OPINION
FIDEL, Chief Judge.
In February of 1988, a child suffered a cardiac arrest in surgery and emerged revived but neurologically impaired. This special action arises from discovery disputes in the medical malpractice suit brought by the child and her parents against the hospital and two physicians.
Shortly after surgery, at the direction of the hospital’s legal department, a nurse paralegal interviewed four operating room witnesses. The witnesses now claim at depositions to recall little or nothing of the event. Though the paralegal’s summaries might refresh the witnesses’ recollection, the hospital declines to provide these summaries to the witnesses or release them to plaintiffs’ counsel.
Plaintiffs (real parties in interest) moved to compel disclosure. The hospital and its fellow petitioners responded that the summaries are absolutely protected by the attorney-client privilege and immune from discovery under the work product doctrine. The trial court ordered the summaries to be produced for inspection in camera, found only portions to be privileged, and ordered the remainder—“the functional equivalent of a witness statement”—disclosed.
In these consolidated special actions, the petitioners allege that the trial court abused its discretion. We take jurisdiction because issues of statewide importance are presented and because there is no adequate remedy by appeal. Church of Jesus Christ of Latter-Day Saints v. Superior Court, 159 Ariz. 24, 25-26, 764 P.2d 759, 760-61 (App.1988) (“When a trial court orders disclosure that a party or witness believes to be protected by a privilege, appeal provides no remedy. Special action is the proper means to seek relief.”).
FACTS AND PROCEDURE
Phoenix Children’s Hospital (“PCH”) is housed within Good Samaritan Regional Medical Center, which is owned and operated by Samaritan Health Services (“Samaritan”). PCH and Samaritan share some facilities and personnel, and, according to the affidavit of Samaritan’s general counsel, the Samaritan Legal Department advises PCH and its employees “with regard to professional liability incidents and adverse patient occurrences.”
In March 1988, anticipating litigation, petitioner Cathey Milam Chester, an attorney employed by Samaritan’s Legal Department, directed an investigation of the incident that underlies this case. At Chester’s direction, petitioner Elaine Fraiz, a Samaritan nurse paralegal, interviewed various witnesses to surgical and post-surgical events in the operating room and intensive care unit. This special action concerns Fraiz’s interviews with four of those witnesses—three nurses and a scrub technician who were present in the operating [429]*429room during surgery.1 Each witness signed a form consenting to representation by Samaritan if a claim were filed against her. The form directed the witness not to discuss the event without legal department approval except “1) with health care providers involved with continuing patient care; 2) with Peer Review Committees upon request; 2 and 3) [if required by the employee’s participation as] a student ... in a formal SHS teaching program.”
Fraiz summarized her interviews of the four witnesses in written memoranda, which were maintained by the legal department. When plaintiffs ultimately deposed the four witnesses, the witnesses claimed to recall little of the surgery and surrounding events. Yet petitioners declined to show them the interview summaries to refresh their recollection, as to do so would have waived whatever privilege might otherwise attach. See Samaritan Health Servs., Inc. v. Superior Court, 142 Ariz. 435, 438, 690 P.2d 154, 157 (App.1984).3
Plaintiffs’ counsel, who had learned of the interview summaries through interrogatories and depositions, sought to depose Chester and Fraiz and subpoenaed documents relating to their investigation. Samaritan moved for a protective order, arguing that Chester and Fraiz knew only what they had learned through investigation and that the attorney-client privilege and work product rule shielded their investigative product from disclosure. As Samaritan was not a party, plaintiffs also moved that defendant PCH be compelled to produce the interview summaries, and PCH likewise asserted the attorney-client privilege and work product immunity in response.4
The trial court directed Samaritan and PCH to provide copies of the summaries for in camera review, but to
line through on the summaries any and all material which constitutes matters of attorney opinion, attorney theory, attorney interpretation or attorney surmise. Those portions of the summaries which indicate the witnesses statements of what occurred during the operation ... shall be left without lines being drawn through.
The court indicated that it would extract from each summary and provide plaintiffs “the functional equivalent of a witness statement.”
Upon receipt of the trial court’s order, petitioners brought this consolidated special action and sought a stay, which this Court granted only in part. We directed the trial court to proceed with in camera inspection and designate what, if any, summary portions it intended to release. Petitioners’ counsel were ordered to file under seal a copy of the summaries with this Court. Release of the designated portions to plaintiffs’ counsel was stayed.
