Agent Patterns

What an agent engagement looks like.

Four representative agent patterns, anonymized and drawn from how agents actually get scoped, modeled, and built for relationship-led businesses. They show the shape of the work. The situation, the system, the math it was designed against, and what the customer keeps.

We publish design targets here, not invented results. Named, attributable outcomes will replace these stories as pilots complete. That standard is part of what you are buying.

Customer Business / Pipeline

The consultancy where every Monday started with a scramble

A two-partner consultancy with strong delivery and a pipeline that lived entirely in the founders’ heads. Follow-up went quiet every time a project heated up, and new business arrived in waves. The partners described it as feast or famine, because that is what it was.

What we built

A pipeline agent that reads the full relationship history before any touch. It resurfaces every open conversation older than two weeks, drafts the follow-up in the lead partner’s voice with the last real exchange referenced, and prepares a Monday pipeline brief ranked by margin, not just deal size. Every draft waits for a two-minute approval.

The math it was designed against

Scoped against the math of conversations that go quiet. The model behind the build assumed that consistently touching every open thread, instead of the half that survive busy months, is worth one to two additional engagements a quarter for a business this size. The agent exists to make that consistency free.

What the customer keeps

The repo, the prompts, the customer-memory store, and the playbook for tuning it. The business runs it on its own accounts.

Customer Business / Reactivation

The advisory practice sitting on six years of dormant customers

An advisory practice with a CRM full of past customers nobody had spoken to in years. Everyone agreed the list was worth money. Nobody ever got to it, because reactivation done properly takes research per name, and the partners’ hours went to billable work.

What we built

A reactivation agent that mapped the dormant list, watches for buying signals like role changes, funding, and company news, and drafts re-engagement notes that pick up the last real conversation rather than announcing we should catch up. A quarterly brief shows which relationships warmed and why.

The math it was designed against

Scoped against the cheapest revenue a relationship-led business owns. The working assumption, drawn from how dormant lists behave, is that a worked list responds at a meaningfully higher rate than cold outreach because the trust already exists. The build is judged on conversations reopened and engagements signed from the list, not emails sent.

What the customer keeps

The signal watchlist, the relationship map, and the full system. The dormant list intelligence compounds for the business, not for us.

The free gap audit shows which of these patterns fits your business, as a PDF in minutes.

Get a Free Gap Audit

Customer Business / Retention

The advisory team whose customer touchpoints stopped at delivery

An advisory team with strong project work and weak relationship rhythm after delivery. Customers left happy, then heard nothing until the partner needed a referral or a renewal conversation. The relationship was valuable, but the touchpoints depended on memory.

What we built

A customer-retention agent that turns engagement notes, milestones, meeting history, and outcomes into customer memory. It prepares quarterly relationship reviews, flags upcoming milestones, drafts useful check-ins, and reminds the partner why each customer may have a reason to talk now.

The math it was designed against

Scoped against expansion, referral, and renewal conversations that usually die from neglect. The model treats a retained or expanded customer relationship as cheaper revenue than net-new acquisition, because the trust and context already exist.

What the customer keeps

The customer-memory layer, milestone rhythm, review briefs, and every workflow, running on the business’s own stack. The intelligence about the relationships compounds for the business.

Customer Business / Proposals & Content

The agency where everything bottlenecked on one person

A founder-led agency where every proposal and every LinkedIn post waited on the owner. Presence mattered, proposals won work, and both kept losing to delivery. The cost was invisible because it never showed up as a line item, only as silence.

What we built

A proposal and content agent drafting from discovery notes, engagement history, and the owner’s own writing. Proposals arrive as reviewable drafts within a day of the discovery call. A week of posts in the owner’s voice gets prepared each Friday, sources cited, ready to approve or kill.

The math it was designed against

Scoped against the owner’s hours, the most expensive in the building. The build is modeled on returning the proposal-writing and content hours to billable work and closing, and on proposals going out while the conversation is still warm instead of two weeks later.

What the customer keeps

The voice model, the proposal library, and the system. The owner’s voice stays the owner’s asset.

Your version of this starts with the free gap audit. It shows which of these gaps your business is carrying, or bring your numbers straight to a Growth Audit Call.

Get a Free Gap Audit Book a Growth Audit Call

Next Step

See the customer-growth gaps before competitors close them.

Start with the free gap audit or go straight to a working session with Jake.

Get My Free Gap Audit Book a Growth Audit Call

Email Jake directly at jake@northsignal.studio