The 2014 case on CGI’s SPARK innovation centres has aged better than almost anything else in the book. Re-read in 2026, it is one of the earliest worked examples of what the current research calls Value Design as a structured discipline — and a quiet lesson in how to fund it.
Prof. dr. Régis Lemmens — Future of Selling research programme — May 2026
Somewhere between roughly 2005 and roughly 2015, IT buyers in European industry lost their patience with the slide deck. They had spent a decade investing in new technologies on the strength of vendor presentations, watched a meaningful share of those investments fail to deliver the benefits promised on the slides, and concluded that the slide was the wrong instrument for the decision. One of the integrators we interviewed for From Selling to Co-Creating put it bluntly. You have to show them the money.
CGI — at the time still trading as Logica — had concluded the same thing and built an answer. The answer was a network of physical innovation centres, distributed across the firm’s offices around the world, where customers could come and actually touch, use, and brainstorm around new technologies before deciding whether they were worth the investment. The firm called them SPARK centres. Sander Van den Born at CGI was the interviewee who walked us through the model when we wrote the 2014 book.
I have been re-reading that case lately. It is one of the few in the book that has, if anything, become more relevant rather than less. What CGI was doing in 2012 looks, twelve years on, like an early worked example of what the current research now calls Value Design as a structured discipline. The same case also contains a small piece of organisational architecture that explains why most B2B suppliers attempting something similar fail at it. I want to talk about both pieces.
The customer’s actual problem
Before the SPARK centre is interesting, the customer’s problem has to be on the page. CGI’s IT-buyer customers were not short of vendor presentations. They were drowning in them. A large enterprise IT function in the early 2010s would be approached every week by suppliers pitching new platforms, frameworks, architectures, paradigms. The supplier’s slides always claimed the new technology would transform the customer’s business. Some of those claims turned out to be right. Many did not. The IT function carried the cost of the wrong ones — in shelfware, in failed implementations, in user resistance, in opportunity cost.
By the time CGI was designing the SPARK centres, IT buyers had developed an entirely rational defensive posture. They did not want another slide presentation. They wanted to see what the technology actually did, with their own data, in something close to their own context, before they committed to anything. They also wanted some evidence that the supplier proposing the technology actually understood it deeply enough to deploy it, rather than having added it to the slide deck because the competitors had.
That is the problem the SPARK centres were designed to solve. The brief, in the language we would use today, was something like: build the venue and the discipline that lets the customer co-create their own use case for a technology they do not yet trust, with a supplier they are not yet sure can deliver it.
What the SPARK centre actually was
Each SPARK centre was physical, located in one of CGI’s offices, and managed by a dedicated team. The centres were not a single global facility; they were a network, with each centre specialising in a set of themes and running its innovation portfolio across three horizons — short term (under one year), medium term (under two years), and long term (under three years). Only the short-term innovations were demonstrable in the centre, because only those had working demonstration prototypes. The medium and long-term work was done in collaboration with academic institutions and partner firms, behind the visible layer.
Each centre had a small group of consultants assigned to it as thought leaders. The detail that matters about these consultants is how they were evaluated. They were not measured on billable days, the way the rest of the CGI consulting workforce was. They were measured on the number of presentations delivered, white papers written, customer workshops run — the activities directly related to promoting the innovation centre. That single piece of HR architecture is, in my reading, what made the rest of the model possible. I will come back to it.
The customer journey through the centre had nine steps. A thought leader identified a business possibility at a customer. The customer was invited to a SPARK session. CGI prepared the session internally. The customer arrived, saw a curated demo of relevant technologies, and participated in a brainstorm on possible applications to their own business. The most promising ideas were worked up by the thought leaders into a proof-of-concept proposal. The customer attended a second workshop to refine the scope. The proof of concept was implemented at the customer’s site. The customer evaluated the outcome. If it worked, the proof of concept became a full project.
Two things are worth noticing about this sequence. The first is that the structure is symmetrical: the supplier brings what the supplier knows, the customer brings what the customer knows, and the centre is the venue designed to make those two knowledges productive together. The second is that nothing in the sequence requires the customer to commit to anything until step nine. The proof of concept is small enough for an executive to authorise without a procurement committee. The full project is authorised only after the proof of concept has delivered real evidence — at the customer’s site, on the customer’s data — that the technology works for the customer’s use case.