The trial court followed our directive, enabling us to review the proposed deletions by the trial court, and the trial court’s finding that, without access to its designated portions, plaintiffs would be precluded from proving significant elements of their [430]*430case. The trial court commented that Samaritan, as soon as an incident is reported, signs up its “employees as ‘clients’ of the legal department of the hospital, does extensive interviews and summaries and thereby wraps this material in a mantle of peer review, attorney/client privilege and attorney work product” The court concluded that
a reasonable reading of Humana Hospital v. Superior Court, 154 Ariz. 396, 742 P.2d 1382 ([App.] 1987) and John C. Lincoln Hospital v. Superior Court, 159 Ariz. 456, 768 P.2d 188 (App.1989) indicates the Court must provide reasonable alternatives____ Where plaintiffs’ opportunity to prove [their] case is as effectively blocked as defendants’ theories would claim in the fourteen motions before this Court, due process would seem to call for some form of remedy.
I. WORK PRODUCT RULE
Petitioners argue that the trial court abused its discretion because the work product rule immunizes the interview summaries from discovery.5
Arizona’s work product rule is embodied in Arizona Rule of Civil Procedure 26(b)(3), which provides qualified discovery immunity for material prepared in anticipation of litigation:
[A] party may obtain discovery of documents ... otherwise discoverable6 ... and prepared in anticipation of litigation or for trial by or for another party ... or for that other party’s representative ... only upon a showing that the party seeking discovery has substantial need of the materials in the preparation of the party’s case and that the party is unable without undue hardship to obtain the substantial equivalent of the materials by other means. In ordering discovery of such materials when the required showing has been made, the court shall protect against disclosure of the mental impressions, conclusions, opinions, or legal theories of an attorney or other representative of a party concerning the litigation.
The witness interview summaries that Elaine Fraiz prepared for Samaritan and PCH fall within the general ambit of Rule 26(b)(3). See Longs Drug Stores v. Howe, 134 Ariz. 424, 429, 657 P.2d 412, 417 (1983) (finding that summaries of witness interviews concerning an event that creates a substantial risk of legal exposure are documents prepared in anticipation of litigation).
The summaries contain to a limited degree the protected impressions of Ms. Fraiz; however, the trial court followed the approved process of in camera inspection and redaction to “protect against disclosure of the mental impressions, conclusions, opinions, or legal theories of an attorney or other representative of a party concerning the litigation.” See Ariz.R.Civ.P. 26(b)(3); Longs Drug Stores, 134 Ariz. at 430, 657 P.2d at 418 (Rule 26(b)(3) does not afford absolute protection to a document containing a mix of mental impressions and facts; if the trial court can separate the facts from protected impressions and the like, the court is free to determine whether the isolated facts should be disclosed.); see also Samaritan, 142 Ariz. at 438, 690 P.2d at 157 (“An in camera examination by the trial court can adequately strike the balance between a party’s interest in discovering any evidence favorable to him and his opponent’s interest in protecting the attorney-client relationship.”).
[431]*431The question narrows to whether plaintiffs have shown, as Rule 26(b)(3) requires, substantial need for the redacted summaries and an inability to obtain their equivalent by other means. The trial court, whose grasp of the evidence surpasses ours, found that without the summaries, plaintiffs would be blocked from proving significant portions of their case.
Our dissenting colleague quarrels with this finding. Stepping outside our record on special action, he counts eighteen volumes in the superior court file, reports the number of depositions noticed in that file, and assumes that, in so intensely processed a case, equivalent information must be somehow available from some other source. Our colleague overlooks the fact that the trial judge who concluded otherwise has administered these extensive superior court proceedings. “[Our] standard of review is whether the trial court abused its discretion in determining that the requirements of Rule 26(b)(3) ha[ve] been met.” Longs Drug Stores, 134 Ariz. at 429, 657 P.2d at 417. We have no warrant to usurp the trial court’s role.
The record before us does not suggest that the trial court was derelict in applying the substantial need and unavailability standards to the materials tendered for review. To the contrary, the trial court denied plaintiffs two of Fraiz’s witness summaries altogether, commenting, “there certainly was no smoking gun in there. There wasn’t even a smoking match.” The trial court ordered release of the remaining summaries only in carefully edited form.
