“Spend limits, approval gates, and an audit trail on every write.”
That’s now on Synter’s homepage — in the hero, next to the Get started button, in the same font as the rest of the pitch.
It’s the first time in fifteen weeks of weekly tracking that any tracked competitor has used Ott’s exact differentiation vocabulary on a product surface. Synter is the closest architectural match to what Ott does — write-capable MCP, AI agents, 21 platforms. When the closest match starts selling the compliance lane in your words, two things are true at once:
- The market finally sees that “audit trail” is the feature to fight over.
- Copying the words is not the same as shipping the product.
Here’s what a real audit trail requires — and where Synter’s claim stands against it, verified live this week.
Where the Claim Actually Lives
First, a tour of Synter’s surfaces, because the claim has no fixed address:
- Homepage hero: “READ ACCESS FREE · 21+ platforms · Spend limits · Audit trail”
- Homepage feature list: “Spend limits, approval gates, and an audit trail on every write” — plus “Human-in-the-loop iMessage & WhatsApp mobile approval gates”
- The MCP page: “Synter executes, with approval gates on every change” — in the FAQ schema. The words “audit trail” don’t appear on the MCP page at all.
- Pricing page: an “Audit Logs & Governance” row in the plan comparison table.
Four surfaces, three phrasings, zero documentation. There is no audit-trail spec, no compliance documentation, no jurisdiction feature, and — after 240+ blog posts — no finance or regulated-vertical content anywhere on the site.
That’s what marketing language looks like before it becomes a product: it moves around. A shipped audit trail doesn’t need to live in the hero copy. It lives in the product.
What a Real Audit Trail Requires
An audit trail is an engineering term with an engineering definition. For a regulated agency, six properties are non-negotiable:
- Actor attribution. Every entry names the actor — a person, a named tool, an API key. “An agent made a change” is not an audit trail; it’s a confession that nobody was accountable.
- Immutability. Append-only. No edit, no delete, no “oops” button. If anyone on the vendor side can retroactively change the record, the record is worthless in front of a regulator.
- Before/after values. “Budget changed” is a log line. “Budget changed from $10,000/day to $500/day at 14:03 UTC by [name], and here’s what the agent proposed” is evidence.
- Independent storage. The log can’t live inside the system being written to. When a Meta account gets restricted, the record of what happened in it has to survive outside it.
- Jurisdiction awareness. The same action means different things under the FCA, MiCA, ESMA, or an Alberta license. A trail that can’t tell those apart can’t prove compliance for any of them.
- Retention and export. A regulator’s request is answered with files, not a UI session. Seven days of history is not record-keeping — it’s a browser history.
Now score Synter’s shipped surface against those six. The pricing page tells you most of what you need to know: “Audit Logs & Governance” is a paid-tier feature. Solo keeps 7-day history. Scale keeps 90 days. Only the sales-led Custom plan gets “Unlimited / SIEM Export” — the tier you have to book a demo to see the price of.
An audit trail that’s a pricing tier isn’t an audit trail. It’s an upsell. And a 7-day window doesn’t satisfy the accountability standard the regulator has been repeating all year — more on that below.
The Regulator’s Bar Is Already Set
This isn’t hypothetical. The pieces are on the table:
- FCA (June 24): Nikhil Rathi, the FCA’s chief executive: “Accountability for regulated activities and outcomes must remain clear.” The speech that opened the agentic-AI conversation for UK financial services.
- IAB Tech Lab (July 30): AAMP 2.3 — “Trust and Safety Move from Optional to Enforceable.” And the line every regulated agency should have bookmarked: agentic autonomy “that respects a hard approval boundary is a critical requirement for regulated advertisers.”
- FCA enforcement (Aug 18–24): a senior-manager ban under SMCR, a former SVS Securities CEO fined and banned, a consumer warning on mini-bonds and loan notes — three actions in one week, all about who is accountable for what.
- FCA Handbook API (Aug 6): the regulator made its rulebook machine-readable and free. Its own FAQ admits the API “does not provide past versions of Handbook content.” The regulator ships current and future rules — the historical record of what your agency actually did is your job. It always was.
That’s the convergence: the regulator demands accountability, the standards body demands approval boundaries, and the enforcement cadence is running. Meanwhile the AI-disclosure regime went live in three jurisdictions this month — IAB v2, EU AI Act Article 50, California SB 942. The compliance surface for finance agencies is expanding, and “7-day history” is on the wrong side of every one of those obligations.
The Gift in Synter’s Own Comparison Page
Synter had its busiest week on record last week — ten posts in four days, including a direct attack on Windsor.ai. “Why Pay for Ad Data Plumbing When Reporting Is Free?” The CEO’s argument: Windsor “is completely read-only — it cannot create ads, optimize budgets, diagnose fatigue, or manage campaigns.”
One problem: that’s false. Windsor’s MCP has been write-capable since May 22 — campaign updates from Claude and ChatGPT, budget edits, pausing and enabling Google and Facebook Ads, and creating Facebook campaigns since late June. Windsor’s own MCP page says it plainly: “The end of the read-only era — if Claude flags a leaking budget at 2 AM, you fix it right there.”
So the two horizontal competitors with the loudest claims about each other are both write-capable, and neither ships a proof layer. The write era is real — the market learned it months ago — and the only thing nobody has built is the record.
Six Questions Before You Trust an Audit-Trail Claim
Whatever you evaluate next — Synter, Optmyzr, Supermetrics, Windsor, or the platform pitching you in-house — run the claim through six questions:
- Can anyone delete or edit a log entry? (If yes, it’s not an audit trail.)
- Does the log attribute every action to a named actor?
- Does an entry carry before/after values, or just “something changed”?
- Where does the log live — and does it survive an ad-account restriction?
- Which jurisdiction does each entry know it operates under?
- What’s the retention, and what does the export look like?
If the answer to any of those is “we’re building it,” “it’s on the roadmap,” or “that’s on the enterprise plan,” you’ve found the difference between copy and product.
That’s the whole point: everyone can now say “audit trail.” The bar for a regulated agency is that the trail survives contact with a regulator.
What Ott Ships
Ott was built around the proof model, not the pitch. Activity Logging records every Meta ads campaign change — pauses, budget edits, creative swaps, bid changes, review and account status — timestamped, attributed to the person or tool that made it, with before/after values, stored independently of Meta so it survives account restrictions, unified across every client, brand, and account, overlaid on performance charts, and exportable in about 90 seconds. Jurisdiction tags know the difference between an FCA-regulated broker and a MiCA-licensed exchange.
The write era demands a write log. Ott is the write log — for changes made by humans, by automation tools like Madgicx and Revealbot, and by AI agents.
“Spend limits, approval gates, and an audit trail on every write.” Great sentence. Everyone’s writing it now. Only one platform’s audit trail survives a regulator’s request — and it’s the one that was built to prove, not to promise.