Token Observe vs doing nothing
With three agents, no regulated data and no incident, doing nothing is often the correct decision. This page is about what changes it.
Premature governance has a real cost, and it lands on the people you have.
Match governance to the decisions your agents can make. Routing a request through Token Observe adds a dependency that fails closed: if the gateway is unavailable, that request cannot reach a model. The hosted cloud pilot manages the workspace infrastructure and has shared capacity limits. Optional customer hosting needs your own deployment, database, backup owner and recovery process. In either mode, assign someone to review the evidence and decide what to do when a policy blocks work.
Start with a free 30-day cloud trial if you want to explore supported billing sources and optional routed-agent governance. Create an account, confirm email and create a workspace; no card or automatic paid conversion applies. Customer-hosted evaluations and design-partner arrangements are scoped separately with a named engineer. Both paths are for bounded evaluation, with no availability SLA, SOC 2 or ISO certification, or independent penetration test. Keep production-critical workloads outside the pilot.
Token Observe and doing nothing, row by row
Where these two differ, one row at a time.
Zero. No deployment, no dependency, no on-call decision.
A free 30-day cloud trial, with provider usage charged separately. Optional customer hosting has separately quoted licence terms and customer-operated backups. Routed agents depend on the gateway in either mode.
Whatever the provider console shows: spend by key, roughly, with no owner attached.
Which agent, whose owner, under which policy, approved by whom, at what cost, with the trace id on every response.
The investigation and the record-keeping start on the same day, and the record you want is the one nobody kept.
The record exists from the first governed request, including the blocked ones: a trace id is minted before the verdict, so a refusal is recorded rather than absent.
Nothing recorded. The EU AI Act’s high-risk obligations, Article 12 among them, apply from 2 December 2027 for the uses listed in Annex III and from 2 August 2028 for products under Annex I.
A flight recorder for the governed request lifecycle: a control that helps evidence a clause, not a certification. Retention is an operator setting, and left unset it keeps traces indefinitely.
You find out what is running when the invoice arrives, or when someone leaves.
Five radar evidence sources, and only for what you feed it. What it does not see, it reports as unseen rather than clean.
- beat 01Name the caller
- 1Authenticate
- 2Resolve the agent
Nothing is decided for an anonymous caller. The key resolves to one registered agent with a named owner, and its roles, its ceilings and any stop switch in force are read live from the store on every request, not from a cached copy, so a suspension takes effect on the next call instead of after a redeploy.
- beat 02Start the record, then read what is being sent
- 3Open the trace
- 4Sanitise
- 5Scan
The trace is opened before anything can refuse the request, which is why a blocked call is recorded as thoroughly as one that went through. Only then is the text normalised and scanned, in that order, so the detectors see what the model will read rather than what a human reviewer can see.
- beat 03Decide once
- 6Govern
- 6bIntersect with the named human
- 7Enact
One function returns one verdict: allow, block, or hold it for a named person to approve. Every check lives in that one place rather than scattered along the path, which is what makes it possible to read your rules and know what they do.
- beat 04Send it, and only then spend the money
- 8Route
- 9Call upstream
Routing fixes which provider and model will serve the call, so where an agent has a USD ceiling it is tested against the price you will pay, not the one the caller typed. Routing is also the last step that can still refuse locally: after it, the request has left your network.
- beat 05Govern the answer, then count the cost
- 10Govern the response
- 11Meter and record
A tool call the model proposes is judged before it reaches your agent. Metering runs last and on every ending, including the refusals and the streams a client abandoned halfway, because the tokens were spent either way and a partial trace is still evidence.
Step names and their grouping are the product’s own, and the order is encoded in the evaluator rather than drawn for the page.
When it makes sense to wait.
Three tests, and if all three hold, wait. Can one person name every agent your organisation runs, without asking anyone. Is every action an agent takes reversible by the person who notices it. Has nobody outside engineering asked you a question about them with a date attached. An estate that passes all three is governed adequately by the fact that it is small.
The advice that goes with waiting is cheap: write down the list, in whatever your team already reads. An agent, a named human owner, one sentence of purpose, and which provider it calls. What waiting does not buy is a free option on the past: records that were never kept cannot be reconstructed, and retrofitting logging onto agents that have run ungoverned for a year is the expensive path. That is an argument for starting the record early, not for buying a platform early.
The four things that change the answer.
None is a headcount threshold or an agent count. They are changes in kind rather than in scale, and any one of them on its own is usually enough.
- An agent can now do something irreversible
- A refund, a deployment, an email that leaves the building, a row written to production. Once an action has a business effect, the question stops being what did the model say and becomes what happened, and who authorised exactly that.
- A second provider appears
- The moment the same behaviour has to be constrained on two providers, a rule maintained in two places starts to diverge, and the divergence is invisible until it matters.
- Someone outside engineering asks
- Counsel, a data protection officer, an auditor or a customer’s security team. Their questions are field-shaped — who owns this agent, what may it touch, who approved this action — and they are answerable from a record or not at all.
- Spend is real and unattributed
- The invoice arrives and nobody can decompose it by agent and owner. An explicitly unbudgeted agent can meter at zero, so an unmetered estate and an idle one look identical, and nothing prompts you to check.
Which of the two to put in.
Read the left one first. If two of those describe you, buy the other thing. If the right column describes you, the next step is thirty days with the source on your own hardware.
When to choose doing nothing
- One person can name every agent you run, and every action they take is reversible by whoever notices.
- Nobody outside engineering has asked you a question about them, and no date is attached to anything.
- The work of configuring and reviewing a fail-closed gateway exceeds your current risk. Customer hosting also needs a deployment and backup owner.
- Your workload requires general availability, service credits or independent assurance that the current cloud pilot and optional customer-hosted arrangements do not offer.
When to choose Token Observe
- An agent can now do something that moves money or changes a customer’s record, and somebody would have to explain it.
- You cannot say how many agents you have without asking, or who owns two of them.
- Spend is material and cannot be attributed to an agent and an owner, which is the point where an unmetered estate and an idle one become indistinguishable.
- Someone has asked for evidence with a date on it, and the answer today is that the record was not kept.
If the left-hand column describes you, that is still worth an email: a short reply costs both of us less than an evaluation that ends in the same place.
See the thirty-day evaluationThe other comparisons
The other 6 categories, each opening with the case for the alternative.
We have three agents and no incident. Should we do anything at all?
Probably not, beyond writing down the list. An agent, a named human owner, one sentence of purpose and which provider it calls, kept wherever your team already looks. That costs an afternoon, it is what every regulatory framework asks for first, and it makes the eventual decision straightforward instead of archaeological. Revisit when one of those agents can do something irreversible.
What is the smallest useful step if we do want to start?
Connect one supported billing source in a cloud workspace and review its periods, currencies and collection status. If you also want agent governance, configure one supported client and a scoped agent key. Start policies in shadow mode, inspect the recorded decisions and acknowledge the required backtest before enforcement. Billing alone does not recover historical conversations.
Is Token Observe ready to buy?
You can start a free 30-day cloud trial today, with no card or automatic paid conversion. Customer-hosted licensing, design-partner support and assisted work are scoped separately. The cloud service is a bounded pilot with shared resource and traffic limits. It has no availability SLA, multi-node high availability, SOC 2 or ISO 27001/42001 certification, or independent penetration test. Use it for evaluation and keep production-critical workloads outside it.
Tell us which way you are leaning, and why.
Create a cloud account and workspace. No card or server installation. Model calls and Google BigQuery scans can incur separate provider charges.