Why the asymmetry is the whole point
The line that has stayed with me from a more recent description of the model — found in our own knowledge base when I went back to the source material — is the cleanest version of the argument I have read anywhere. It captures the structural insight CGI was operating on, twelve years ago, in twenty-six words.
CGI brings the technological knowledge and the customer brings their use case knowledge in a session whereby they both search for new projects.
Future of Selling knowledge base — CGI case note
The supplier has the technology knowledge. The customer has the use-case knowledge. Neither party has both. The slide deck is the wrong instrument for the decision because it carries only the supplier’s half of the knowledge into the room. The brainstorm at the SPARK centre is the right instrument because it carries both halves into the same room at the same time. The proof of concept is the right instrument for the contract because it tests the joint hypothesis the two halves produced — not the supplier’s claim, and not the customer’s hope.
That distinction, plain as it sounds, is the one most B2B suppliers still get wrong. They walk into the customer’s office with a deck describing what their technology does, and ask the customer to figure out whether it solves a problem the customer has. The customer, who knows their problems but not the technology, cannot answer with any confidence. So they either over-commit and regret it later, or under-commit and the supplier loses the deal. The SPARK centre reframes the meeting. Instead of supplier-pitches-customer-evaluates, the meeting is supplier-and-customer-jointly-investigate. The relationship the meeting builds is structurally different from the relationship a pitch builds.
The architecture that funds it
Here is where the case gets unusual, and where I think most of the lesson for B2B sales leaders sits.
CGI explicitly runs the SPARK centres as cost centres. Not as profit centres. Not as sales channels with revenue targets attached to them. As cost centres, on the company’s books, with the firm bearing the cost of operating the venue, paying the thought leaders, and running the customer sessions, without expecting that cost to be recovered through a direct revenue line.
That is the architectural choice that most B2B suppliers attempting something similar refuse to make, and the refusal is the reason the model usually fails to take root. When the innovation centre is set up as a sales channel with quarterly revenue targets, the thought leaders revert to behaving like salespeople. They start steering the brainstorm toward the supplier’s existing services. They stop running sessions with customers who are unlikely to convert in the current quarter. The customer notices, very quickly, that the conversation is no longer joint investigation but disguised pitching. Trust collapses. The model stops working.
CGI’s design pre-empts this. The thought leaders are not measured on revenue. They are measured on white papers, presentations, and workshops — the leading indicators of intellectual credibility, not the lagging indicators of revenue. The centres are funded centrally as a strategic investment in the firm’s positioning, not as a sales experiment whose viability has to be proven quarter by quarter. That single piece of financial architecture is what gives the model the room to actually be what it claims to be.
The honest qualifier is that this architecture is expensive. Running a network of innovation centres with dedicated thought leaders evaluated on non-billable activities is a material cost on the P&L. CGI carries that cost because the firm has concluded — correctly, in my view — that the alternative is worse. The alternative is competing on the strength of slide decks against competitors with comparable slide decks, in a customer environment where slide decks no longer earn the decision. The cost of the SPARK centre is the cost of being in a different conversation.
What this looks like through the 2026 framework
Re-read with the current research vocabulary, the SPARK case maps cleanly onto three pieces of what we now call Systemic Value Selling.
It is a worked example of Value Design as a structured discipline, rather than as an activity individual salespeople improvise on their own initiative. CGI took the question of how value is engineered with the customer and gave it a venue, a team, a service blueprint, a set of evaluation criteria, and a horizon-based portfolio. Most B2B suppliers I work with talk about value engineering as a competence; CGI built it as a facility. That is a different thing, and the difference is where the discipline becomes teachable across a sales organisation rather than dependent on individual talent.
It is also one of the earliest worked examples of what the current framework treats as the proof-of-concept as a contractual stage, not as a project-management activity. The 2026 conversation with B2B buyers — Kaneka, Chevron Phillips, Delaware — keeps returning to the same point. Buyers in a permacrisis economy will not authorise large investments on the strength of a vendor’s claim. They will authorise small, time-bounded proofs that test the joint hypothesis on the customer’s site. The supplier who structures their selling around that reality — rather than around the traditional pitch-then-propose sequence — is in a different commercial conversation than the supplier who has not.
