Messages & email
Team chats, email threads, review notes and approvals.
Slack · Teams · Gmail · Outlook
Boutique data advisory · Established companies
AI developers and researchers pay to learn from real business workflows. Your team’s records may qualify. We find buyers, negotiate the terms and run the deal for you.
$0 upfront. Our success fee applies when an agreed deal closes and pays you.
Head Managing Partner’s
prior M&A deal experience
In investment banking.
Now applied to your data deal.
Deal skills. Operating experience. Direct access.
Explore the opportunity
Explore a planning range in three short steps. See what shapes it, then decide whether to talk. No email or upload needed.
A broad range with stated assumptions
Factors that could raise or lower it
Your answers stay in this browser
The tool uses published buyer-program pricing as market context. It has not reviewed your records or any buyer bids. See how the range works.
Illustrative gross licensing ranges, before fees, costs, and taxes. Actual results can be $0 or outside the range. No buyer has reviewed your records.
The asset already exists
Your own work can show how real jobs get done. Start with the systems you use—not a finished dataset.
Team chats, email threads, review notes and approvals.
Slack · Teams · Gmail · Outlook
Process guides, working models, reports and revised drafts.
Drive · SharePoint · Word · Excel · Notion
Work orders, support tickets, project changes and repair notes.
Project, job and support systems
Risk reviews, quality checks, exception notes and work logs.
Finance, risk and operations tools
Code reviews, issue threads, test results and change history.
GitHub · GitLab · Private repositories
Approved meeting transcripts, training videos and task walkthroughs.
Meeting tools and training libraries
You control the scope
Approved specialists anonymize the agreed buyer package. We review their reports and evidence before you approve release. Any specialist access to raw records is separately approved.
Customer lists, third-party-owned files, personal details, credentials and privileged advice stay out, along with anything you exclude. Mixed records need review.
Read the data and privacy answers ↗Why Azoth
We connect commercial experience with the technical work needed to deliver your data deal.
The partners
Direct access to the people who lead your deal.
Head Managing Partner
Leon spent seven years in investment banking, working on $7 billion in M&A deals—the buying and selling of companies.
He then served as Head of Risk and Investment at a healthcare group that grew by buying and combining businesses. He later became COO and helped grow the group into one of the top five in its U.S. sector.
His work spans compliance, governance, data security and the day-to-day demands of running a business. At Azoth, he brings that experience to finding value, negotiating terms and checking how each deal is carried out.
Partner · Technology and delivery
Giorgio brings six years of software engineering experience across fintech and banking, including work with the central bank of Latvia. His work spans modernizing legacy systems, integrating business platforms and managing large-scale data warehouses, combining hands-on development with leadership of engineering teams.
His background also includes coordinating releases, quality assurance and acceptance, giving him a practical understanding of what it takes to move complex technical work into production. At Azoth, he brings that delivery experience to assessing business systems, understanding their data and preparing it for commercial opportunities.
Your data. Your choice.
Choose labs that test AI against human work, or buyers that build and improve models. We pursue only the routes you approve. Both use anonymized data.
A lab compares an AI model’s answer with work your team has already done. The aim is to see where AI matches people—and where it falls short.
License selected data to AI companies, including frontier model builders. They may use it to train, improve or test models—within the terms you approve.
We compare fixed-term and perpetual rights where available. Perpetual rights have no fixed end date for the agreed data. Future records are agreed separately, and we explain what remains usable after a term ends before you sign.
We test interest with summaries you approve. A research-only deal does not allow model training. The signed license sets the buyer, permitted use, term and sharing limits.
How we get your deal done
We lead the process. Specialists handle the agreed technical work. You make the key decisions.
We review your business and work history, then agree what to explore, what stays out and which buyers to consider.
A clear starting scope.
We assess records, rights and buyer feedback. Where evidence supports a range, we explain the value, fees and key gaps.
A first view of value and the key gaps.
We approach suitable buyers with approved summaries and seek competing offers where possible.
Buyer feedback and any offers.
You approve the price, permitted uses, license length, exclusions and responsibilities before the deal proceeds.
A deal with clear terms and responsibilities.
Approved specialists prepare the data. We review the agreed anonymization checks before you approve release, then track delivery and payment.
Your deal record, review findings and any open items.
Future opportunities
Selected clients may qualify for ongoing programs covering fresh records or agreed updates. We review buyer fit, then agree the scope, workload, fees and payment schedule separately.
A first deal does not guarantee repeat payments.
Start with a conversation
Tell us about your business, then choose a time to discuss the fit.
$0 upfront. We agree our success fee before work starts. No raw records are needed for the first call.
Provide a short description. Do not attach files or grant system access.
Before we talk
What stays private.
Which route you choose.
How you get paid.
Your data, your terms. Clear answers before you choose a deal.
Already have an offer? Talk to us ↗Agreed use of selected, anonymized records that show how your team works. Think decisions, project notes, process guides and solved problems—not customer lists or vendor-owned data. A license sets out who may use them, how and for how long.
Yes. Anonymization is required before the buyer can use the data. Approved specialists remove or change identifying details and check for other identity clues. We review the agreed checks and flag gaps before you approve release. The buyer may know which company it contracts with; that is separate from the anonymized dataset. Removing identifying details reduces risk; it is not a zero-risk guarantee.
Yes. Our research-only route lets approved labs compare AI results with human work. It does not allow model training or fine-tuning. Any testing service must follow those limits too. We agree which labs and testing services may see the data, and any sharing or publication rules, before release. A new use needs a new approval.
Yes. We can approach model builders, research-only buyers or both, based on your choice. For a model-building deal, the contract states whether training, improvement or testing is allowed. Selecting a route on this page does not grant any data rights.
Customer lists, identifying personal details, account numbers, passwords, private legal advice and records you have no right to license stay out. So does anything outside your approved scope. We look at your own work records and assess whether useful parts can be separated from mixed files.
Approved specialists handle collection, anonymization and transfer. Any raw access they need is limited and approved first. We review their reports and the agreed evidence against your scope, then flag gaps. Your Deal Review Summary separates what we checked from what a provider reported. It is not a technical certificate of perfect anonymity.
We can seek one-year research options, subject to buyer availability. Other buyers offer fixed-term or perpetual licenses. Perpetual means the agreed rights have no fixed end date. We compare the terms and explain what happens to the files, test results and any derived work after the term. A one-year license does not, by itself, erase all results created during that year.
Select clients may qualify for recurring data programs after working with us. These can pay for fresh records or agreed updates. We review fit and buyer demand, then agree the scope, schedule and fees before the program starts. This is not automatic. A perpetual license alone does not create recurring payments.
There is no upfront fee. We agree a success fee before work starts. It applies when an agreed deal closes and pays you. Any outside costs need your approval first. Your engagement sets the scope, rate and any fees for later deals or ongoing programs.
Explain the business, connect us with the right people and make the key approvals. We handle buyer talks and coordinate the work. You do not need to clean years of files to start. Any later input from your team is agreed before that work begins.
You can. We run your side: find the useful records, compare buyer routes, negotiate the payment and rights, manage the work and check the agreed process. Judge the result by what you keep after fees, the rights you retain and the work taken off your team.
Some may pay us for sourcing and coordination. Those fees are separate and do not automatically reduce your fee. We explain relevant arrangements and material conflicts before the affected decision. We do not advise a buyer on how to negotiate against you.
No. You choose whether to accept a data license. Appointing Azoth gives us an agreed job; it does not itself approve data use or a release. Your engagement sets any continuing obligations. We do not promise a buyer offer or a minimum payout.
The work is already done
You do not need a data project. Start with a clear view of the deal.