Most B2B wholesalers and distributors use AI to cut costs. The larger opportunity sits on the revenue side of the P&L: in our experience, companies in this sector can reach 5 to 10% incremental topline growth within three years by putting data, AI and automation to work in sales and marketing.
That was the argument of our webinar on August 25. Avaus CEO Emma Storbacka and Ilona Vigren, Director of Strategic Business Development at Avaus, were joined by Sara Nygren, Business Developer at Dustin, and Carl Command, Group Manager Analytics at Dustin. More than 70 registered participants heard what one year inside Dustin’s Commercial AI Program has produced, and what it took to get there.
Watch the full webinar recording on demand here.
Why wholesale and distribution is built for AI-driven growth
Emma opened with why this sector, of all sectors, is so well placed to grow with data and AI. A transactional sales model usually means omni-channel selling, multiple buyers inside each account, and many touchpoints. All of that generates data. The customer base runs from a long tail of small, occasional buyers to large contract-bound accounts with negotiated prices and dedicated logistics. Sales teams know transactional selling well but are often new to omnichannel and solution selling. Marketing and CRM exist but carry unused upside. And many companies in the sector grew through acquisition, then captured synergies on the cost side while leaving growth operations unaligned across markets.
The result is a set of performance pockets that data and automation can address: acquiring customers with targeted outreach instead of broad campaigns, raising average order size and margin through product mix and cross-sell, expanding the share of each customer’s wallet into new categories, adding services on top of low-margin product sales, and cutting the cost of sales by shifting transactions to self-serve channels. Footprint expansion is often the largest of these: the customer already buys from you in one category and buys the next one from a competitor. As data maturity grows, new revenue streams open on top, such as retail media, circularity services and predictive fulfillment.
The reason to act now is AI itself. A team of four used to have the capacity of four people. When repeatable work shifts to agents quarter by quarter, the same team’s capacity keeps growing, which shortens the path from a long list of use cases to automations in production.
Inside Dustin’s Commercial AI Program, one year in
Dustin is one of the largest IT partners in the Nordics and Benelux, selling hardware, software and services to everything from small businesses to large enterprises and the public sector, through both digital channels and a relational sales force. The company handles around 2 million orders a year.
CAP, the Commercial AI Program, is Dustin’s multi-year program for accelerating its commercial transformation through data, AI and automation. It concentrates on marketing and sales, aims at gross profit uplift, and moves step by step toward autonomous commercial operations. Three objectives guide it: winning market share through better prospecting and acquisition, winning customer loyalty through retention and share of wallet, and winning in commercial excellence through more effective sales and marketing. The first year went into foundations: the operating model, the ways of working, and the first waves of use case development.
The operating model is the machinery that produces use cases
Ilona framed the program as two things built in parallel: a portfolio of data-driven use cases, and the operating model that produces them. The operating model is where technology investments succeed or stall. Avaus structures it in eight areas: targets, strategies, structures, use case implementation, capabilities, change management, governance and measurement. The framework doubles as a maturity map, a traffic-light view of where the organization is strong and where to focus next. We covered the same territory from the leadership angle in our Stockholm breakfast recap on building an agentic enterprise.
Sara showed how the framework runs in practice at Dustin. Governance splits into a strategic steering group for direction and priorities, an operational steering group for dependencies, blockers and adoption, and a core team that builds. The program runs on a quarterly cadence with two big room planning sessions a year: the core team drafts a six-month hypothesis plan, both steering groups pressure-test it in one room, the plan gets validated with the business, and the second session further evaluates feasibility, designs a solutions and locks the scope for the coming quarter. Work is organized in seven streams and four interconnected backlogs covering use cases, capabilities, the operating model itself and change management.
Measurement respects the reality of B2B sales cycles. The financial effect of an automation arrives slowly, so the program tracks the number of live automations as its leading indicator, alongside program-specific KPIs and general commercial performance.
Use cases start concrete and scale into automations
A use case is the Data-Algo-Action loop applied to one specific, defined business problem. In CAP, the terminology is precise: a use case is an end-to-end commercial process, designed for scale from the start but proven small. Once it works, it scales into automations by adding markets, offerings, customer types and channels.