One of these summaries—that of Fraiz’s interview with operating nurse Kim Wilson—exemplifies what the trial court found to meet the criteria of Rule 26(b)(3). Wilson actively participated in the operating room response to the “code arrest.” On March 16, 1988, according to Fraiz, Wilson explicitly detailed the effort to revive the child, describing the child’s external appearance, the color and appearance of her internal organs, the actions and conversations of the operating room physicians and staff, the compressive techniques that they employed, the medications that they discussed and administered, and what the monitors showed. Wilson also compared this code response to others in her experience. At deposition in May of 1990, by contrast, Wilson’s recollection was summarized in the following exchange:
Q: Anything else you remember specifically going on while you were present other than you helped with the supplies, you assisted with the code cart, you scrubbed up and gave whatever help they wanted at the surgical site? Anything else?
A: No. I don’t remember anything other than that. That pretty much sums it up.
The other witnesses, like Wilson, remembered significantly less at deposition than when interviewed. A significant witness’s inability to remember can constitute substantial need. Klaiber v. Orzel, 148 Ariz. 320, 323, 714 P.2d 813, 816 (1986); see also Longs Drug Stores, 134 Ariz. at 429, 657 P.2d at 417 (“[T]he trial court may order production upon a showing that the statements ... are unique because they were taken soon after the event.”); 8 Wright & Miller, supra, § 2025, at 220 (“No one doubts that production should be ordered if the witness has a faulty memory and no longer remembers details of the event.”). Here, by declining to make Fraiz’s interview summaries available to Wilson and the other three witnesses, the petitioners effectively vetoed the possibility of refreshing their recollection. The trial court, by its order, attempted to overrule this veto so that the truth-finding process might profit from witnesses with refreshed recollection of critical moments of the case.7
We find no abuse of discretion. Because plaintiffs showed substantial need, and because the trial court took proper measures to protect the working impressions of Samaritan’s legal staff, we hold that the trial [432]*432court’s order met the requirements of Rule 26(b)(3).
II. ATTORNEY-CLIENT PRIVILEGE
Petitioners and amici argue that, even if work product immunity might yield to plaintiffs’ evidentiary needs, the attorney-client privilege absolutely bars disclosure of communications by Samaritan or PCH employees to a member of Samaritan’s legal staff. This argument poses several subordinate issues:
(1) Does the attorney-client privilege apply when, as here, the conduit of communications is a paralegal?
(2) Did the witnesses’ acceptance of client status in Samaritan’s interview consent forms oblige the court to accord them client status for purposes of the attorney-client privilege?
(3) Should Arizona law permit a corporate client to claim the attorney-client privilege for its lawyers’ interviews with non-control group employees, and, if so, should the privilege be absolute or qualified?
We will discuss each of these issues in turn.
A. Communications to a Paralegal
Arizona’s attorney-client privilege statute provides that
an attorney shall not, without the consent of his client, be examined as to any communication made by the client to him, or his advice given thereon in the course of professional employment. An attorney’s secretary, stenographer or clerk shall not, without the consent of his employer, be examined concerning any fact the knowledge of which was acquired in such capacity.
Ariz.Rev.Stat.Ann. (“A.R.S.”) § 12-2234 (1982) (emphasis added). Paralegals are not mentioned in the statute.
Plaintiffs argue that paralegals are an omitted and, therefore, unprotected conduit for attorney-client communications. See Church of Jesus Christ, 159 Ariz. at 29, 764 P.2d at 764 (“Privilege statutes, which impede the truth-finding function of the courts, are restrictively interpreted.”). Although we accept the precept of restrictive construction and will have more to say about it later in this opinion, we find plaintiffs’ construction too restrictive in this instance.
In section 12-2234, our legislature recognized that lawyers often communicate with clients through agents. See Morris K. Udall et al., Arizona Practice: Law of Evidence § 74, at 139 (3d ed. 1991); see also 8 John H. Wigmore, Evidence § 2301, at 583 (John T. McNaughton rev., 1961) (The common law attorney-client privilege traditionally has included a client’s communications to the attorney’s representative.). We believe that the legislature intended by reference to “secretary, stenographer or clerk” to list a representative, not exclusive, group of agents through whom a lawyer and client might confidentially confer. See Udall et al., supra, § 74, at 139 (“In recognition of the fact that lawyers have employees who will hear or see confidential communications, at a minimum [section 12-2234] means that such communications do not become unprivileged when revealed to clerical employees hired by the lawyer to assist in furnishing legal advice.”).
