Harper Relay / Context Relay

Context survives the handoff.

Context Relay keeps the source, promise, urgency, account state, and last human decision attached to the work. For agencies and multi-account operators, that means freshness checks, escalation rules, and source ownership stay visible before support turns into confident noise.

source owner attached
freshness checked
escalation rule visible
last human decision preserved

The current public next step still runs through the live route-attached message desk at /contact?from=agency-knowledge-bot. The surrounding trust path stays attached through /privacy?from=agency-knowledge-bot and /about?from=agency-knowledge-bot. This page sells the governed rollout shell; it does not claim there is already a client-facing analytics suite or unsupervised support fleet.

Context relay

The promise travels with the work.

source namedlive
freshness checkedlive
handoff owner setlive
reply path clearlive

Glance state

96%

context attached

01

nudge

02

report

03

launcher

Source stays named
Promise stays attached
Handoff stays human-aware

Live route proof

Live route

/contact?from=agency-knowledge-bot#message-desk keeps the rollout context attached through the live Harper Relay desk.

Collection note

The desk asks for first name, last name, email, message, and human verification, keeps the context scoped to the handoff, and states the no-tracking, no-secrets, and no-portal boundaries before submit.

Trust route

Privacy and About stay route-attached so storage, ownership, and escalation context do not fall back into generic company copy.

Trust routing

Sensitive questions still need a named operator-approved owner, so the rollout path stays honest while the context layer stays usable.

Context Relay Standard

Explain source ownership, freshness, and escalation before the handoff gets noisy.

The trust move is governance first. If the buyer cannot see who owns the answers, how they stay current, and when humans take over, the rest of the page reads like commodity bot setup.

01 / Source ownership

Name the answer source before the bot ever starts sounding confident.

Each client or location needs a visible source-of-truth map: which docs, FAQs, policy notes, and operator instructions are allowed to answer, and which questions still belong to a human by default.

02 / Freshness checkpoints

Keep launch changes and stale docs from quietly turning into support debt.

The governed layer shows when answers were last reviewed, what changed since the last pass, and which account needs an operator read before the assistant keeps repeating old language.

03 / Escalation defaults

When the bot is unsure, sensitive, or outside scope, it should stop cleanly.

Policy exceptions, account-specific disputes, sensitive edge cases, and newly launched offers should escalate on purpose instead of being flattened into a polite guess.

04 / Multi-account oversight

Give operators a readable oversight layer without making them babysit another dashboard.

The useful view is simple: which brand is fresh, which brand is drifting, what escalated recently, and where a human needs to step in next.

05 / Rollout discipline

Keep the public promise bounded to branded support, freshness, and human follow-through.

This offer is not an all-purpose agent marketplace. It is a managed support layer with branded behavior, governed source control, and operator-owned escalation rules.

Proof Surface

The proof should look like governed rollout notes, not another shiny support dashboard.

These are sample oversight moments that show the operator-readable shape of the service. The point is to make freshness, account drift, and escalation visible without inventing vanity metrics.

Sample source map

One client brand gets a named answer set before rollout starts.

Support articles, approved FAQs, service boundaries, and escalation-only topics are separated up front so the assistant never has to guess where truth lives.

Freshness slip

The review note says what changed and what still needs a human look.

New offer launch, pricing change, policy edit, or stale location detail all get a visible status instead of hiding behind vague 'AI updated' language.

Escalation rule

Uncertainty has a named owner and a next move.

Sensitive requests, unusual support asks, and account-specific exceptions route to the right human instead of sounding impressive and wrong.

Oversight note

The agency can see which account needs attention without a separate analytics wall.

The useful signal is freshness state, escalation reasons, and operator follow-through. Decorative conversation charts are not the product.

Where It Fits

Keep the service tied to real rollout buyers, not to generic agent-platform ambition.

Harper Relay should position this as a managed support and knowledge-governance service. The rollout makes sense when the operator already feels drift, stale docs, or brand inconsistency across accounts.

Boutique agencies

Resell a branded support layer without reselling fragile chatbot theater.

Use this when the agency needs one governed support pattern it can adapt across clients while keeping source ownership and escalation rules explicit.

See Lead Manager

Multi-location operators

Keep location answers current while routing policy and exception cases to people.

This works for businesses with multiple brands or locations that need answer consistency, local nuance, and clear human takeover when the question gets sensitive.

See scheduling layer

Support-heavy retainers

Turn knowledge upkeep into a recurring management lane instead of one-off bot setup.

Freshness reviews, escalation tuning, and account-level governance make the recurring service clearer than selling an assistant once and hoping it stays true.

Request rollout review

Verified Inputs

who owns the source of truth for each client or brand
how freshness gets checked before stale answers spread
which questions escalate by default and to whom
what the agency can review without a separate analytics layer

Boundaries

unsupervised support autonomy sold as if no human ever needs to step in
generic white-label chatbot language with no source ownership model
conversation charts or scorecards that do not change an operator decision
multi-agent platform claims that outrun the actual rollout and review process

Start Here

Keep context moving without making support own every answer.

Use Context Relay when the business already has docs, customer promises, offer notes, and account knowledge scattered across the day. The message desk keeps the handoff human-owned while Harper Relay quietly builds the source map, escalation rule, and follow-up trail behind it.

context mapfreshness checkhuman takeoveraccount governance

Bring This

which client accounts or locations are first in the rollout
where the current support answers drift or go stale fastest
which requests must escalate instead of being guessed at
who owns approvals when docs, offers, or policies change

Trust Notes

The public next step is the live message desk at /contact?from=agency-knowledge-bot, not a bot signup or hidden operator console.
This page sells the governed rollout shell. It does not claim a live public analytics suite, autonomous support fleet, or finished client-facing dashboard.
Privacy notes and company context stay attached at /privacy?from=agency-knowledge-bot and /about?from=agency-knowledge-bot while the broader public owner boundary remains visible.
Review Lead Manager

What Leaves The Review

A useful first pass should hand back concrete rollout guidance, not a vague promise to tune the bot later.

operator-ready outputs

Source map

A named answer set for the first account or location, plus the topics that stay off-limits until a human approves the next pass.

Freshness watch

The docs, policy notes, or offer changes most likely to create drift, with the next review point written down instead of implied.

Escalation sheet

The exact questions that stop the assistant, the human owner who takes over, and the account-risk note that should stay visible.