Dustin’s offering expansion model shows the pattern. It went live in one market, for one software offering, on three channels: LinkedIn, email and relational sales. That proved the workflow. Scaling the same use case across three markets, three offerings and five channels generates 45 automations from one build. This means the discussion about use case prioritization only needs to happen once, but can still result in dozens of automations.
To see what a candidate portfolio could look like for your own business, our use case navigator lets you filter more than 300 commercial AI use cases by business challenge and industry, including distribution and wholesale. It’s free and needs no sign-up.
Three example use cases
Carl walked through three examples spanning the funnel.
Finding new prospects. Machine learning models trained on Dustin’s transactional history, enriched with third-party data, identify ideal customer profiles among companies that aren’t yet customers. Every channel then targets the same prospects with the same offering, so relational sales, paid marketing and email stop working separate lists.
Offering expansion. For existing customers, a probability model picks the next best offer: which software or service a hardware customer is most likely to need next. The recommendation activates across CRM for the sales team, marketing channels, and the web, so a customer who logs in sees the same offer their account manager is calling about.
Proactive order management, the agentic case. Today an inside sales team manually monitors orders for risks: products going end of life, distributor problems, delayed shipments, repricing needs. Dustin is building an agentic process that scores each order for risk and value, proposes the next best action per order, and can execute parts of the flow itself, from product swaps to proactive customer outreach. Against 2 million orders a year, that is a material amount of margin and customer satisfaction to protect. For the broader shift this represents, see our piece on the rise of AI agents.
Key learnings after running the program for a year
Each speaker named a different lesson.
- Carl’s: good use cases usually die between departments, when one team prioritizes the work and another doesn’t. CAP bakes cross-department prioritization and resourcing into the operating model up front, so a build that hits a technical hurdle or needs a new channel doesn’t restart the argument.
- Ilona’s: operating model work is never finished, and shouldn’t be. The program enters its second year with a target of 2 to 3x higher productivity, a moving target on purpose.
- Sara’s: unlike classic programs, there are no work streams to check off, only ways of working that get replanned every quarter.
The Q&A surfaced two more. Asked what surprised her most, Sara pointed to how fast the stakeholder group grew, to over 50 people across sales, marketing, IT and analytics, which complicates coordination but builds the shared ownership the program needs. Asked about data location and European sovereignty, Carl was matter-of-fact: Dustin’s contracts require its data to be stored and processed in Europe, so the program works with the large American model providers under EU hosting. And the next hurdle, per Ilona, is adoption: getting the daily sales process to pull full value from what the program ships. That theme will follow the program for years, and it connects to what we argued in our recap of the AI strategy webinar: the plan matters less than the system that executes it.
Common questions
Why is B2B wholesale and distribution well suited to AI-driven growth?
The sector combines rich data with untapped demand for it. Transactional, omni-channel selling across multi-buyer accounts generates large volumes of usable data, while data-driven ways of working still trail B2C retail. In Avaus’s experience, that gap translates into 5 to 10% incremental topline growth attainable within three years.
What is the difference between a use case and an automation?
A use case is the Data-Algo-Action loop applied to one specific, defined business problem, built as an end-to-end commercial process and proven in one market, offering and channel set. An automation is each scaled instance of that use case. One use case scaled across three markets, three offerings and five channels generates 45 automations.
How do you measure a commercial AI program when sales cycles are long?
Track leading indicators first, above all the number of automations live in production and the effort each new use case takes to build. Financial KPIs follow as the portfolio matures and sales cycles complete. Dustin runs this as a stack: three program objectives, program-specific KPIs, general commercial KPIs, and lead indicators underneath.
Where will your growth come from?
AI is delivering efficiency in most organizations. The growth question is still open, and it is the subject of our executive workshops this autumn in Stockholm (September 15) and Helsinki (September 30), with second dates in November. Half a day, a curated peer group of around ten senior commercial leaders, complimentary and application-reviewed. Apply here.
If any of this raises questions about where your organisation stands, get in touch and we can work through it together.
Related reading
- Case: Commercial AI Program at Dustin
- What it takes to build an agentic enterprise: lessons from the Stockholm breakfast
- Why your AI strategy shouldn’t exist: and what to build instead
- Data Driven Voices: omnichannel automation at Dustin, with Emma Spjut
Contact us
PS. Watch the full webinar recording and explore the Use Case Navigator, operating model assessment, and Data–Algo–Action framework.