AI in Private Equity Deal Sourcing: Motherbrain and the Associate's Job

AI in private equity deal sourcing does the finding and the first sift. It tracks far more companies than a deal team could follow, picks up signals such as a sudden rise in a company's hiring, and ranks possible targets against a fund's criteria before anyone picks up the phone. EQT's Motherbrain is the best-documented example. For an associate, the change is where sourcing work begins: with a ranked list to judge rather than a blank one to build.

Put simply, AI deal sourcing is the use of machine learning on data about companies, people and markets to find, screen and rank investments before a deal team makes contact.

What does AI change in private equity deal sourcing?

Sourcing is the front end of a buyout: choosing where to hunt, finding the companies that fit, and getting to the good ones early instead of waiting for them to arrive as incoming leads. EQT tells associate applicants they will work "from thesis development and origination through to due diligence, execution, and portfolio engagement." AI now does much of the work at the front of that funnel.

Step by step, the swap looks like this:

The machine takes the steps that grow with the number of companies in a market. The associate keeps the ones that grow with judgment.

What an associate does with a ranked list

Starting from the list, not the phone

Quoted on EQT's own ThinQ site in October 2025, EQT partner Chris Litchford, a technology specialist, described tools that automate much of the firm's sourcing by ranking potential targets on a range of criteria. Then he set them against the old way:

"Instead of some of our competitors who might hire analysts for $85,000 a year to cold call and gather that information, [we're asking] if some tooling gives you 85 percent of that… and we can use it across the entire group."

The last clause matters as much as the number. A list built by one analyst sits in one inbox. A list built by software sits where every team in the group can use it, and it keeps updating after the analyst moves on.

What does the associate do instead? EQT gave a concrete answer as early as 2020. In its Ventures and Growth teams, the firm said, every investment professional relied on Motherbrain "to manage the daily deal work", using its algorithms to prioritize which companies to assess. A working day built that way starts from the queue: why did this company score where it did, is the data behind it right, and which three names deserve a call?

A worked example: one week of screening

Bain & Company's 2024 private equity report gives the clearest numbers, from one large investor. That fund's professionals looked at about 10 deals to find one worth a deeper look, and spent a full day on most of those looks, working from seven criteria tied to the fund's strategy. Bain wrote that generative AI can bring the screening time per company down from a day to an hour.

Run that through one associate's week:

• By hand, at one company a day, five working days give five looks. On the fund's ten-to-one ratio, that is half a company worth a deeper look.
• With AI, at an hour a company, and assuming an eight-hour screening day, the same week gives 40 looks and roughly four companies worth a deeper look.

By hand, a day per company

5

With AI, an hour per company

40

Illustrative. The day, the hour and the ten-to-one ratio come from one fund described in Bain's 2024 private equity report; the eight-hour screening day is an assumption.

Nobody screens for 40 hours straight, and a ratio measured on a short list will not hold at the far end of a long one. The point survives the rounding. The bottleneck moves from how many companies you can look at to how well you judge the ones you do.

What the saved days buy

Bain's own answer is that faster screening frees a team "to focus on the more qualitative work involved in analyzing the potential gems that make it through the funnel." EQT makes the same pitch to applicants: associates use AI-powered tools "to move faster, go deeper, and spend more time on the thinking that matters."

In a sourcing team, that thinking has a shape. It is the thesis that decides which criteria the machine screens on, the judgment about which ranked names are real, and the first conversation with an owner, which stays with people.

What is EQT's Motherbrain?

EQT describes it in one line: "Motherbrain is EQT's dedicated team driving the integration of AI and data within the private equity industry." It "supports EQT's entire investment lifecycle", from finding companies to working with the ones the funds already own.

The name covers two things. One is a team, bringing together "expertise in data science, machine learning, user experience, and strategy", with people in North America, Europe and Asia-Pacific. The other is the Motherbrain Platform, the "proprietary tool" where sourcing work lives: EQT says it "redefines how we track deal pipelines, uncover insights, and develop deep sector expertise", and it holds the lessons from past deals in one place.

The person running it is Alexandra Lutz, who joined EQT in 2019 and is Head of Motherbrain. Her route in was not an engineer's: she had been a chief strategy officer, a start-up president and a management consultant, after a first career as an Emmy-winning journalist for ABC News. At EQT she first ran digital growth in the US, helping deal teams develop thematic ideas and source targets.

