Should Halden buy its way into the US?

A European dental materials maker has hired a strategy firm to plan its US entry. A week before the steering committee, the case team's draft recommends a $105M acquisition. Everything needed to test that is already in the firm's files, in eight documents that nobody has read side by side.

Scroll through to watch the files become a context graph, see where the recommendation breaks, and see why the answer differs from what a general AI assistant would give you.

and one thing the partner knows that no file says

Eight files, one entry decision

Halden Dental wants to know how to enter the US. A week before the steering committee, the case team holds a proposal, a market model, 24 interview notes, a draft pre-read recommending an acquisition, three past engagements from the firm's archive and a partner's email.

Each file says something specific

Every page, slide and tab is read, and the figures, claims and precedents that bear on the decision come out as typed entities, each one tied back to the passage it came from.

They connect into a context graph

Typed relationships show how the evidence depends on each other: what argues for the acquisition, what prices it, who buys through a formulary, and which past engagement looks like this one. That connected structure is a context graph, and it is what Tenuto Labs builds for each client.

The market is not as fragmented as the slide says

The pre-read's 84% counts dentists, including associates and public-health dentists. The model counts spend: DSOs buy 21% of the category today and drive two-thirds of its growth to 2030, and 6 of the 9 DSOs interviewed buy from a formulary. The acquisition case rests on the shrinking share of the market.

And its proof case did not hold up

The pre-read cites Solenne's 2022 board recommendation deck: coverage in 12 months. The firm's 2024 retrospective says 22 of the 54 acquired reps left within a year, a third of the revenue went with them, and Solenne wrote down $21M. The partner's 2024 email told the practice to read the follow-up.

The number nobody wrote down

Twice, a rep or an agency held the customer relationship, and twice about a third of the revenue walked when ownership changed: 29% at Norvik, 34% at Solenne. Where a contract with a central buyer held it, at Pellam, it stayed. No file says that is a rule. No AI model, however capable, can read a rule that was never written down, which is why this question goes to the partner next.

The partner answers one question

The pattern becomes a single question for Elena Marsh. Her answer takes two sentences and becomes a rule in the context graph, linked to the business case it reprices and the precedent it points to. The next case team that asks gets her answer without needing her time.

Question generated from the context graph

In both past engagements where a rep or agency held the customer relationship, about a third of the revenue left when ownership changed. Where a contract with a central buyer held it, it stayed. Is that a rule, and what does it mean for Larchmont?

Answer, Elena Marsh

It's a rule. I price anything a rep holds at two-thirds, and I'd rather win the person who signs the formulary than buy the person who visits the office. For Halden that means DSO contracts now, and Larchmont only if it works at two-thirds, which today it doesn't.

E.M.

Added to the context graph: 1 rule · 2 relationships · source: partner, Oct 2026

Same question, asked twice

Ask the hero question both ways. A general AI assistant searching the engagement files retrieves passages that point in different directions, and it has no way to know the partner's rule. The context graph follows the evidence and returns an answer with its reasons.

A general AI assistant retrieves

PPTXPre-read: acquire Larchmont; 84% of US dentists are independent.reads as yes

PPTXPre-read, citing Solenne 2022: coverage in 12 months.reads as proof it works

XLSXModel: DSO share of category spend rises from 21% to 28% by 2030.reads as a trend, not an answer

finds the passages, not the reasons

It can summarize these. It cannot know the two-thirds rule, because no file contains it.

The context graph answers

No, not now.

The 84% counts dentists, not spend. DSOs drive two-thirds of category growth, and most buy from a formulary. XLSXDOCX
The proof case kept 66% of its acquired revenue and took a $21M write-down. The pre-read cites the 2022 deck, not the 2024 retrospective. PPTXPDF
Priced at two-thirds retention, with the line pulls Larchmont's own rep predicted, the acquisition earns about $1.7M a year on $114M. XLSXDOCX
Pellam is the closer analog: central buyers with formularies, and its direct contracts renewed. PPTX
Counter: contract direct with five DSOs, let distributors fulfill, and revisit Larchmont at two-thirds pricing. Marsh rule

Your turn: choose the entry route

Pick a route. Capital, year-3 EBITDA and how much revenue sits on contracts Halden holds all move with it, because each is derived from the route and the partner's rule rather than typed in.

Contract direct with DSOs (Elena Marsh's route): Capital committed $14M; Year-3 US EBITDA $5.3M; Revenue on contracts Halden holds 70%. Win the formulary at the DSOs driving two-thirds of category growth, with distributors fulfilling and covering independents. It adds $8M of capital to the distributor-only route and puts 70% of revenue on contracts Halden holds. If only three DSOs sign, year-3 EBITDA is about $1.3M.
Acquire Larchmont (The pre-read's route): Capital committed $114M; Year-3 US EBITDA $1.7M (pre-read $15.8M); Revenue on contracts Halden holds 0%. The pre-read's $15.8M assumes every Larchmont relationship survives the sale. Price rep-held revenue at two-thirds and add the line pulls Larchmont's own rep predicted, and it is $1.7M on $114M. Owning a distributor also puts at risk the national distributors Halden needs everywhere else.
Distributors only (The Norvik route): Capital committed $6M; Year-3 US EBITDA $4.3M; Revenue on contracts Halden holds 0%. Cheapest start, and sound for year 3, but Halden would hold no customer relationships and little of the DSO growth. Norvik took that route, and converting later cost it 29% of the revenue it was trying to own.

this is what Tenuto Labs builds

A context graph built on your engagements and your partners.

This is a worked example on a synthetic engagement. The companies, people and figures are invented; the method is the one we use. In an engagement we model your practice, connect past engagements, interviews and models into a context graph, and put a few short questions to your partners each day, so the firm's judgment stays in the system when the people who hold it move on.

General AI assistantContext graph, built by Tenuto Labs

Where the knowledge livesWhatever it retrieves or you paste in, for one conversationEvery engagement, interview and retrospective, connected across the firm's history

Expert judgmentOnly what someone wrote downCaptured from partners as rules, like pricing rep-held revenue at two-thirds

Showing its workA fluent answer you have to checkAn answer with its path, each step tied to a page or a cell

When a precedent changesHope someone remembers to mention itAdd the retrospective once, and every recommendation that cited the deal is flagged

Start a conversation

The eight files behind this example

PDFCorwin Hale proposal and SOW, Halden US entry
XLSXHalden US market model v4
DOCXCustomer and channel interview notes
PPTXSteering Committee 3 pre-read, draft v3
PPTXNorvik: US entry 2019 and conversion review 2023
PPTXPellam: US entry 2021 and three-year review 2024
PDFSolenne acquisition retrospective, September 2024
EMLElena Marsh email to the practice, October 2024

Worked example on a synthetic engagement. The context graph was drawn from the eight invented files listed. All companies, people and figures are fictional; market statistics are illustrative estimates.

See the legal example