The law has recognized in other contexts that an attorney may properly and efficiently act through a paralegal. This Court, for example, has found paralegal services compensable in attorneys’ fee awards, reasoning that lawyers should not devote time to tasks more economically assigned to legal assistants, “solely to permit that time to be compensable [in a fee award].” Continental Townhouses East Unit One Ass’n v. Brockbank, 152 Ariz. 537, 544, 733 P.2d 1120, 1127 (App.1986); accord Missouri v. Jenkins ex rel. Agyei, 491 U.S. 274, 288 & n. 10, 109 S.Ct. 2463, 2471 & n. 10, 105 L.Ed.2d 229 (1989). Similarly, lawyers should not retain information-gathering tasks more efficiently delegated to paralegals, solely to protect the client’s privilege. We hold that a lawyer does not forfeit the attorney-client privilege by receiving otherwise privileged client communications through the conduit of a [433]*433properly supervised paralegal employee.8
We caution that this holding does not create an automatic “paralegal-client privilege.” There may be circumstances in which a nominal paralegal serves an investigative function independent of the attorney-client relationship. See Longs Drug Stores, 134 Ariz. at 427-28, 657 P.2d at 415-16 (Communications to insurance investigators are not privileged, as they may be used by the investigator’s insurance company employer for purposes other than securing legal advice for the insured.); Butler v. Doyle, 112 Ariz. 522, 525, 544 P.2d 204, 207 (1975) (Because insurance carriers may utilize statements received from their insureds for purposes independent of their insureds’ legal representation, they are not agents of the attorneys hired to represent their insureds.). Such circumstances, however, are not established here.9 Ms. Fraiz was an employee of Samaritan's Legal Department, acting solely at its direction, in anticipation of litigation against PCH. Her paralegal status did not strip the communications she received of whatever attorney-client privilege might otherwise attach.
B. Client or Witness
We next consider what client may assert the privilege to protect the communications that Fraiz received. Like the trial court, we look beyond Samaritan’s self-serving designation of each witness as a client. The trial court recognized the impediment to the truth-finding process if, simply by signing up an employee-witness as a client, a corporation could “wrap[ ] this material in a mantle of ... attorney/client privilege.”
There is no indication in the record that the interviewees sought, needed, or received personal legal advice. They were merely employees to whom the corporation had ready access to gather facts. Essence, not label, controls; the interviewees were witnesses, not clients, in this case.
We recognize, however, that a corporation’s lawyers must sometimes seek facts from corporate employees in order to “give sound and informed advice” to the corporate client. See Upjohn Co. v. United States, 449 U.S. 383, 390, 101 S.Ct. 677, 683, 66 L.Ed.2d 584 (1981). We therefore consider whether PCH and Samaritan may assert a corporate attorney-client privilege to protect their lawyers’ gathering of facts from corporate employees.
C. Attorney-Client Privilege and the Corporate Client
No recorded Arizona case has yet applied the attorney-client privilege to communications between a corporation and its lawyer. See Udall et al., supra, § 74, at 138-39. However, Arizona’s attorney-client privilege statute, A.R.S. § 12-2234, does not exclude the corporate client. Nor would such an exclusion fit the privilege’s supporting rationale. Udall writes, “[t]he purpose of the privilege is to encourage candor by the client so that the lawyer’s advice will be based on reliable information.” Id. § 74, at 136; see also State v. [434]*434Holsinger, 124 Ariz. 18, 22, 601 P.2d 1054, 1058 (1979). Because lawyer-client candor is important in corporate, as in individual, representation, we hold that an attorney-client privilege is available to corporations under Arizona law.
To define the proper scope of that privilege poses harder questions. Who speaks confidentially for the corporate client? Does the privilege attach to every communication funnelled to the corporate lawyer from a corporate employee? Does the privilege need to extend so far to meet the purpose of promoting well-informed legal advice? Within a large corporate client, information useful to the lawyer may be scattered among many employees at various corporate tiers. Some employees, at the policy-making or advisory level, will participate in formulating the corporate response to legal advice. Others, like the nurses in this case, are mere eyewitnesses whose observations the lawyers and corporate decision-makers seek to discover, as they would the observations of any eyewitness, corporate employee or not. Courts and commentators, sensing a qualitative difference, have struggled to define a corporate privilege broad enough to promote the flow of lawyer-client communications, but not so broad as to block external access to internal corporate witnesses and facts. As the Supreme Court of Illinois explained:
The purpose of the attorney-client privilege is to encourage and promote full and frank consultation between a client and legal advisor by removing the fear of compelled disclosure of information____ Under some circumstances, however, the privilege poses an absolute bar to the discovery of relevant and material evidentiary facts, and in the corporate- context, given the large number of employees, frequent dealings with lawyers and masses, of documents, the “zone of silence grows large.” ... That result, in our judgment, is fundamentally incompatible with this State’s broad discovery policies looking to the ultimate ascertainment of the truth ..., which we continue to find essential to the fair disposition of a lawsuit____ Its potential to insulate so much material from the truth-seeking process convinces us that the privilege ought to be limited for the corporate client to the extent reasonably necessary to achieve its purpose.
