About the Company
Riverwalk Holdings is an investment firm that needed to dramatically reduce the time its investment team spent on manual research and monitoring of target and portfolio companies. The investment thesis at Riverwalk is built on signal - changes in regulation, leadership, news, and ecosystem activity that move the needle on their portfolio - and the existing process of tracking that signal was bandwidth-bound.
Riverwalk partnered with Tequity to build an AI-driven monitoring platform that would continuously track companies across diverse data sources, surface real-time alerts and insights, and let the investment team focus on judgment rather than data gathering.
The Challenge
Investment intelligence platforms are deceptively hard to build well:
- Diverse data sources to integrate: Tracxn, K-Scan, LinkedIn, Medium, Substack, regulatory filings, and tens of thousands of links across the public web - all with different schemas and reliability.
- Quality signal at scale: Once data is ingested, the platform has to surface what matters to a human analyst, not drown them in noise.
- AI summarization and scoring: Summaries had to be accurate, useful, and tunable, with relevance scoring the team could trust.
- Workflow integration: The platform had to slot into Riverwalk's existing workflows, including Zoho, rather than become another silo.
- Iterative product discovery: As the team used the system, the requirements would shift. The platform had to be built for fast iteration on prompts, scoring, and UI.
Solution Implemented by Tequity
Tequity ran the engagement as an MVP-driven design and engineering build, with the platform now scaling toward full production rollout.
Discovery and Foundation:
- PRD and wireframing completed up front to lock the v1 scope.
- Designs closed and approved by the Riverwalk team.
Data and Scraping Layer:
- Connected and standardized all major data sources, including Tracxn, K-Scan, Apify-based public profile scraping, LinkedIn, Medium, and Substack.
- Built a Tracxn link scraper across ~21,000 company links for initial company-data ingestion.
- Implemented a K-Scan webhook for handling missing Tracxn data gracefully.
AI Layer:
- Summary engine with a tunable summary style.
- Relevance scoring with iterative prompt optimization.
- Dynamic addition of public profiles for people on Medium and Substack.
- Semantic search across Market Pulse and the People module.
Product Modules:
- Market Pulse - news, regulatory, and circulations scraping with a UI tuned to investment workflows.
- People - end-to-end add and LinkedIn scraping flow with feedback-driven scoring.
- Company - tracking module backed by the Tracxn scraping pipeline.
- Chat - research interface, fully built and pending only organization verification for final testing.
Workflow and Integration:
- Zoho implementation completed.
- Taghash integration shipped.
- Bug fixes, demo support, and feedback iterations as the team began real usage.
The platform is now moving into production after multiple successful client demos and feedback rounds.
Results
The platform delivered measurable time leverage to the Riverwalk team:
- 3 hours per day saved by Associates and Partners on research and analysis
- Multiple data sources unified into a single intelligence surface
- Live AI summarization, scoring, and semantic search powering daily usage