And it is a clean illustration of the principle our research now states as: a strong sale increases the customer’s ability to succeed after the sale. The SPARK session, even when it does not lead to a proof of concept, leaves the customer with a clearer understanding of which emerging technologies actually matter for their business and which are noise. That outcome is valuable to the customer regardless of whether they buy anything from CGI. The supplier that consistently delivers that outcome is structurally indispensable to the customer’s thinking, in a way the supplier that consistently delivers slide decks is not.
What this teaches sales leaders now
If you are running a B2B sales organisation in 2026 and your customers are sophisticated enough to have seen through the slide-deck pitch, the SPARK case is worth reading carefully. Three things stand out.
The first is that the right unit of analysis is the asymmetry of knowledge. The supplier and the customer each hold half of what is needed to design a useful application of any technology. A sales process that does not bring both halves into the same room — visibly, on equal footing — is solving the wrong problem. Most sales processes do not. They are organised around the supplier explaining, the customer evaluating, and the supplier winning or losing on the strength of the explanation. The SPARK design treats the meeting as joint investigation, and the contractual sequence as a series of bounded tests of joint hypotheses. That is a different sales process. It produces different outcomes.
The second is that the institutional architecture matters more than the rhetoric. Many B2B suppliers talk about co-creation. Few build the venue, the team, the evaluation criteria, the funding model, and the service blueprint that would make co-creation actually happen at scale. CGI built all five. The difference between talking about a discipline and building the architecture that makes it possible is the difference between most B2B suppliers and a small number of unusual ones.
The third — and this is the harder one — is that the funding model has to follow the architecture. If you set up a SPARK-equivalent function as a revenue centre, you have built a sales channel wearing the costume of a co-creation venue. The customer will see through the costume on the first or second meeting. The function has to be funded as a strategic investment in the firm’s positioning, with the cost taken on the central P&L and the leading indicators measured rather than the lagging ones. That is an executive decision, not a sales-design decision. It is also why the lesson of the SPARK case is, in the end, not really about sales. It is about how the firm decides to spend its money.
Twelve years on
The CGI case has aged the best of any in the 2014 book. The technologies the centres were demonstrating in 2012 — early machine learning, mobile platforms, the first wave of enterprise cloud — have been superseded several times over. The discipline of the centres has not. Generative AI is, today, exactly the kind of technology IT buyers are wary of investing in on the strength of slide decks. The same buyers will commit to a small, time-bounded, on-their-data, joint-investigation engagement that gives them evidence rather than claims. The instrument the SPARK centres were built around in 2012 is, in 2026, more relevant than ever.
What CGI saw earlier than most was that the customer’s resistance to the slide deck was not a sales objection to be overcome. It was a rational response to a decade of being burned by slide-deck-driven investments. The supplier who treats that resistance as a sales problem will keep trying to perfect the deck. The supplier who treats it as a signal about what the customer actually needs will build something quite different. The SPARK centre is one of the cleanest examples we have of a supplier who got the signal right and built accordingly.
Most B2B suppliers will not build a SPARK centre. They do not need to. What they do need to import is the underlying architecture: a venue for joint investigation, a team funded on intellectual contribution rather than revenue, and a contractual sequence that tests joint hypotheses on the customer’s site rather than the supplier’s claims in the customer’s office. Whether that venue is physical, virtual, or hybrid is a second-order question. Whether it exists at all is the first-order one. And on that first-order question, twelve years on, far too many B2B suppliers are still answering with a slide deck.
About the Future of Selling research programme. Co-led by Prof. dr. Régis Lemmens (Solvay Brussels School / AMS Antwerp Management School) and Prof. dr. Javier Marcos (Cranfield School of Management). Continuing the work first published in From Selling to Co-Creating (Lemmens, Donaldson and Marcos, 2014), which is the source of the CGI SPARK Centre case re-read in this article. The original CGI case material in the 2014 book was developed with Sander Van den Born at CGI. Companion articles in this series re-read further cases from the same book — MLP on capability-building selling, and the two Bekaert cases on textile carding and Dramix concrete reinforcement — and more recent interviews at Kaneka, Chevron Phillips Chemicals and Delaware.
Future Of Selling 
Leave a comment