In a 2024 essay she put the bet in one line:

"Algorithms and machines won't necessarily replace humans, but I do believe that humans with algorithms will outperform humans without algorithms."

From EQT Ventures to the whole firm

Motherbrain did not start in buyouts. It began where the problem was sheer volume: EQT said in 2020 that more than 4,800 start-ups were being founded every day, and "no human can rate them all." From the venture team it moved through the growth fund into private equity.

The venture numbers are a snapshot of the tool's first home. What carried over to buyouts was the working method, and it rests on three ideas.

How does Motherbrain find and rank companies?

Each idea answers a different sourcing question: which companies exist, which ones resemble each other, and what the firm already knows about them.

Tracking companies, not deals

A normal deal process starts when a company comes up for sale and ends when a fund exits. Motherbrain was built the other way round. EQT's 2020 report said it "supports the tracking of company life cycles rather than deal life-cycles", following businesses long before and after any transaction.

Elin Bäcklund, then EQT's Deal Engine Lead, set out the ambition in the same report: "The idea is that we want to model the world. We want to have all the companies out there in Motherbrain." The payoff, she said, was sourcing investment opportunities "rather than being reactive to incoming leads."

Scale is what makes that possible. Lutz wrote in 2024 that the system sits on "a vast proprietary database containing hundreds of millions of data points about companies, people, and the connections between them."

Similarity and ranking

Once a company is in the system, the next question is what it resembles. In 2021 EQT announced that a paper by Motherbrain's data science team had been accepted at EMNLP, one of the leading academic conferences in natural language processing. The method, called PAUSE, turns company descriptions into numbers so the closeness of any two companies can be measured, which helps with jobs such as competitor mapping.

That matters for sourcing in a specific way. If a fund likes one company, similarity search surfaces its peers, its competitors and possible add-ons without an investment professional labeling every pair by hand. EQT's 2020 report described Motherbrain tools for "analyzing peers and competitors as well as sourcing add-ons and analysing markets for M&A opportunities."

Victor Engelsson, then a partner in EQT Growth's advisory team, said in 2021 that the growth fund used Motherbrain and algorithms like PAUSE "to assess and make faster and more substantiated decisions." Ranking is the step after similarity. The tools Litchford described in 2025 rank potential targets by attractiveness across a range of criteria, which is what puts an ordered list in front of the associate.

The deal team's memory

The third idea is the least glamorous and possibly the most useful. On top of external data, Lutz wrote, sits "a collection of all the thought that goes into a private equity investment": "meeting notes, people in common, jotted-down numbers, lengthy powerpoints or not-so-lengthy emails." Motherbrain "lets anyone manage pipelines, add assessments, and track advisors."

Two details from the 2020 report show how that memory was built without extra typing:

• An automated note parser pulled key metrics out of meeting notes, so EQT captured them "without changing users' behavior."
• Contact data was combined with "140k unique connections uploaded by EQT employees", so the firm's network sat in the same place as its data on companies.

For an associate, this is the part that changes the handover. The next person to look at a company inherits every past assessment of it, which raises the bar on what a new assessment has to add.

Which other private equity firms use AI to source deals?

EQT is the most documented case, not the only one. Blackstone and TPG have each built an in-house team, and each has described in its own words where AI fits at the front of a deal.

Blackstone: data science beside the deal teams

Blackstone's data science team describes itself as "an integral part of the firm's investing process, applying data science and machine learning to solve questions that matter to our businesses." Matt Katz, its global head of data science, wrote in 2023 that the team "works with Blackstone deal teams around the world to make better investment decisions."

In 2024 he went further: AI and data science are "often a part of the way we determine which companies to invest in." The sourcing statement came in May 2026 from John Stecher, Blackstone's chief technology officer:

"We're also using AI at the very top of the funnel. Because of Blackstone's scale, we see a lot of deal flow. AI allows us to evaluate opportunities faster, say no faster, and spend more time on what we believe are the highest-quality investments."

"Say no faster" is the phrase to keep. At a large firm most of sourcing is rejection, and a tool that makes rejection cheap changes what a junior spends the week on.

TPG: themes first, then companies

At TPG the order starts with themes. Nehal Raj, co-managing partner of TPG Capital, set it out in December 2025: "Before we make any new investments in a space, we like to identify the themes that we're most interested in, and then we go and find the most interesting companies within those theme areas."

