A classic dilemma in industry-academia-research collaboration: the hard problem is stuck on the production line, the experts are in universities, and in between lies "not knowing whom to ask."
I. Two Ways Collaboration Fails
A technical lead at a manufacturing company told me about their experience: there was a process problem on the production line; after working on it internally for six months without cracking it, they decided to seek collaboration with university experts.
On the first attempt, they asked an industry friend to introduce a professor—after chatting for half an hour at the meeting, they realized the professor's field only marginally touched their problem, and both sides felt awkward. On the second attempt, they searched a bunch of names on academic websites and picked the one with "the most impressive title," but the person had long focused only on basic research, had never collaborated with industry, and had neither the willingness nor the experience to engage.
Both failures point to the same problem: before making contact, the company knew almost nothing about the candidate experts.
II. Screen First, Then Contact
Later they took a different approach: first screen people using MuchSci's talent profiles, then visit with materials in hand.
Step one: translate the hard problem into "search language." Break the process problem down into research fields and key technical points, and turn them into keywords searchable in the profile database.
Step two: search expert profiles by research field. The database contains public profiles of academicians, Distinguished Young Scholars, and others that have already been included; first use the profile database's scale statistics to understand how big the "pool" is in this field, then pull up the public resumes of candidate experts one by one.
Step three: pass through two filters. This is the step that most tests the craft in the entire process:
· First filter: research direction relevance. Look at the keywords and disciplines in the project details—whether the person's projects truly cover the company's technical problem, rather than merely "looking similar" in the title.
· Second filter: signs of industry-academia-research collaboration. Look at project types and undertaking institutions—whether the person has undertaken projects jointly with enterprises or local governments, and whether they are still active recently based on years and funding. What gets filtered out are experts who "understand both the technology and industrial collaboration."
These judgments are all based on hard fields: project details undergo "exact matching of the principal investigator's name + verification of the public institution"; those that fail verification are marked "pending verification," and information not included is directly written as "not provided in the current profile." Better to leave it blank than fill it in with impressions.
Step four: conduct a collaboration-basis analysis. For shortlisted experts, first verify whether there is any public collaboration between the two parties—verified direct collaboration, public relationship exists but collaboration is not proven, no direct evidence found, or unable to determine; the four levels are unambiguous. Then assess fit from four dimensions: research direction, complementarity of results, project trajectory, and overlap of networks. One detail worth mentioning: the tool explicitly states that "a collaboration-basis score is not the probability of successful collaboration"; it provides a reference for match, not a guarantee.
III. Sitting Down at the Negotiating Table with Four Sets of Materials
Once screening was complete, the materials the company had were completely different from two months earlier:
The expert's public resume and eight-dimensional score, verified project details, a collaboration network with an evidence chain, and an analysis of the basis for collaboration between the two parties. With these, the opening line changed from "we came because of your reputation" to "we saw in public data that a certain project you undertook in the past three years matches this problem of ours exactly"—every sentence has a source, and the expert can immediately judge whether this is worth discussing.
Of course, the tool's boundaries are equally clear: it does not produce a recommendation list of "whom to approach and whom to exclude." The materials are a starting point, not a promise; whether a deal can be reached still depends on the two sides sitting down and talking.