Consolidation Coal Co. v. Bucyrus-Erie Co., 89 Ill.2d 103, 59 Ill.Dec. 666, 672-73, 432 N.E.2d 250, 256-57 (1982) (citations omitted).
The Supreme Court of Illinois resolved this dilemma by adopting a corporate attorney-client privilege but limiting the privilege to communications with “control group” employees—those who personify the corporation by their managerial or advisory authority to shape corporate decisions in response to the corporate lawyer’s advice. Id. 59 Ill.Dec. at 673-74, 432 N.E.2d at 257-58; see also City of Philadelphia v. Westinghouse Elec. Corp., 210 F.Supp. 483, 485 (E.D.Pa.), petition for mandamus and prohibition denied sub nom. General Elec. Co. v. Kirkpatrick, 312 F.2d 742 (3d Cir.1962), cert. denied, 372 U.S. 943, 83 S.Ct. 937, 9 L.Ed.2d 969 (1963).
The Illinois Supreme Court’s approach in Consolidation Coal is helpful to our effort to determine the proper scope of the corporate privilege in Arizona. Before discussing that case further, however, we examine the United States Supreme Court’s very different approach in Upjohn, 449 U.S. 383, 101 S.Ct. at 679, a case favored by petitioners and the dissent. There, one year before Consolidation Coal, the Supreme Court rejected the control group test for the federal courts, finding it inadequate to protect the flow “of information to the lawyer to enable him to give sound and informed advice.” Id. at 390, 101 S.Ct. at 683.
1. An Examination of Upjohn:
In Upjohn, corporate counsel had conducted an internal investigation to determine whether corporate employees had bribed foreign officials. The Internal Revenue Service (“I.R.S.”) later sought Upjohn’s investigative product, including the lawyers’ interview memoranda and notes. The Sixth Circuit held the attorney-client privilege inapplicable to communications by non-control group employees “for the sim[435]*435pie reason that the communications were not the ‘client’s.’ ” Id. at 388, 101 S.Ct. at 682. The Supreme Court reversed, refusing to restrict the attorney-client privilege to the corporate managerial or policy-making tier. The Court explained:
Middle-level—and indeed lower-level— employees can, by actions within the scope of their employment, embroil the corporation in serious legal difficulties, and it is only natural that these employees would have the relevant information needed by corporate counsel if he is adequately to advise the client with respect to such actual or potential difficulties.
Id. at 391, 101 S.Ct. at 683.
Petitioners urge us to draw from Upjohn an expansive rule of privilege for the Arizona courts. Petitioners argue, and the dissent apparently agrees, that the privilege should cover whatever fact-finding a corporation’s legal staff undertakes from any corporate employee as long as the employee’s communications are sought in anticipation of litigation or to enable the lawyer to formulate informed legal advice. We decline to adopt so broad a rule.
First, Upjohn does not bind the state courts; it establishes federal evidentiary, not constitutional, law, and we need follow it only to the extent we find it persuasive. See Udall et al., supra, § 74, at 138-39; see also Consolidation Coal, 59 Ill.Dec. at 672-73, 432 N.E.2d at 256-57 (rejecting Upjohn as preserving too broad a corporate “zone of silence” in frustration of the state’s “discovery policies looking to the ultimate ascertainment of the truth”).
Second, evidence scholars have not found Upjohn persuasive. One study describes it as “an opinion largely characterized by unexplicated conclusory language.” John E. Sexton, A Post-Upjohn Consideration of the Corporate Attorney-Client Privilege, 57 N.Y.U.L.Rev. 443, 444 (1982). Another criticizes the Court for overstating the investigative benefits of the attorney-client privilege in the corporate context and for ignoring
the special costs inherent in applying the privilege to corporations. One such ignored cost is that the corporation is more likely than an individual client to manipulate the privilege to conceal information used for nonlegal purposes____ An expansive attorney-corporate client client privilege may encourage corporations to conduct legal investigations into the details of any potentially embarrassing incident, not as a result of legal need, but because of the secrecy the privilege provides to business decisions.
