RADAR
updated 2026-07-20
↻ Refresh
Auto-updates daily. To refresh now, tell Claude “run the radar”. Every change is version-controlled, so nothing is lost.

Pulse — who is getting closer to us · last 30 days

8high impact
10approaching ▲
12on the radar
37field total · see field →
Close to usOn watchBackgroundRead AI — Approaching: Shipped MCP/Claude connectors + 'Ask Read Actions for Sales' and an 'Ada' exec twin — moving from notes to agentic cross-surface actions, edging into real-time meeting→execution. Read AIFireflies.ai — Approaching: Expanded to 200+ AI Skills/agents + 'Live Assist' real-time coaching + an MCP for Claude/ChatGPT/Devin — broadening from notetaker toward agentic, real-time, programmable actions. Fireflies.aiOtter.ai — Approaching: Productized autonomous agents (SDR Agent runs live demos & books meetings 24/7; Sales Agent does follow-ups + CRM sync) — a clear shift from passive notes to agents that act from conversations. Otter.aiFellow — Steady: Recent output is mostly SEO comparison posts + recording/consent/Channels features — consolidating its privacy-first notetaker lane, not pushing into goals or decision governance. FellowSembly AI — Approaching: Semblian agent now generates multi-session roadmap/status docs with prioritized initiatives, revised timelines, dependencies and scope-affecting decisions across planning sessions, leadership syncs and quarterly reviews. Sembly AIMicrosoft Planner — Approaching: Planner Agent in M365 Copilot hit GA 15 Jun 2026 + a new 'goals bucket' for objective-tied tasks + Copilot setting measurable goals. MS PlannerCisco Webex Assistant — Steady: Q1 2026 GA of a Task Agent that creates Jira tickets + updates Salesforce, plus a full agent suite announced at Cisco Live EMEA (Feb 2026). Cisco Webex AssistantZoom AI Companion — Approaching: AI Companion 3.0 added agentic workflows, action-items→Zoom Tasks, cross-source retrieval/memory and a Jira connector; May 2026 added in-meeting workflows (CRM updates from notes). Zoom AI CompanionAtlassian Rovo — Approaching: Team '26 (May 2026): Jira Agents GA assigned to work items with audit logging + new multistep reasoning + a 'Memory' capability + beta goal/project status auto-updates — moved from search/chat to agentic execution. Atlassian RovoMicrosoft (Copilot / Facilitator) — Approaching: Active 2025–26 shipping: GA expansion + public-preview Planner task-sync + Word/Loop drafting + extended to Teams Rooms; Learn doc updated Apr 2026. MicrosoftProductboard — Approaching: Spark agent reached GA ~22 Jun 2026 — ties findings to evidence/citations and closes the spec→shipped→learnings loop, moving toward decision governance and goal-linked provenance (our adjacent turf). ProductboardSpinach — Approaching: Added real-time in-meeting Q&A + live notes and positions itself as turning standup talk into assigned Jira/Linear tickets automatically — deeper into real-time meeting→execution (our turf). SpinachCOLLABIX
colour = impact on us (🔴 high · 🟠 medium · 🟢 low) closer ring = competes more directly bigger dot = stronger company tail = where it came from (◦ start → ● now): outward tail = approaching us, inward = backing off on the radar = closeness 4+ OR a shipped move into our loop · hover a dot for why
Summary · last 30 days
Rivals keep racing to be the meeting 'memory' other AIs read — Otter's meeting notes are now searchable inside Glean, days after it plugged into OpenAI's Codex. The goal held across meetings, with human approval, is still ours.
12 on radar · 8 high-impact · 17 new signals in last 30 days
Run health · 2026-07-20 · coverage 70% · 2 need a look

