Where the word comes from

In a paper dated autumn 1978, Gifford Pinchot III and Elizabeth S. Pinchot wrote that large corporations were "suffering from size". They proposed giving selected employees "a status within the corporation akin to that of the entrepreneur within the larger society", with the corporation's technological, financial and informational umbrella still over them, and called these intra-corporate entrepreneurs "intrapreneurs".

The paper already raised the questions this guide answers. Rewards were to be shared between the company and the intrapreneur "in a well-defined and equitable way". A would-be intrapreneur would present and defend a business plan to get money. A product no other division would buy could become a new division, or a new corporation largely owned by the parent.

The American Heritage Dictionary (fifth edition) now defines the intrapreneur as "A person within a large corporation who takes direct responsibility for turning an idea into a profitable finished product through assertive risk taking and innovation."

Step 1: Decide what the program is for

In our view, the majority of corporate innovation programs, startup partnership announcements and corporate venture capital programs end in PR without results. The cause is a mismatch of incentives: what leadership looks for from the program differs from what building something new requires.

Steve Blank, writing in 2018 about the programs he and Pete Newell had helped build, saw the same split: incubators and hackathons "got coffee cups and posters", while senior leadership "sat up and took notice" when operating groups changed how they worked. Global Corporate Venturing (GCV) wrote in August 2026 that there is now "more of an expectation of results", and that success in venture building is measured "by new business revenues generated, new product launches and new strategic options created".

Write the program's purpose as one sentence with a number and a deadline in it, and give it to every team, judge and sponsor: "Within four years of launch, two ventures earn revenue inside our business units." Leave ideas submitted and demo days held out of it. They count activity.

Step 2: Agree what the sponsor accepts

In our view, building new businesses from inside takes commitment, dedication and time, and large companies seldom give those to the leaders who run their programs. Results come, but they come in time.

GCV wrote that executives "still struggle to answer basic questions around team and programme size, investment commitment, timings and how many failures are within the boundaries of acceptability." The C-level sponsor answers those four in writing, before launch.

Step 3: Change the approval flow before launch

Blank wrote in 2018 that "all the established procedures and processes in a large company are primarily designed to execute and support the current business model." In our view, a venture that has to pass through those structures unchanged will not be deployed on time or on budget. Four functions need a conversation before the first team starts.

In our experience, procurement pushes back on a venture's suppliers; that is its mandate. What a venture needs from these functions, as we see it, is speed of execution, room to fail at parts of what it tries, room to work in grey areas without breaking the law, and the authority to decide quickly who it buys from. Ask each of the four leaders for exceptions, within limits. Where they cannot give them, build the venture a separate structure whose rules come from its market.

If your business units mainly need to buy from startups, that is a different model. In venture clienting, as GCV describes it, teams "focus on becoming customers of startups without taking minority equity stakes". We compare the options in Venture clienting, corporate accelerator or CVC: Which one your company is ready for.

Step 4: Pick the teams and protect their time

In the same 2018 post, Blank split the people who do innovation in large organisations into innovators, who invent new technology, products, services or processes, and entrepreneurs, who have worked out how to get innovation adopted and delivered through the organisation's existing procedures. Any successful innovation program, Blank wrote, "requires an individual or a team with at least these two skill sets." Two months earlier Blank had put the gap plainly: "it's rare that the smartest technical innovator is the most successful entrepreneur."

In our view, the usual mistake is to staff a venture team with innovators and hope the entrepreneur turns up. Name that second person at the thesis stage: someone who has already pushed something new through this company's procurement, legal, IT and budget cycle. Give them the landing plan, the receiving owner's sign-off and the list of approvals the venture will need, and pay them for the landing. We set out their first ninety days in Innovators Shift.

Then protect their time. Blank saw teams trained outside their organisation go home and find that "execution of existing priorities took precedence." We would give a venture team time that is more patient than normal corporate reporting allows, and pay to match. Write down each member's share of time on the venture and put it into their line manager's targets, or the day job wins.

