The Forrester Wave™: Document Mining And Analytics Platforms, Q2 2026 highlights a market that is broad, fragmented, and rapidly evolving — where success depends less on vendor selection alone and more on precise alignment to use cases, document types, and architectural choices. While innovation around agentic AI and LLMs is accelerating, most organizations are still early in adoption and must balance ambition with practical realities around time to value, cost, and accuracy. The research findings and recommendations below focus on helping leaders cut through the noise, make grounded decisions, and design for flexibility as the market continues to mature.
Understand that DMAP is not a single market. Document mining and analytics platforms (DMAP) capabilities span multiple segments, including enterprise content management (Hyland, Iron Mountain, OpenText), intelligent automation (Automation Anywhere, EdgeVerve, Rossum, UiPath), search/knowledge platforms, and other niche specialties. Start with your use case, then shortlist vendors within the right segment — not the other way around.
Refine your use case early because document type matters more than you think. Long-form documents (e.g., contracts) and high-volume transactional documents (invoices, purchase orders, shipping labels) require very different capabilities. Vendor strengths diverge significantly here, so clarity up front avoids costly mismatches later.
Pay attention to platforms’ agentic AI architecture flexibility. The ability to plug different ML models into different DMAP functions is critical — not every task needs the most expensive LLM. Use premium models for complex extraction (e.g., images, cursive writing), lighter models for summarization, and custom ML where needed (e.g., complex forms and tables).
Plan realistically for time to value. Expect about six months to reach MVP, not a few weeks, as some vendors suggest. Timelines vary widely based on document complexity, languages, and geographic scope.
Understand pricing dynamics before scaling. Costs typically range from approximately $0.05 per page at high volumes (millions annually) to approximately $0.20 per page at lower volumes. Your business case will hinge heavily on volume assumptions and document mix.
Set pragmatic expectations for accuracy. Accuracy often starts around 60%-plus and improves (to the high-90% range) with tuning, but it varies by document complexity, structure, and language. “Human in the loop” processes remain essential for most production deployments.
Last but not least, as agentic AI pushes teams toward build vs. buy decisions, download the evaluation spreadsheet included with the Forrester Wave. It outlines the criteria and key questions needed to guide both platform selection and in-house builds. Bottom line: Think carefully before taking on that level of complexity.
If you have more questions about this or any other research, please do not hesitate to set up a call with me.