DAILY FULL pass, 2-day delta (last_checked was 18 Jul). NotebookLM Deep Research WORKED this pass (auth restored) and was the primary engine — one ephemeral notebook, deep-research web sweep returned 103 sources across all 12 on-radar competitors + discovery; ephemeral notebook deleted after. Every figure verified against a real source URL via WebFetch. LinkedIn via CDP-connect (real Chrome, verified logged in on /feed/, authwall:false) read cleanly for otter-ai + semblyai, then LinkedIn rate-limited the session into an authwall for the remaining 9 profiles (browse server also went unresponsive twice and was restarted) — logged in failed[], NOT treated as quiet. RESULT: 2 new material our-lane signals; the rest quiet/deduped/noise. NEW: (1) Otter meeting knowledge now searchable in Glean (14 Jul, Otter blog; part of Glean's 14 Jul sales-connector expansion) — the memory-layer-other-AIs-plug-into land-grab reaching Glean, after Otter→Codex and Read AI→Claude [amber, our-lane; corroborated live on Otter LinkedIn]; (2) Microsoft Planner boards now live inside Outlook + the M365 Copilot app (MC1309747, 'late July 2026' rollout) — native-suite gravity deepening, distinct from the June Planner-Agent signal [amber, our-lane]. VERIFIED-DROPPED: the JotMe claim that Teams meetings connect directly to Planner plans is NOT in the official MC1309747 message — dropped to avoid fabrication. Atlassian Rovo July 'Feature Focus Vol 8' (Profile Memory GA, Smart Links for Goals, MCP v2 preview, @Jira-in-Slack) = monthly roundup, no firm post-18 dates, Rovo MCP/Memory already logged → folded to noise. Read AI / Fellow / Sembly MCP servers = same context-layer theme already logged (dedup). Zoom July updates = admin/mobile polish on logged ZoomMate. Cisco/Webex AI Notes for in-person = capture/translation, not our lane. DISCOVERY: Lamponi (Tokyo) launched a recall/transcription assistant — NO task execution or goal tracking, low proximity, ambiguous date → dropped from intake per 'default to DROP'; Tana remains the standing intake candidate. All 12 on-radar last_checked refreshed to 2026-07-20.

  • ⚠️ LinkedIn — 9 on-radar profiles authwalled after rate-limit Session was logged in (read otter-ai + semblyai OK) but LinkedIn threw an authwall for the rest after rapid company-page loads; browse server also went unresponsive twice. Not covered live this pass — eyeball manually: read-ai (/company/readinc), atlassian (/company/atlassian), productboard (/company/productboard), microsoft (/company/microsoft), zoom (/company/zoom), spinach (/company/spinach-io), fellow (/company/fellow-ai), fireflies (/company/fireflies-inc), webex (/company/webex). Substance for all 12 was still covered by NotebookLM Deep Research + WebFetch. · open ↗
  • ⚠️ X / Twitter Not run this pass — LinkedIn prioritized per runbook step 2 and browser was fragile; X login state unverified. Surfaced so it is not a silent skip. Top handles to eyeball: @otter_ai, @ReadAI_, @sembly_ai, @Zoom, @Atlassian, @firefliesai. · open ↗
  • ⏳ GoRetro changelog no update since May 2023 — product likely stalled (off-radar/background)

A failed source shows here in red — “no news” is never assumed safe.

Look here first

On the radar = closeness 4+ OR a shipped move into our loop. Sorted by impact, then closeness.

