Skip to content
AnonymizedFilm · 4 min

They asked for agents

An industrial holding group asked us for AI agents: agents that would find prospects, reach out to them, and hand its sales reps leads that were ready to go. We met the people doing the work first. What the company needed was a redesign of who does what, with agents where the work was ready.

Chapters

At a glance

What we did, and what it meant

Situation
A holding group of industrial businesses, where one team sold to and served the customers. There was a CRM, and nobody logged their activities in it.
Mandate
AI agents that would find prospects, reach out to them, and hand the sales reps leads that were ready to go.
Our role
Interviewed both teams, redesigned who sells and serves, and built the agents, the automation, and the reorder model ourselves.
What we did
  • Listed what each person on both teams was responsible for.Commercial lane
  • Built an AI agent and a workflow automation for service and quotes, with one person reviewing every quote.Software and agents
  • Built a model that sets a reorder window for every customer, and opens a task when an order is late.Data science
  • Split sales into setters and closers, with agents that find and research leads, and a power dialer.Commercial lane
  • Logged every call and email in the CRM automatically.Data engineering
Outcome
One commercial function, on one system, with clear roles, and additional revenue from the same team.
What it meant for the owner and lender
A customer who stops ordering is now a task with an owner, and every call and email is on record: an early warning for a lender, and less key-person risk.

What we found

A process and organization problem, not a technology one

One team sold to and served the customers. An inside team worked the phones: calls, service questions, and quotes. A field team visited prospects. There was a CRM, and nobody logged their activities in it.

Before building anything, our principals interviewed both teams and listed what each person was responsible for. Three things came out of it.

  1. The inside team had room in its day, and most quotes were for standard parts.
  2. Nobody could see what the field team was doing.
  3. People were in the wrong seats.

The plan

Two rules, three jobs

Remove manual work for the least spend. Put people where they earn the best return.

  1. Serve the customers you have

    In a week, we built an AI agent and a workflow automation for the service and quote work. The agent now handles most of it, and one person reviews every quote before it goes out. That freed people the company was already paying for.

    Software and agents

  2. Keep them coming back

    A model learns each customer’s order history and sets a reorder window, updated every month or quarter. When an order does not come in on time, a salesperson gets an alert and a task opens in the CRM.

    Data science

  3. Win new ones

    The people freed from service joined sales, split into setters, who call all day, and closers, who take the opportunities a setter has qualified. Agents find and research each lead before a setter sees it, a power dialer keeps the calls coming, and every call and email is logged in the CRM automatically.

    Commercial lane

The result

Same team, new seats

Six months in, the company ran one commercial function, on one system, with clear roles. The activity log nobody filled in now fills itself, and the company brought in additional revenue on top of what it usually brings in.

For the owner
More revenue from the team it already had, with every role clear and every activity on record.
For a lender
A customer who stops ordering is no longer silence. It is a task with an owner, and an early warning.
For anyone underwriting it
Every call and email sits in the CRM, so the business keeps the knowledge and owns its relationships. That is less key-person risk.

They asked for agents, and they got them. But the agents were only the power source. The redesign was the value creation.

More work

Other engagements

AnonymizedFilm · 11 min

The economics of the stockout

A specialty retailer carrying thousands of items, each planned with one forecast number per month, typed into a workbook.

Anonymized

A lender-backed transformation in temperature-controlled logistics

A temperature-controlled third-party logistics company in a lender-backed transformation, with much of its operation still run on paper.

Anonymized

A system of record for a pharmaceutical outsourcing facility

A pharmaceutical outsourcing facility moving its commercial work onto one system.

Next step

Bring us a situation

If you back a company like this one, or lend to one, a principal replies within one business day and holds a first call.

Bring us a situation