Users, Market, and Competitors

Users, Market, and Competitors

The Strategic Signal

The market is moving beyond transcription. Plaud is pushing AI note-taking hardware toward always-on, multi-Agent, personalized wearables. Granola is turning meeting history into an enterprise Context and MCP layer. Town is financing a personal assistant that “learns how you work.” ChatGPT, Gemini, and Apple are making personal Context part of the model or operating system.

Memova’s opportunity is not to produce another meeting summary. It is to turn handwriting, voice, and real-world Context into a calibrated, exportable, Agent-readable personal semantic layer.

Plaud: Hardware Is Becoming a Real-World Context Layer

Plaud has demonstrated a hardware-to-subscription business loop at significant scale. Its narrative has expanded from recording and summarization toward an upstream source of truth for professional Context. Team workspaces, shared meeting insight, automated processing, and a future wearable point toward continuous capture and personalization.

This is a direct threat because “real conversations become the data layer for AI action” overlaps with Memova. It is also validation. Plaud remains primarily voice-first and professional. Memova can differentiate through handwriting plus voice, a model of people, projects, commitments, and preferences, and a feedback loop that lets the user correct the resulting memory.

Town: The Closest Personal-Alignment Narrative

Town describes an assistant that learns how a person works across email, calendar, messaging, documents, and business tools, then operates routines such as inbox handling, outreach, schedule optimization, and meeting briefings.

This makes Town the clearest narrative competitor to personal superalignment. It is not merely a note-taking product; it promises proactive action based on learned work patterns. Memova must therefore make its distinction explicit: an Agent-readable knowledge base the user owns, physical-world capture, source-level traceability, feedback writeback, and approval boundaries for action.

Granola: From Meeting Notes to Company Context

Granola has expanded from an AI notepad into shared Spaces, APIs, MCP access, and enterprise controls. Its meeting transcripts, decisions, and action items can become a data source for Claude, ChatGPT, Cursor, and other Agents.

Granola’s strength is company Context and Agent-readable meeting history. Its center of gravity is not a person’s long-term preferences, goals, relationships, or private life. It may nevertheless own the “meeting memory to Agent” software entrance. Memova should avoid competing as a better meeting-notes application and instead connect multiple input surfaces into a personal semantic model.

Read AI: Meeting-to-Action and the Digital Twin

Read AI’s Ada and Digital Twin direction shows how meeting capture can expand naturally into email and scheduling action. The system can draw on transcripts, summaries, decisions, and action items while using approval flows and guardrails.

This validates the move from memory to action. Memova’s defensive advantage must be a deeper, user-owned semantic layer with correction, portability, and Context beyond meetings, rather than a stronger meeting bot alone.

Passive Memory Wearables and Platform Absorption

The acquisition of Limitless and continued investment by major platforms in wearable AI indicate that passive memory capture is becoming a system-level interface. This creates a structural warning: a hardware-only memory company may eventually be absorbed by the model or operating-system layer.

Memova should own what remains valuable across devices—the portable memory schema, source provenance, personal semantic model, correction history, and access boundaries—rather than placing its entire moat in one piece of hardware.

Memory Infrastructure

Companies such as Supermemory and Personal AI validate “memory as infrastructure.” They build engines, filesystems, or model layers that help Agents retrieve and use long-term memory.

These systems may become suppliers, partners, or substitutes. Memova’s key decision is whether it defines and owns its Note-to-memory schema and feedback loop. If long-term memory is entirely outsourced to a third-party API, the product risks losing the layer that allows one person to audit, correct, export, and reuse their memory across Agents.

Platform-Level Personal Context

ChatGPT, Gemini, and Apple are all making preferences, goals, prior conversations, messages, email, photographs, and application actions part of the default assistant experience.

Memova cannot rely on the generic promise that “AI remembers you.” Platforms will increasingly provide that. The independent value is a personal semantic asset that the user can inspect, correct, export, and authorize for any Agent—not a hidden set of model memories controlled by one vendor.

Who Memova Serves

Memova is most useful for people whose work and lives produce more Context than existing tools can help them understand:

• founders running multiple evolving projects;
• researchers who need provenance and decisions to survive across long timelines;
• investors connecting meetings, companies, theses, and commitments;
• creators whose work process can become public expression;
• families preserving stories without surrendering an entire private archive;
• long-running teams that need continuity beyond individual tools and employees.

The initial wedge should be narrow enough to prove repeated value. Founders building in public are strong candidates because they create dense Context, need action, and benefit when verified work becomes shareable Pages.

Positioning

The positioning should move from “AI notebook” to:

A personal semantic layer for Agents: Notes become structured memory, and structured memory becomes calibrated action.

The product’s differentiators are:

1. capture across handwriting, voice, documents, meetings, and Agent work;
2. a user-owned, local-first, exportable knowledge layer;
3. explicit models of people, projects, commitments, preferences, and open questions;
4. source-level provenance and the separation of Archive from Interpreter;
5. user correction that writes back into future understanding;
6. approval boundaries between suggestion, drafting, and consequential execution;
7. interoperability across models and Agents.