Michael L. Waldman, Beyond Upjohn: The Attorney-Client Privilege in the Corporate Context, 28 Wm. & Mary L.Rev. 473, 493-94 (1987) (footnote omitted); see also 24 Charles Alan Wright & Kenneth W. Graham, Jr., 24 Federal Practice and Procedure § 5476, at 144-49 (1986) (criticizing on similar grounds “the Supreme Court’s flawed vision of the privilege and the world in which it must be applied”); Vincent C. Alexander, The Corporate Attorney-Client Privilege: A Study of the Participants, 63 St. John’s L.Rev. 191, 381. (reasoning that Upjohn applies the privilege more broadly than is logically, practically, or empirically required to meet the investigatory needs of corporate counsel); Stephen A. Saltzburg, Corporate and Related Attorney-Client Privilege Claims: A Suggested Approach, 12 Hofstra L.Rev. 279 (1984); Note, The Supreme Court 1980 Term, 95 Harv.L.Rev. 91, 273 (1981) [hereinafter 1980 Term],
Third, behind Upjohn’s broad language lies a narrow, fact-specific, holding. See Alexander, supra, at 378-79. Although Chief Justice Burger, in a separate concurrence, urged the Court to articulate a standard to guide future cases,10 the remaining justices expressly declined to do so, stating:
[436]*436[W]e decide only the case before us, and do not undertake to draft a set of rules which should govern challenges to investigatory subpoenas. Any such approach would violate the spirit of Federal Rule of Evidence 501. See S.Rep. No. 93-1277, p. 13 (1974) (“the recognition of a privilege based on a confidential relationship ... should be determined on a case-by-case basis”).
449 U.S. at 396, 101 S.Ct. at 686.
In a later decision, the Supreme Court reiterated the fact-specific confines of Upjohn, stating “we [there] considered whether the privilege covers only communications between counsel and top management, and decided that, under certain circumstances, communications between counsel and lower-level employees are also covered.” Commodity Futures Trading Comm’n v. Weintraub, 471 U.S. 343, 348, 105 S.Ct. 1986, 1991, 85 L.Ed.2d 372 (1985) (emphasis added). This is no charter for the blanket rule favored by petitioners and the dissent.
If Upjohn is indeed confined to its facts, the most salient is the I.R.S.’s failure to show that it could not obtain the information that it sought through alternative, less intrusive means. The Court observed that Upjohn had turned over a list of the employees in question and that the I.R.S. already had interviewed twenty-five of them. 449 U.S. at 396, 101 S.Ct. at 686. The Court did not analyze Upjohn as a case in which application of the privilege would block the adversary’s or factfinder’s ultimate access to essential employee-witnesses, but merely as a case in which the adversary would be put to the inconvenience of developing their evidence on its own.
While it would probably be more convenient for the Government to secure the results of petitioner’s internal investigation by simply subpoenaing the questionnaires and notes taken by petitioner’s attorneys, such considerations of convenience do not overcome the policies served by the attorney-client privilege.
Id. This case is far different, and we cannot do it justice by dislodging Upjohn’s broadest pronouncements from their factual context and accepting them indiscriminately here.
2. Some Arizona Precepts and the Illinois Approach:
To shape a rule to fit this case and guide similar cases in the future, we return to some fundamental tenets of Arizona’s evidentiary privilege jurisprudence. First is the limited purpose of the attorney-client privilege—to encourage client candor so that the lawyer may give well-informed advice. Udall et al., supra, § 74, at 136. Our supreme court has written, significantly, that this purpose is accepted in Arizona “not to protect the client, but to ... promote the administration of justice.” Holsinger, 124 Ariz. at 22, 601 P.2d at 1058.
The administration of justice also depends, however, on ascertainment of the truth. Accordingly, the Arizona courts have adhered to the tenet of narrow construction. See, e.g., Church of Jesus Christ, 159 Ariz. at 29, 764 P.2d at 764 (quoting Morris K. Udall & Joseph M. Livermore, Arizona Practice: Law of Evidence § 71, at 123-24 (2d ed. 1982) (“ ‘[Privileges, being in derogation of the truth, are to be strictly construed.’ Exceptions are few and narrow to the proposition that ‘the public ... has a right to every [person’s] evidence.’ ”)); see also 8 Wig-more* supra, § 2291, at 554 (The benefits of the attorney-client privilege “are all indirect and speculative; its obstruction is plain and concrete____ It ought to be strictly confined within the narrowest possible limits consistent with the logic of its principle.”) (footnote omitted).
