Platform

Context

What agents generate and learn from

Every agent and every person works from the same operational context: your entities, your procedures and the history behind each case. It is built while the work is done, so what one agent learns, everyone knows.

A connector gives an agent data; the context layer gives it an understanding of your operation. It is also why the second agent costs a fraction of the first.

Built while doing the work

The systems you connect, the cases your agents carry, the calls your people make.

  • Your systems, structured

    Connect the ERP and its records come across as entities: customers, suppliers, orders, items, work orders, with the relationships between them and the history behind them. The system of record stays exactly where it is.

  • Every case leaves something behind

    The supplier that answered in two days, the line that went short, the exception and what was done about it. The operational history that never had a field in the ERP now has somewhere to live.

  • Corrections become procedure

    Someone changes an agent's call and says why. The reason is captured, proposed as a change to the procedure, and from then on it applies to every case of that kind.

Teams, purchasingPO-90214 · line 4
  1. PO agent

    Festo moved line 4 of PO-90214 to 28 Mar. WO-4412 needs that part on the 24th, so the line stops.

  2. PO agent

    Festo has run 2.1 days late on what it confirms since January, so I would read 28 Mar as the 30th. RS Components has the part and took a rush in two days in February.

  3. M. Røed

    Split it. Never hold a whole order for one part.

  4. PO agent

    40 pcs stay with Festo, the rest goes to RS on PO-90401, released. Planning has the new dates and the frame agreement price holds.

  5. Saved as procedure v14 · every date slip from here, on every order

What is in there

One picture of the operation, joined across every system it touches.

Every system you connect teaches it something

A connector carries more than data. What comes through it teaches the platform how your operation behaves.

  1. Structure from the system itself

    Master data, code lists, order types, the names your company uses. The model takes its shape from your setup rather than from a template, so an agent reads a picking list the way your people do.

  2. Behaviour from real transactions

    Which supplier delivers on the date it confirmed, which customer orders at month end, how long that operation really takes against the routing. The pattern behind the field.

  3. The second system joins the first

    An order in the ERP, a case in the CRM and a thread in the inbox stop being three records. They become the order, with everything around it, and an agent working one side can see the others.

  4. Live on the record

    Context points at the record where it lives, so a date changed in the ERP is in the context the same minute, and agents act back through the system's own business functions.

The second agent costs a fraction of the first

Most of the work in putting an agent live is context, and context is built once.

  1. First agent

    The first one lays the ground

    It maps the entities it touches, turns the way that work is done into a procedure you can read, and puts the tools it needs into the shared library. Our team is beside yours for it.

  2. Next agent

    The next one starts already knowing

    Your entities, your suppliers, your procedures, the tools already built and the history of what happened. All that is left is its own job: its queue, its rules, its channels.

  3. Next interface

    Interfaces come out connected

    Describe the view your team needs and it is built on the same context: live records, the agreements behind them, what agents have already done on each one. Connected in its first version.

Shared the moment it exists

One copy, read by every worker on the job and every tool you point at it.

What Riff is not

  • Not an ERP replacementRiff sits on top of your systems of record.
  • Not a general-purpose assistantAgents have specific operational roles.
  • Not a workflow toolAgents have judgment and context, not just trigger-action rules.
  • Not an analytics platformThe context layer serves operations, and your BI tools plug into it.

We leave capability, not dependency

Our forward-deployed team works as part of your organisation from day one, deploys the first agents, and stays as the backstop until your own capability exists. Your builders learn on real use cases with our team alongside, the platform carries the standards, and your domain experts end up owning, extending and building agents around your own strategic goals.

More of the platform

One workflow, agents live in weeks.

Book a demo