Seven products already sell thesis tracking, evidence citations and decision journals. None of them connects the whole chain — from the filing that changed your mind to the outcome that proved you right or wrong. That chain is ClassicInvestor.
Competitors already ship almost every feature that could superficially define this product. Any pitch built on a single capability is already answered by someone with a live product and paying users.
“AI monitors your thesis.” · “Chat with company filings.” · “Keep an investment journal.” · “Version-control your thesis.” · “Connect your brokerage account.” · “Get alerts when the thesis changes.”
Helm Terminal already combines linked portfolios, thesis pillars, automatic event monitoring, exact source evidence, prioritised actions and tax-aware context. subThesis already combines version-controlled theses, assumptions, source-linked claims, decision records, scenario branches and outcome review. Journalytic already captures decisions, predictions, probabilities, feelings, checklists, opportunity cost and behavioural feedback.
The gap is not a missing screen. It is the absence of a dependable link between what an investor knew, what they believed, what they did, and what happened. Each competitor owns a segment of that chain and stops.
Gold = no reviewed product was publicly evidenced as reaching these links from authoritative company evidence.
The name is the position. The classic discipline of fundamental investing — read the filing, write down what must stay true, size the position, review honestly — is not new. What has never existed is a system that keeps that discipline intact across years and proves whether it worked.
A permanent, evidence-anchored record of how an investor decides. Every conclusion keeps the filing it came from, the assumption it rested on, the risk that could break it, the position it justified, the events that followed and the outcome that resolved it.
Not another portfolio news brief. Not another Git-backed thesis notebook. Not another chat-with-filings box. Not another dashboard of financial tables — those are table stakes and already well supplied.
A point-in-time investment case: the evidence available then, the claims drawn from it, the investor's assumptions and confidence, the risks and invalidators, the valuation and expected return, the decision and the alternatives rejected, the capital committed, and the resolution rule that will later settle it.
Every resolved case improves the next one. Calibration, failure patterns and management track records accumulate into something a competitor cannot copy by shipping a feature — because it is the customer's own history.
Exact citations · thesis pillars and assumption tracking · portfolio context and alerts. These are now expected. They buy us the right to compete; they do not win.
Who it is for. Self-directed investors with holdings spread across accounts who want a daily, portfolio-aware answer to “what changed and what needs action?”
What it does well. Low-friction brokerage connection, whole-portfolio visibility, ETF look-through, exact sourced evidence, daily prioritisation, action-oriented monitoring, and tax intelligence that runs all year.
Where it stops. It answers “what changed in what I own?” far better than “what did I believe, why did I act, and what did I learn?” No publicly evidenced thesis version history, original decision record, process-versus-outcome review, or investor calibration.
Match the portfolio-aware monitoring and the exact citations — then connect every alert to a versioned assumption, the prior decision, the position size, the expected outcome and the later review. Helm tells you what moved. We keep why you were there in the first place.
Who it is for. Thesis-driven fundamental investors and independent researchers who want one persistent research record instead of fragmented notes and spreadsheets.
What it does well. Research continuity, version control, evidence-linked claims, scenario branches, structured risks, decisions and outcomes — with an open, exportable, Markdown/Git-backed history.
Where it stops. It preserves a company thesis well, but does not publicly demonstrate portfolio-wide risk governance, capital-weighted calibration, opportunity cost, behavioural analytics, executive commitment tracking, or cross-company learning.
Adopt the open, portable, versioned research model — then extend it into portfolio allocation, resolved outcomes, personal calibration and cross-company evidence relationships. subThesis is the benchmark to beat, and the closest thing to a direct competitor in this set.
Who it is for. Investors who want to improve their decision process, recognise behavioural patterns, and separate repeatable skill from luck.
What it does well. Decision journaling, probabilistic predictions, emotions, checklists, consequences, opportunity cost and behavioural feedback — the richest process capture in the set.
Where it stops. No primary-source company ingestion, no exact filing evidence, no integrated financial analysis, no automatic thesis monitoring. It can explain the investor's process but cannot independently reconstruct the company evidence that should have informed it.
Build Journalytic-quality process capture directly into the evidence and portfolio workflow. Journaling must never be a separate destination the user has to remember to maintain — that is precisely why most investment journals die.
Four further products compete directly for the same user. Each is operationally simpler, and each stops earlier.
Does well: simple setup, clear thesis pillars, continuous event scanning, understandable Supported / Challenged verdicts.
Stops at: the verdict. Thin decision history, limited calibration, limited source-text depth.
We differ: use verdicts only as summaries. Keep the exact evidence, the changed assumption, alternative readings and the resolution criteria underneath.
Does well: conviction journal, structured assumptions, exit criteria, AI challenge, bias scanning, thesis-health dashboard, audit trail.
Stops at: source-level provenance. Its bias framing narrows part of our proposed whitespace, so this one is worth watching closely.
We differ: anchor the same discipline to authoritative evidence rather than to self-report alone.
Does well: mobile-first, low friction, thesis attached to each holding, goals and timelines, no broker login required.
Stops at: lightweight tracking. No research depth.
We differ: a different user entirely — we should not chase this simplicity at the cost of the record.
Does well: measurable assumptions, numeric thresholds, explicit invalidation rules, background monitoring, risk limits.
Stops at: the threshold alert.
