provider list. As in previous editions, our analysts have highlighted five start-ups that excite them the most. This year, the five companies they see as potential future trendsetters are Flowie, Pavus, Tamarin AI, Vallor and Zapro.
To earn such recognition, these vendors must have a product that is ideally two to five years old, is used by more than five customers and showcases an innovative use of technology. Additionally, these vendors should have revenue below $10 million, and our analysts must determine that they are both sustainable and growing with clear momentum.
In this series of articles, we explain what these vendors offer, why they are likely to become future members of our 50 to Watch and 50 to Know procurement provider lists, and what challenges they might encounter.
Get to know all the procurement tech providers that made our lists this year.
Here’s why we chose Pavus.
What is Pavus
Pavus AI is an all-in-one sourcing and procurement analytics platform that combines spend visibility, cost intelligence and supplier discovery with sourcing execution tools. The platform aims to unify multiple procurement functions into a single, data-driven system powered by machine learning.
At its core, Pavus converts unstructured procurement documents (purchase orders, invoices and other transactional data) into organized spend cubes, regardless of the coding standards or formats used by companies. This data foundation supports detailed procurement analytics that identify cost savings opportunities by comparing actual spending against market benchmarks and commodity indices.
The platform’s key feature is its cost analysis, which breaks down any product into its component materials and labor. Users upload product specifications in PDF format, and Pavus deconstructs the item into its constituent materials with their respective weights. The system then links these materials to commodity indices in specific regions and uses benchmarking data to determine a target price, showing whether customers are paying above or below market rates.
For supplier discovery, Pavus consolidates data from various third-party sources and custom-built tools that leverage OpenAI’s advanced search features. The platform delivers combined results from these sources to assist users in finding suppliers based on criteria such as location, revenue, sustainability certifications and product capabilities. Users can view detailed supplier profiles and product catalogs directly within the platform.
The sourcing execution module uses should-cost analysis as a target price benchmark during competitive events. Pavus supports multi-round bidding, allowing procurement teams to hold multiple rounds with suppliers to gradually lower prices. The platform includes built-in communication tools that function like messaging apps, removing the need for external email exchanges, as well as document management features.
Why we chose Pavus
We chose Pavus because it demonstrates a truly AI-native approach to procurement technology, built from the ground up to utilize recent advances in machine learning rather than simply adding AI features to legacy systems.
The platform’s AI-driven design is clear in its key features. The data conversion engine uses machine learning models to handle unstructured procurement documents in any format or coding standard, automatically turning purchase orders, invoices and transactional records into organized spend cubes without manual mapping or extensive setup. This eliminates the data normalization bottleneck that has traditionally restricted spend analytics deployments.
The system’s cost analysis demonstrates how AI enables capabilities that previously required extensive human expertise. It automatically extracts product specifications from PDF documents into materials and weights, then links these components to real-time commodity indices by region. This material breakdown and cost modeling, which was typically performed by specialized engineering and procurement consultants, is now accessible to procurement teams through AI-powered analysis. The platform integrates purchased commodity databases with web scraping algorithms that continuously gather open-source pricing data, providing daily market intelligence updates.
The supplier discovery feature demonstrates effective AI integration. Pavus uses custom tools that utilize OpenAI’s powerful search capabilities layered over traditional supplier databases like Veridian. These tools employ machine learning algorithms to preselect and rank the most suitable suppliers based on user needs. The system filters across multiple criteria at once, such as manufacturer versus distributor, sustainability certifications, revenue thresholds, geographic presence and product capabilities. It then displays supplier profiles with product catalogs in response to natural language queries.
What sets Pavus apart from competitors adding AI features to existing platforms is how these capabilities are connected. The spend analytics inform the should-cost models, which generate target prices that feed directly into sourcing events, where multi-round bidding data flows back into the benchmark database. This creates a learning system where each transaction improves the intelligence for future decisions. The vendor’s roadmap includes proactively launching sourcing events specifically to gather market pricing data, further illustrating how AI enables procurement operations that would be economically infeasible with manual processes.
The platform extends its AI-driven approach into workflow automation. Built-in communication tools prevent external email exchanges during sourcing events, and document management seamlessly integrates with the bidding process. These are not separate modules added together but rather components of a unified system designed around machine learning from the beginning.
For procurement organizations exploring AI, Pavus is a strong candidate: instead of merely adding AI features to existing processes designed for humans, the platform reimagines procurement workflows based on what AI can enable. The founding team’s pursuit of fintech Series A funding shows both their ambition to expand this vision and their confidence that the market is ready for AI-native procurement infrastructure.
Threats or challenges ahead
It is important to remember that Pavus is a new vendor in a competitive market with well-established competitors across multiple categories. The platform operates in spend analytics, cost modeling, supplier intelligence and sourcing execution, each led by experienced professionals.
The should-cost methodology, while technically interesting, has inherent accuracy limitations. For standardized materials, linking to commodity indices generally works well, but for custom components, the approach requires accurate data on conversion costs and manufacturing processes. Pavus addresses this by using financial and operational reports from companies, but these averages may not reflect the specific capabilities or cost structures of individual suppliers. Should-cost models are best considered as negotiation tools rather than precise price predictions, and customers might have different expectations about their accuracy.
The supplier discovery feature’s reliance on third-party data providers raises concerns about dependency and cost structure. Pavus does not offer direct access to tools; instead, it incorporates external data into sourcing results. The financial impact and sustainability of this integration model remain unproven, especially as the vendor expands and data costs increase with usage.
From a product maturity standpoint, some capabilities are more developed than others. For instance, the company is still identifying the best data sources for specific functions.
Although we are optimistic about Pavus’s integrated approach and its data-driven value proposition, the vendor encounters the typical challenge faced by innovators: proving that unification provides enough value to justify the switching costs, integration complexity, and perceived risks of replacing specialized tools with a platform from a startup.
Join our analyst-led webinar on April 15 to discover more about the selection process.
