Skip to content
MTK Solution

Business operations for every unit on one platform

Each division and team runs its own accounts, opportunities and bids, while leaders see the whole company in one number. AI agents propose the next step; people approve it.

Business operations: company data, unit tree, targets, approval, people

One company, many systems and spreadsheets

Where multi-unit sales operations lose time and accuracy.

One company, many systems and spreadsheets
TodayOn one platform
Visibility or autonomyDivision heads need the whole division, while each unit keeps managing its own records.Every record belongs to a unit, and each role sees its own records, its unit, its division or the whole company.
The same customer, twiceUnits create accounts that already exist in another unit or in the company CRM.Company accounts are read in place and linked at creation, so duplicates are caught before they spread.
Targets in slidesTargets live in presentations, and every reporting period needs queries and currency conversion.Targets are set once at the top, split down the unit tree and reported in one currency.
A new unit, a new systemEach new business or industry adds another tool to connect and maintain.Industry differences are configuration on one shared platform.

How the system works

Company data flows in once, every record sits in the unit tree, and agents work inside the same permissions as people.

  1. Unit tree and access scopes
  2. Company data, linked not copied
  3. KPI builder without SQL
  4. Targets that cascade
  5. Pipeline, bids and approval gates
  6. Operations agents

Illustrative. Numbered points match the key features below.

Key features

Six capabilities, each marked with where it stands today.

  1. Console

    Unit tree and access scopesIn development

    Records carry their unit in a company, division, unit and team tree. Each role sees its own, unit, division or company records, and any record can be shared with a person, role or unit.

  2. Console · data layer

    Company data, linked not copiedIn development

    Accounts, contracts and financials are read from existing systems through one read-only layer, checked for duplicates and linked by their source ID. Nothing is written back.

  3. Console

    KPI builder without SQLIn development

    Metrics are defined by object, measure, date, filters and source, starting from a template of prebuilt definitions.

  4. Console

    Targets that cascadeIn development

    Yearly targets are set at the top and allocated down the unit tree by period, with versions. Amounts convert to one reporting currency at the monthly rate; the original is kept.

  5. Console

    Pipeline, bids and approval gatesIn development

    Leads become opportunities, bids, contracts and renewals through configurable stages, bid or no-bid scoring, and approval gates routed by unit and threshold.

  6. AI agents

    Operations agentsRoadmap

    Agents flag opportunities without a next step, write weekly KPI reports per unit and read bid documents, returning each finding with its source page.

Agents propose, people decide

Agents act with the rights of the person using them and only write proposals. A finding without a source is rejected.

Agents propose, people decide
AI agentWhat it doesDecided by
Sales assistant Drafts activity updates, next steps and new leads from a conversation.The salesperson applies or dismisses each proposal.
Pipeline hygiene Flags stale or overdue opportunities and those without a next step.Unit directors act on each flag.
KPI reporting Writes the weekly achieved-versus-target view for each unit, with variances and risks.Leaders read the report and decide.
Bid review Reads bid documents and returns findings, each citing the file and page.The bid manager runs the approval gate.

What a company measures

Each organisation records its own baseline before the pilot starts.

  • Weighted pipeline and coverage by unit
  • Share of open opportunities with a next step
  • Duplicate accounts stopped at creation
  • Days to close a reporting period
  • Agent proposals applied versus dismissed

Where to start

Each step ends with a decision on the next, measured against the baseline recorded first.

  1. AI Readiness Assessment

    Two weeks to map units, data sources and the KPIs that matter most.

  2. AI Pilot

    Six weeks with two or three units on your data: unit tree, linked accounts and a first target plan.

  3. Roll-out

    More units, the pipeline and bid gates, then agents once their proposals are measured.

Plan this pilot

Plan a pilot of business operations with MTK

Start with two or three units and one target plan, measured against today’s reporting.