We differ: keep the numeric rigour, then carry the breach into a decision and a resolved outcome.
| Platform | What it makes table stakes | Where it stops |
|---|---|---|
| Fintool | Fast AI search across filings and transcripts with cited answers and company timelines | Retrieval and synthesis. Acquired by Microsoft in April 2026 — raising distribution risk for generic “chat with filings” |
| AlphaSense · BamSEC · Tegus | Enterprise document search, filing redlines, broker research, expert transcripts | Institutional research workflow |
| TIKR | Global coverage, estimates, screening, valuation, ownership, transcripts | Research and monitoring, not personal calibration |
| Koyfin | Terminal-style dashboards across fundamentals and macro | Data exploration, not thesis evolution |
| Simply Wall St | Accessible visual fundamentals at global scale | Standardised visuals, not exact evidence or decision records |
The lesson: we should not compete on the number of financial tables, the speed of generic document chat, or the polish of an ordinary company dashboard. Those are necessary and already well supplied.
Full central and operational · Part narrow, manual or secondary · No evidenced absent · Unk insufficient evidence — Unknown must never be read as No.
| Product | Research / evidence | Financial analysis |
Portfolio tracking | Brokerage sync | Event monitoring |
AI analysis | Source citations | Alerts |
|---|---|---|---|---|---|---|---|---|
| Helm | Full | Part | Full | Full | Full | Full | Full | Full |
| subThesis | Full | Part | Part | Part | Full | Full | Full | Part |
| Journalytic | No | No | Part | No | No | Unk | No | Full |
| MyThesis | Part | Part | Part | No | Full | Full | Part | Full |
| SYNESI | Part | Part | Part | Unk | Part | Full | Unk | Full |
| Horyzon | Part | Part | Full | No | Part | Full | Unk | Unk |
| ThesesWatch | Part | Part | Part | Unk | Part | Unk | Unk | Full |
| Product | Thesis creation | Thesis monitoring |
Assumption tracking | Decision journal | Risk monitoring |
Outcome tracking | Post-invest review | Bias feedback |
|---|---|---|---|---|---|---|---|---|
| Helm | Part | Full | Part | Unk | Part | Part | Unk | Unk |
| subThesis | Full | Full | Full | Full | Full | Full | Full | Unk |
| Journalytic | Part | No | Part | Full | Part | Full | Full | Full |
| MyThesis | Full | Full | Part | Unk | Part | Unk | Unk | Unk |
| SYNESI | Full | Full | Full | Part | Part | Part | Full | Full |
| Horyzon | Full | Part | Part | Part | Part | Part | Part | Unk |
| ThesesWatch | Full | Full | Full | Unk | Full | Full | Part | Unk |
Thesis creation is Full for six of seven — it is fully commoditised. Monitoring is Full for five — “monitor my thesis” is a crowded promise. But financial analysis is Full for none, decision capture is uneven, and bias feedback is Full for only two. The right-hand side of this table is where the market thins out.
| Product | Evidence | Analysis | Thesis | Decision | Portfolio | Monitoring | Outcome | Review | Full |
|---|---|---|---|---|---|---|---|---|---|
| Helm | Full | Part | Full | Unk | Full | Full | Part | Unk | 4 |
| subThesis | Full | Part | Full | Full | Part | Full | Full | Full | 6 |
| Journalytic | No | No | Part | Full | Part | No | Full | Full | 3 |
| MyThesis | Part | Part | Full | Unk | Part | Full | Unk | Unk | 2 |
| SYNESI | Part | Part | Full | Part | Part | Full | Part | Full | 3 |
| Horyzon | Part | Part | Full | Part | Full | Part | Part | Part | 2 |
| ThesesWatch | Part | Part | Full | Unk | Part | Full | Full | Part | 3 |
| Product | Its workflow, end to end |
|---|---|
| Helm | position and portfolio state → thesis pillars → monitored evidence → prioritised action. Stops after the action. |
| subThesis | evidence and research → versioned thesis → assumptions and risks → decisions → monitoring → outcome review. Stops before portfolio-wide calibration. |
| Journalytic | investor thought and emotion → decision or prediction → consequence → behavioural review. Never starts from company evidence. |
| MyThesis · ThesesWatch | thesis assumptions → automated monitoring → verdict or threshold alert. Stops at the verdict. |
| SYNESI | thesis and assumptions → AI challenge → monitoring → review. Stops before source-level provenance. |
| Horyzon | portfolio → thesis note → company update → mobile insight. Stops at lightweight tracking. |
| P | Capability | The user problem it solves |
|---|---|---|
| P0 | Canonical evidence & provenance layer | Investors cannot verify AI claims or reconstruct what was known at decision time |
| P0 | Integrated First Pass workspace | Too much time collecting data before deciding whether a company deserves deep work |
| P0 | Versioned thesis, claim and assumption model | Static notes hide how thinking changed |
| P0 | Dynamic risk register | Important risks disappear into narrative and are never monitored |
| P0 | Decision and expected-outcome record | Investors forget why they acted, then rewrite history |
| P0 | Event-to-assumption monitoring | Ordinary alerts create noise and never explain relevance |
| P1 | Portfolio and allocation context | Securities are analysed individually but capital is allocated at portfolio level |
| P1 | Monitoring inbox ranked by decision relevance | Users cannot tell important changes from routine information |
| P1 | Outcome resolver and review workflow | Decisions stay open indefinitely, preventing learning |
“Here is the filing sentence that changed your mind in March, the assumption it rewrote, the position you sized because of it, the event that tested it in August, and whether you were right — measured against what you said would happen.”
Every feature in this briefing can be copied. A competitor can ship citations in a quarter and thesis versioning in two. What cannot be copied is five years of one investor's evidence, decisions, outcomes and calibration, connected end to end — because it is not a feature, it is an accumulation.
Sources. Public product material as of August 2026 — helmterminal.dev · subthesis.app · journalytic.com · mythesis.ai · synesi.app · horyzonapp.com · theseswatch.com · fintool.com · tikr.com. Screenshots captured from live products. Assessments use a conservative public-evidence rubric; they are analytical outputs, not vendor-reported statistics.
Prepared for partner distribution. Confidential.