Step 5: Set gates and release money in tranches

Blank also described funding committees that judge ideas without measurable milestones: "the best presenter and flashiest demo usually win." A gate asks for evidence, and a tranche ties the money to it.

In our experience, the conditions that matter have to be set before the money moves. Once a venture team has its funding it is independent, and the sponsor and the coaches can enforce very little. So write each gate's conditions into the funding decision: what has to be true for the next tranche, who decides, and how the venture's leads break a tie.

Step 6: Decide where a successful venture lands

GCV's 2025 survey suggested that only 25% of companies build ventures to spin out, and a lot of respondents expected their successful ventures to be absorbed back into the parent. ADP Ventures builds for that ending on purpose. Its ventures are always destined to be integrated back into the business, Usman Khan, then a senior vice president there, told GCV in October 2025, and the businesses it had graduated into ADP "collectively today do nine figures ARR combined."

Blank named one reason landings fail: budgets were planned years in advance, so the organisations they saw had no slack to finish and deploy even validated prototypes. Our first change would be to name the receiving owner at the thesis stage, a business unit leader who agrees that the venture lands with them if it hits its milestones and who sits in its gate reviews. Then put the landing year into that unit's budget, build to the parent's standards for architecture, security and support early, and report in the unit's own measures. We took ADP's model apart in Innovators Beacon.

Spin a venture out when the parent's structures would hold it back. In our view, it then needs a new structure with flexible rules, pay to match and patient time; the Pinchots allowed for that in 1978 with a new corporation largely owned by the parent. Before it gets money, write down how its founders break a tie and which conditions travel with each tranche.

Step 7: Choose which ventures get a serious bet

Score every venture the same way, from evidence, before anyone pitches. Ask every member of a venture team the same questions separately, and treat their disagreements as evidence too. At every gate, add one question: who on this team has got something adopted here before, and what did it take? If the honest answer is nobody, we would not give the venture more money, however good the prototype is.

Our venture team assessment works that way. Every team member answers separately, in about 90 minutes, and can save and come back. It scores four dimensions (Current State, Scalability and Fund Raising, Road to Revenue, Competitive Landscape) on five levels from Emerging to Market-Ready, and each program gets its own instance with its name on the reports.

Where we fit

We design and run intrapreneurship and venture programs around a company's culture: leadership alignment, venture frameworks, team training and day-to-day delivery. A program typically takes 3 to 6 months from design to first rollout and starts with cohorts of 10 to 20 teams. Our corporate work includes Siemens Startup School (co-led since 2022), advice to Ford Turkey, Akbank+ at Akbank LAB (advisor and lead mentor of its full-time intrapreneurship and spin-off program) and sessions in the five-day AI and Innovation Hackathon of the Financial Academy and Medgulf.

If an earlier program stalled, start with the Strategic C-level workshop: a half day to a full day in which your leadership team assesses readiness, finds out why earlier programs stalled and decides the next step. Programs and the workshop are quoted after a discovery call. The venture team assessment starts from 4,500 EUR in the first year for up to 25 teams, VAT included.

What each model buys

We use the definitions VentureClient.org published in January 2026, edited by Gregor Gimmy, which set venture clienting apart from accelerators and corporate venture capital. Its own summary is short: venture clienting focuses on adoption and use, CVC on equity investment, and accelerators on startup development and preparation.

VentureClient.org dates the term to BMW R&D in 2014. In its account, general functions such as partnerships, technology scouting and procurement could not consistently bring in startup solutions of the quality and volume needed, and CVC could reach only a limited number of startups, because venture financing is capital-intensive and selective. BMW formalised its answer as the BMW Startup Garage, launched in 2015.

A fourth option sits inside the company: venture building, where the company founds and usually controls new ventures itself. VentureClient.org calls this internal corporate venturing. Our design guide for intrapreneurship programs, next in this cluster, covers it.