The Supreme Court of Illinois, reasoning from these same premises, sought a rule that would adequately promote consultation between lawyer and corporate client without unduly frustrating the truth-seeking process of the courts. Consolidation [437]*437Coal, 89 Ill.Dec. at 673, 432 N.E.2d at 257. The court concluded that the control group test “strike[s] a reasonable balance” toward that end. Id. Although one year earlier the United States Supreme Court had criticized that test as unpredictable in Upjohn, 449 U.S. at 393, 101 S.Ct. at 684, the lower Illinois courts had been using the test for nineteen years,11 and the Illinois Supreme Court applauded its “predictability and ease of application.” Consolidation Coal, 89 Ill.Dec. at 673, 432 N.E.2d at 257. Other courts and commentators have found the same. See In re Grand Jury Investigation, 599 F.2d 1224, 1235 (3d Cir.1979) (The test “draws as bright a line as any of the proposed approaches.”); Note, Attorney-Client Privilege for Corporate Clients: The Control Group Test, 84 Harv.L.Rev. 424, 430 (1970) (“[T]he test provides an easily applicable bright-line rule to facilitate judicial decisionmaking.”).12
The Illinois Supreme Court defined the control group to include not only top managerial decision-makers, but also those with pertinent advisory responsibility:
We believe that an employee whose advisory role to top management in a particular area is such that a decision would not normally be made without his advice or opinion, and whose opinion in fact forms the basis of any final decision by those with actual authority, is properly within the control group. However, the individuals upon whom he may rely for supplying information are not members of the control group. Thus, if an employee of the status described is consulted for the purpose of determining what legal action the corporation will pursue, his communication is protected from disclosure.
Consolidation Coal, 89 Ill.Dec. at 674, 432 N.E.2d at 258.
By confining the attorney-client privilege to control group communications, the court left lawyers’ communications beyond the control group to whatever protection the work product rule might afford. With one reservation, which we will come to shortly, we find considerable merit in this approach.
First, as Udall points out:
[Bjeyond the control group, what is really being protected are the statements of witnesses who become clients solely because of the happenstance of employment by the corporate client.
Udall et al., supra, § 74, at 139. The dissent maintains that lawyers will not interview such witnesses without an unqualified attorney-client privilege. Yet lawyers, when representing corporate clients, frequently gather information from witnesses beyond the corporate sphere, even though such communications, however vital to sound advice, are not subject to the attorney-client privilege, but to the more flexible protection of the work product rule. We see no reason why Udall’s “happenstance of employment” should require a more rigid form of protection for routine witness interviews within the corporate sphere. Udall adds:
While one can understand a lawyer’s desire to keep everything concerning his client’s case secret, that function has traditionally been performed by the work product doctrine, prohibiting discovery of trial preparation materials absent a showing of special need, rather than the attorney-client privilege which is an absolute bar to discovery.
[438]*438Id. at 140; see also Saltzburg, supra, at 303 (“[T]he premise of the nonabsolute character of the work product doctrine is that lawyers who need information to provide legal advice will seek it from knowledgeable witnesses, even without an absolute privilege protecting their work.”).
Further, empirical research suggests that, beyond the control group, an absolute privilege is unnecessary to encourage candid communications between employee witnesses and the corporation’s lawyer. See Alexander, supra, at 381 (reporting that nearly two-thirds of a surveyed group of corporate lawyers agreed that a qualified privilege—one surmountable, like work product immunity, upon a showing of special need—would not diminish the candor of non-control group employees); see also 1980 Term, supra, at 276-77 (“[I]n seeking legal advice on future corporate moves, the corporation, ... [even] without the privilege, [has] little incentive to withhold relevant information from an attorney. Similarly, investigations of past corporate behavior ... would continue even in the absence of the privilege—although they might be performed by someone other than the attorney—because corporate management needs the information to correct past inefficiencies [and] to reward and reprimand employees____”).
These reasons lead us to join the Illinois Supreme Court in concluding against extending an unqualified attorney-client privilege to non-control group employees. We recognize, as did the Supreme Court in Upjohn, that such employees may have information material to a lawyer who wishes to properly advise the corporate client. This insight warrants some discovery protection, but does not compel Arizona to define that protection in absolute terms that preclude practical accommodation.