Ayanna Clunis, a TPG partner and its head of operations, described Samantha, the tool TPG created after setting up Lab 39, as a layer "that allows us to interrogate the data", drawing on outside tools such as OpenAI's models and Copilot so teams can research, query information and compare documents.

Tim Millikin, a TPG partner who co-leads its software and enterprise technology investing, said what the tools free his team to do: "focus on theme identification, building relationships, finding companies to transform". He added that "AI can help power each one of those steps along the way."

What do AI sourcing teams hire for?

For an engineer or a data scientist, these teams are a way into private equity that does not run through a banking analyst class. EQT's version reads less like a research lab and more like consulting with code.

A 2023 EQT posting for Motherbrain Labs, based in New York, London or Singapore, described the group as technologists who "partner directly with EQT's deal professionals and portfolio executives to provide data-backed, strategic decision support." It added: "Labs technologists are a combination of applied data scientist, deal professional, information architect, and researcher."

What the posting asked for, in practice:

• Advanced SQL and Python for data analysis, plus four or more years in finance or user-facing data roles.
• Time with deal teams: meeting them to understand the problem, sourcing new data to answer it, and presenting the result in tools such as Hex, Airtable and Softr.
• A taste for simple methods: "we don't build black boxes when a simple algorithm would suffice."

EQT Technology, which is responsible for the platform, describes its team as user experience researchers, software engineers and data scientists. Two neighboring routes in are the forward deployed engineer role and the AI engineer seat at a hedge fund.

Will AI replace private equity analysts in deal sourcing?

EQT, the best-documented case, says no, though the tools are changing which analyst work a firm pays for.

A 2025 piece on EQT's ThinQ site drew the line inside the deal: "Screening and early diligence are ripe for automation, while confirmatory checks and negotiation are not." Most of sourcing's list work sits on the first side of that line. The first call to an owner, and the negotiation that may follow, sit on the other.

Bain, in the report behind the screening example above, was blunt about the intent: "The goal here isn't to fill seats with less expensive robo investors but to make investment professionals smarter and faster at what they do." EQT's digital co-heads, Sven Törnkvist and Petter Weiderholm, wrote in 2025 that "it's still unlikely that machines will fully automate investment decisions or replace human judgment."

What changes is the bar for the people who stay. Törnkvist and Weiderholm also wrote that EQT had "developed a roadmap requiring AI literacy across our team of investment professionals."

For a candidate, that is the useful reading: the associate seat stays, a tool now does much of its list-building, and the AI literacy on EQT's roadmap is something you can show before the first interview. The same question for investment banking analysts turns on the size of banks' entry-level classes rather than on sourcing tools.

How AI sourcing could come up in a private equity interview

EQT describes its associate interview process as "around five stages, combining competency-based interviews with case or analytical exercises." None of these firms says its own AI system is an interview topic, so the questions below are ones a sourcing case might raise, given what the firms say they value: a thesis before a list, judgment over the tool, and AI literacy.

How to prepare

Four steps cover it, and each one leaves you with something to talk about:

1. Read the firms' own words. EQT's Motherbrain page and one of its ThinQ essays, plus one other firm's description of its data team, give you accurate language and the dates that matter.
2. Screen ten companies by hand and time it. Pick a sector, write five criteria, score ten companies from public information, and note how long each one takes. That number is your own version of the day-per-company figure.
3. Repeat the exercise with an AI assistant. Give a general-purpose model the same criteria and compare its list with yours: which companies it adds, which it gets wrong, and which facts it cannot source. Keep the log. How banks use these assistants day to day is covered in ChatGPT and Claude in banking.
4. For engineers, build something explainable. A small similarity search over company descriptions, with a plain account of why two companies score as close, shows the instinct EQT's 2023 posting asked for.

For a structured run-through before the real thing, use the interview readiness tool, and browse the wider guides for what individual banks and firms have built.

The bottom line

AI in private equity deal sourcing now does the finding and the first sift. At EQT, Motherbrain has been in front of every private equity investment professional since 2020, and other large firms have built in-house teams of their own. The associate's week starts from a ranked list rather than a blank one.

What a firm still pays a junior for is the judgment the list cannot supply. A candidate who can show that judgment on a real list, and describe the tools accurately without overclaiming, is ready for the version of the job that remains.

The later stages of the deal are moving the same way: in due diligence AI takes the first read of the data room, and once a deal closes it goes into the value creation plan.