Why programs end as press releases

In our experience, corporate innovation programs, startup partnership announcements and CVC units end in press releases more often than in results. The cause we see is incentives. What leadership looks for and what the work requires are two different things: the work needs commitment, dedication and time, and the people in charge of these programs are rarely given that time.

So the label on the program matters less than two things the company decides before launch: how it rewards the people who run it, and whether its own rules let a startup move at a startup's speed.

The four functions decide which model you can run

In our view, all three models run into the same four functions, and a company that wants its people to build or adopt new things has to sit down with each of them before launch. HR has to design a separate way to reward the people doing it. Procurement will resist parts of the work, because that is its mandate. Legal protects a large organisation with contracts that a startup cannot carry. Finance attaches approval steps and signatory powers that slow decisions down.

What each model needs from them is different, and that difference is the readiness test.

Read the table row by row with each function in the room. The column with the fewest answers of "not yet" is the model your company can run in the next twelve months.

What leadership has to accept

Venture clienting puts a young company inside your operations. VentureClient.org's principles describe ventures as operating under extreme uncertainty, with strong growth and exit pressure and limited organisational resilience. That makes their products valuable and fragile at once, so a business-unit owner has to want the product and accept that the supplier may change or disappear.

A corporate accelerator prepares startups and leaves adoption to later. VentureClient.org notes that adoption by the sponsor, if it occurs, usually comes after the program, and that because accelerators often carry significant cost and organisational effort, whether one feeds adoption needs to be evaluated on purpose. If nobody owns that question, the demo day becomes the result.

CVC typically puts capital into minority stakes. When a CVC unit chases financial return and the business units want a product adopted, the two aims can pull apart, and someone has to decide which one wins. VentureClient.org treats that conflict as something to manage.

All three need patience. In our experience, results from a new technology field or a new product come in time, and they come later than a normal corporate reporting cycle expects.

If your model puts money into a team

We helped a venture spin out of a corporate entrepreneurship program, and the investment committee funded it because the founders had field experience and passion, even though we had seen signs that the team did not work together. The two remaining co-founders had no agreed way to break a tie, so whoever pushed harder won. Before the investment, the team did what had been agreed. After it, the founders were independent, we could no longer do anything, and the team collapsed.

Write the decision rules, the milestones and the conditions for each tranche of money before the money moves. In our experience, a rule that was not agreed before the money moves cannot be enforced after it.

Signs your company is not ready yet

• Procurement cannot say how long it takes to buy from a company founded two years ago.
• No business unit can name a problem it would pay a startup to solve in the next twelve months.
• The accelerator's measure of success is the number of startups or the size of the demo day.
• The people running the program are rewarded on the same reporting cycle as everyone else.
• An investment committee is ready to fund a team on experience and passion while the warning signs about how the team works are on the table.

A readiness check for one meeting

1. Write down the business problem in one sentence, and the business unit that owns it.
2. Put HR, procurement, legal and finance in one room and go through the table above, row by row.
3. For every "not yet", ask for an exception within limits. Where none is possible, decide whether the team needs a separate structure with its own rules, pay and timeline.
4. Choose the model whose column needs the fewest exceptions.
5. Agree what the first result is (a product in use, an adopted startup after the program, or an investment made on written terms) and the date leadership will look at it.

Where we fit

We design and run intrapreneurship and venture programs around a company's culture: leadership alignment, venture frameworks, team training and day-to-day delivery. A program typically takes 3 to 6 months from design to first rollout and starts with cohorts of 10 to 20 teams. When leadership first has to decide which model fits, the Strategic C-level workshop takes a half day to a full day to assess readiness, find why earlier programs stalled and decide the next step. Our past work includes Siemens Startup School, which we have co-led since 2022, and Akbank+, Akbank LAB's full-time intrapreneurship and spin-off program, where Ozan Sonmez is advisor and lead mentor. Both the programs and the workshop are quoted after a discovery call. See how we design intrapreneurship and venture programs.