3. Work Product Immunity or Qualified Privilege:
We have one reservation, however, about leaving non-control group communications wholly to the protection of the work product rule. Work product protection is traditionally confined to lawyers’ efforts in anticipation of litigation, whereas corporations seek legal advice on a broad range of matters, in many of which the litigation prospect is remote. Timely and informed legal advice may enhance corporate compliance with the law and prevent later litigation. We believe that the flow of information from the corporation to its lawyer should be promoted whenever useful to this advisory role, and not confined to instances where litigation is in view. The work product rule falls short of this mark.
We meet this concern, however, by importing the familiar discovery hurdles of the work product rule into a qualified attorney-client privilege for communications with non-control group members. See Alexander, supra, at 377-81 (proposing a qualified privilege—a privilege surmountable in limited circumstances of special need—to achieve the right balance of protection for non-control group members); 24 Wright & Graham, supra, § 5476, at 191 (same). This qualified privilege includes the following elements:
a. The qualified privilege applies when management, in order to secure legal advice or assistance for the corporation, directs a corporate employee outside the control group to communicate in confidence with the corporation’s legal staff about matters within the scope of employment.
b. Unlike work product immunity, the privilege does not require a showing that the communication was undertaken in anticipation of or preparation for litigation.
c. Like work product immunity, however, the qualified privilege may be overcome “upon a showing that the party seeking discovery has substantial need of the materials in the preparation of the party’s case and that the party is unable without undue hardship to obtain the substantial equivalent of the materials by other means.” Ariz.R.Civ.P. 26(b)(3); see also Alexander, supra, at 401 (proposing that these requirements for overcoming work product immunity be adopted as requirements for overcoming the qualified privilege for non-control group communications).
d. Additionally, to overcome the privilege, the party seeking discovery must [439]*439demonstrate that its substantial need outweighs the corporation’s interest in maintaining confidentiality. Alexander, supra, at 391-400.
We appreciate that the notion of a qualified attorney-client privilege is unfamiliar. At the root of much discussion lies the all- or-nothing assumption that the attorney-client privilege, if recognized at all, must wholly trump any competing interest of an adversary or the court.13 Yet courts have begun to recognize that certain circumstances may warrant a more flexible approach.14 And in Arizona, where we must advance the law of privilege along lines of reason and experience,15 we should hesitate to define new areas of privilege in absolute terms if the purposes of privilege may be served adequately by a more flexible rule.
We opt for flexibility in this case. We attempt, by adopting a qualified privilege for lawyers’ communications with non-control group employees, to reconcile, without needlessly subordinating one objective to the other, our law’s commitment to full client disclosure and its commitment to full adjudicative access to the facts.
Our rule, like the work product rule, requires the trial court to prudently evaluate claims of substantial need and unavailability. The dissent chooses to attack these standards as too pliant to provide meaningful protection. Yet these are case-hardened standards, long familiar to the courts, imported directly from a rule drafted expressly to “give[] strong protection to work product.” Ariz.R.Civ.P. 26(b) State Bar Committee note (1970 amendment); see also Alexander, supra, at 386 (“Experience under the work product doctrine suggests that adversaries are seldom capable of making the necessary showing that they are unable to obtain the evidence they need from nonprivileged sources.”). We believe that these criteria should place a heavy burden on those who seek discovery and adequately bolster already ample corporate [440]*440incentives to provide legal advisers access to material facts.16 See Saltzburg, supra, at 303 (explaining that corporations have practical incentives, even without the attorney-client privilege, to assure their lawyers sufficient access to the facts to provide well-informed advice).
D. Application of the Privilege
Because the trial court invoked due process in its ruling, we point out that we do not. This area of law requires flexibility, not the increased calcification that would follow a ruling on constitutional grounds. In other respects, however, we conclude that the trial court properly approached and resolved the privilege question: First, as stated earlier in this opinion, the trial court reasonably found.that plaintiffs had proven substantial heed and the unavailability of equivalent evidence from non-privileged sources. Second, the trial court reasonably found that petitioners’ confidentiality interests could be protected adequately through a process of redaction and in camera review. Third, after extending such protection, the trial court properly treated the petitioners’ privilege as a qualified privilege that had been overcome.
CONCLUSION
Although the interview summaries are protected by both the work product doctrine and by a qualified attorney-client privilege, plaintiffs have made the requisite showing that their need for disclosure outweighs the corporation’s interest in confidentiality. Because the trial court did not abuse its discretion in ordering limited disclosure, the relief that petitioners have requested is denied.
CLABORNE, J., concurs.