Immediate Product Actions

Memova should make the following actions concrete:

1. Tighten the public claim around the personal semantic layer rather than another AI notebook.
2. Maintain a Plaud battlecard focused on hardware plus subscription, always-on capture, developer access, and the difference between voice capture and correctable personal memory.
3. Maintain a Town battlecard comparing input surfaces, memory ownership, auditability, approval, action success, and Agent interoperability.
4. Prioritize “confirm before action” and “write feedback back after action” in the MVP.
5. Preserve the local-first, exportable, Agent-readable architecture as platform memory becomes stronger.

Signals to Watch

For Plaud, watch whether future wearables confirm independent connectivity, always-on operation, multiple Agents, personalization, and developer interfaces.

For Town, watch for a public memory schema, auditable preferences, an approval model, and demonstrated cross-application action.

For Granola, watch whether MCP and API access expand from reading notes to writing back into CRM, project systems, and decisions.

For Read AI, watch whether Ada becomes a broader chief-of-staff Agent and whether users can correct long-term preferences.

For major platforms, watch whether personal memory becomes exportable, source-auditable, and readable by third-party Agents.

The market is validating capture, memory, and action separately. Memova’s opportunity is to connect them around one person while keeping that person in control.

Agent and AI Notebook Market Research - 2026-05-14

Key Findings

• US workplace AI use is mainstream, but true agent use is still earlier: Gallup reports 50% of US employees use AI at work at least a few times per year, 28% use it weekly or more, and 13% daily as of February 2026. Gartner reports 75% of surveyed orgs are piloting/deploying some form of AI agent, but only 15% are considering/piloting/deploying fully autonomous agents.
• Developer coding agents are much more mature than general workplace agents. Stack Overflow 2025 reports 31% of developers currently use AI agents, 17% plan to, and 69% of agent users report productivity gains. JetBrains January 2026 reports 74% of developers worldwide use specialized AI developer tools, with US/Canada Claude Code work adoption at 24%.
• Codex appears to be the fastest-growing named coding-agent product with official user counts: OpenAI reported 3M weekly active Codex users in early April 2026 and more than 4M two weeks later. Earlier public milestones were 1M+ desktop app downloads in the first week and 1.6M WAU after GPT-5.3 Codex, then 3M WAU in early April.
• AI meeting/notebook apps are growing quickly but remain smaller than traditional workflow platforms by users/customers. Plaud reported 2M global cumulative users by April 2026. Granola disclosed 10% weekly user growth in May 2025 but not absolute users. Read AI reported 5M MAU and 50K daily signups in February 2026. By contrast, Notion reports 100M users and monday.com reports 250K+ customers.

Implications For Memova

• The market signal supports Memova's notes -> memory -> actions thesis: broad AI use is already present, but workflows have not transformed for most users.
• Coding agents show the most advanced adoption pattern: users accept agents when they operate inside an existing workflow, can take action, and remain inspectable.
• Meeting-note tools validate capture and summary demand, but most public growth is still around transcription and notes. Memova should differentiate on actionable memory, people/project context, and confirmed next actions rather than "better notes" alone.

Read AI Case Study

• Founded in 2021 in Seattle by David Shim, Elliott Waldron, and Rob Williams. The founding team came from Placed/Foursquare/Snap; Shim had been founder/CEO of Placed and later CEO/president at Foursquare.
• Initial wedge: meeting analytics and AI summaries inside Zoom/Google Meet/Microsoft Teams. The product used a meeting bot and free utility to enter existing meeting workflows.
• Development path: meeting summaries -> cross-meeting search and coaching -> "connected intelligence" across email/messages/meetings -> agent-like assistants such as Ada for scheduling, answers, and action handling.
• Funding: about $10M seed in 2021, $21M Series A in 2024, $50M Series B in October 2024. Total disclosed funding is roughly $81M.
• Growth markers: by late 2024 Read AI claimed 100K new accounts weekly and 75% Fortune 500 penetration. By February 2026 it reported 5M monthly active users, 50K daily signups, and an internal goal of 10M MAU.
• Memova lesson: Read AI entered through a high-frequency, low-friction capture surface, then expanded horizontally into workflow context. Memova should consider a similarly narrow entry point, but differentiate by owning personal memory and confirmed follow-up actions rather than only meeting capture.

Sources

• Gallup AI Indicator, updated April 2026.
• Stack Overflow 2025 Developer Survey.
• JetBrains AI coding tools survey, January 2026 wave.
• OpenAI Codex enterprise posts, April 2026.
• TechCrunch Granola Series B, May 2025.
• Plaud press release, May 2026.
• TechCrunch Read AI Ada launch, February 2026.
• GeekWire and TechCrunch Read AI funding/growth coverage, 2021-2026.
• Notion 100M users post.
• monday.com investor relations and FY2025 results.