High impact ▲ approaching Read AI Huge and growing fast, and just moved its message to 'turn talk into action' — straight at our lane. But people hate its always-on bot, which is our opening. → Do: Read can open a Linear issue; Collabix keeps the goal that issue serves — with the decisions and risks behind it — across every meeting, into Jira/Azure DevOps too. shipped a move into our exact loop (red our-lane signal) High impact ▲ approaching Microsoft Planner Infra Collabix could feed, but strategically heavy: it's Microsoft's PM surface tied to Teams + Copilot where we operate. The new goals bucket lands in our goal-tracking positioning; ecosystem lock is the real risk. → Do: A 'goals bucket' is a list; Collabix keeps the goal's why — the decisions, risks and dependencies — alive across meetings and across tools. competes directly with us (closeness 4+) High impact ▶ steady Cisco Webex Assistant The most direct platform-native threat: its Task Agent extracts AND executes action items into Jira/Salesforce (GA Q1 2026) — overlapping our exact meeting→execution loop with huge distribution. We defend on cross-meeting goal memory + provenance. → Do: Webex opens a ticket from one call; Collabix keeps the goal — and the decisions and risks behind it — across every meeting, into your Jira or Azure DevOps. shipped a move into our exact loop (red our-lane signal) High impact ▲ approaching Zoom AI Companion Direct overlap with our core — and as of ZoomMate (Jun 2026) it now executes cross-stack, not just inside Zoom. We no longer defend on 'they stay in-ecosystem'. Our edge is the goal kept across meetings with decision provenance and a human approval gate; Zoom's lock-in is now the call platform (you must meet on Zoom). → Do: Zoom Companion runs work from a call; Collabix keeps the goal across meetings — its decisions and risks — and ships the approved next step into your trackers. shipped a move into our exact loop (red our-lane signal) High impact ▲ approaching Atlassian Rovo Our strongest rival on paper: an AI agent across Jira/Confluence that now executes work (Jira Agents assigned to work items) and added a 'Memory' feature plus auto goal/status updates. But it only sees Atlassian tools and still needs a human to publish goal updates — that handoff + lock-in is our wedge. → Do: Rovo's depth is its cage — it only sees Atlassian. Collabix keeps your goal alive across any stack (Jira + Azure DevOps + Slack + Teams). shipped a move into our exact loop (red our-lane signal) High impact ▲ approaching Microsoft (Copilot / Facilitator) Collabix's biggest direct threat — same meeting-to-tasks motion, free-riding on existing M365 Copilot seats. But its gaps are exactly our wedge: (1) single-meeting only — no goal/decision memory ACROSS meetings; (2) tasks land in Planner/Teams, NOT Jira or Azure DevOps; (3) locked to the Microsoft stack, not cross-vendor. → Do: Copilot summarizes one meeting; Collabix carries the goal across meetings and ships it to the trackers your devs use — Jira/Azure DevOps, not just Planner. shipped a move into our exact loop (red our-lane signal) High impact ▲ approaching Fireflies.ai Massive (≈$1B, 20M users) and very loud on LinkedIn. Different DNA (transcription), so watch for the day they ship real execution — that's when proximity jumps. → Do: Fireflies gives you 200 skills; Collabix gives you the one thing that matters — the goal, governed and remembered across meetings. competes directly with us (closeness 4+) High impact ▲ approaching Otter.ai Now a RED our-lane rival, not a notetaker: it turns meetings into cross-stack auto-execution (Salesforce/Gmail/Jira) + a memory layer other AIs build on (MCP). Overlaps our core loop at huge scale. We defend on the goal kept across meetings, decision provenance, and human-approved (not auto-run) execution. → Do: Otter turns meetings into a searchable knowledge base and fires off tasks; Collabix keeps the goal moving across meetings — its decisions, risks and the approved next step — not a pile of auto-actions. shipped a move into our exact loop (red our-lane signal) Medium impact ▲ approaching Productboard Closest to us on 'goals', and owns the Discovery→Delivery wording. But its goal dies when work hands off to Jira — we keep it alive. → Do: Productboard's provenance lives in product strategy; Collabix's lives in the meeting — every decision traced to the moment it was made. competes directly with us (closeness 4+) Medium impact ▲ approaching Spinach Closest to us in intent — an AI for standups/retros that already ships. Smaller and ceremony-bound; we win with a continuous goal across all meetings. → Do: A generic 'context layer' just stores talk; Collabix attaches context to the GOAL — decisions, risks, dependencies — so it drives work, not just recall. competes directly with us (closeness 4+) Medium impact ▶ steady Fellow Made the boldest move late 2025 (botless + MCP, auto-creates Jira/Asana/Linear tasks). It erodes our old 'no bot' edge — so we lead on goal memory, not 'no bot'. competes directly with us (closeness 4+) Medium impact ▲ approaching Sembly AI One of the most directly overlapping rivals: tracks decisions, dependencies and owners and regenerates roadmap/status deliverables across planning sessions. KEY: its 'Success Plans' are generated DOCUMENTS (re-made each cycle), not a persistent live goal object — so it approximates, but does not equal, our goal-context memory + decision provenance. → Do: Sembly remembers and documents meetings; Collabix drives the goal to done — decisions, risks and the approved next step across meetings and your trackers. competes directly with us (closeness 4+)