YU, J.

¶ 1 This case involves a will contest and whether the will proponents presented sufficient evidence to rebut a presumption of undue influence. The trial court invalidated the will at issue, finding that it was the product of undue influence. The trial court's factual findings were not challenged on appeal, but the Court of Appeals reversed and remanded for a new trial, holding that the trial court failed to make findings of direct evidence to support its conclusion of undue influence, relying solely on the presumption of undue influence to invalidate the will.

¶ 2 The proper inquiry here is whether the trial court's unchallenged findings of fact support its conclusions of law. The Court of Appeals erred by reweighing evidence that sufficiently supported the trial court's conclusions. We reverse the Court of Appeals and reinstate the trial court's judgment invalidating the will as a product of undue influence.

FACTUAL AND PROCEDURAL HISTORY

¶ 3 The following summary of facts is based on the trial court's extensive and detailed findings of fact and conclusions of law. Clerk's Papers (CP) at 1162-89. Eva Johanna Rova Barnes was born on July 17, 1916, in Bellingham, Washington. She died at her home in Poulsbo, Washington, on June 27, 2011, just a few weeks before her 95th birthday. Barnes' will was admitted to probate on July 1, 2011. Respondent Michelle Wells was appointed personal representative with nonintervention powers but was later removed by the court and replaced by her husband, Dennis Wells. Barnes' estate includes an acreage of land located on Rova Road that was homesteaded by her parents. The property contains her residence and a rental property in which the petitioners (the Rovas)1 shared a one-half interest. Barnes' probated will completely disinherited the Rovas in favor of Wells and her husband. Wells became acquainted with Barnes as Barnes' rural mail carrier, and the two became friends after Barnes' husband and daughter passed away. After Barnes suffered a fall in her home, Wells became her caretaker.

¶ 4 The Rovas challenged the validity of Barnes' will for lack of testamentary capacity and undue influence by Wells. After a five-day bench trial, the trial court issued 83 findings of fact and 23 conclusions of law, finding that while Barnes had testamentary capacity when she executed the will in contest, the will was invalid as a result of Wells' undue influence. The trial court found that Barnes' increasing dependence on Wells coincided with Barnes' estrangement from her family and that Wells made numerous false statements that "fanned the flame" of Barnes' unfounded anger and mistrust of the Rovas. Id. at 1180-81 (Finding of Fact (FF) 73). Wells became the only person close to Barnes on a consistent basis, eventually replacing Barnes' niece as her attorney-in-fact and assuming the role of caretaker after Barnes fell in her home. Isolated from her family and friends, physically and mentally impaired,2 and totally dependent on Wells, it is indisputable that Barnes was highly vulnerable to undue influence.3

¶ 5 Throughout her relationship with Barnes, Wells and her husband were struggling financially.4 After Wells became more involved in her life, Barnes began writing checks to Wells and Wells' family members for various services and expenses. Just days before Barnes passed away, Wells paid her own mortgage with a check issued from Barnes' personal bank account. Barnes was in or close to being in a coma when Wells wrote this check. The check posted on the same day that Barnes passed away.

¶ 6 On appeal, Wells did not challenge the trial court's findings of fact but assigned error to the conclusions that the Rovas had established a presumption of undue influence that Wells failed to rebut, and that Barnes' will was invalid because it was a product of Wells' undue influence. In an unpublished opinion, the Court of Appeals reversed and remanded for a new trial, holding that Wells had sufficiently rebutted the presumption of undue influence. In re Estate of Barnes, noted at 186 Wn.App. 1004, 2015 WL 786791, at *5, review granted, 183 Wn.2d 1025, 355 P.3d 1154 (2015). The Court of Appeals also found that the trial court did not make any findings of fact of "`positive evidence,'" but had "wholly relied on the presumption" to find that there was undue influence sufficient to invalidate the will. 2015 WL 786791, at *5.

STANDARD OF REVIEW

¶ 7 When reviewing a will contest, the appellate court's function is to determine whether the trial court's findings are supported by substantial evidence. In re Estate of Kleinlein, 59 Wn.2d 111, 113, 366 P.2d 186 (1961); see also Thorndike v. Hesperian Orchards, Inc., 54 Wn.2d 570, 575, 343 P.2d 183 (1959). We defer to the trial court's determinations of the weight and credibility of the evidence. Kleinlein, 59 Wash.2d at 113, 366 P.2d 186. Unchallenged findings are verities on appeal. In re Estate of Haviland, 162 Wn.App. 548, 563, 255 P.3d 854 (2011) (citing State v. Hill, 123 Wn.2d 641, 644, 870 P.2d 313 (1994)).

¶ 8 The trial court's extensive findings of fact in this case are not disputed. Thus, the only question is if the unchallenged facts support the trial court's conclusions of law. Whether the facts rise to the level of undue influence that is sufficient to invalidate a will is a question of law that we review de novo. Id.

ANALYSIS

¶ 9 The right to testamentary disposition of one's property is a fundamental right protected by law. Dean v. Jordan, 194 Wn. 661, 668, 79 P.2d 331 (1938). A will that is executed according to all legal formalities is presumed valid. RCW 11.24.030. Nevertheless, a will executed by a person with testamentary capacity may be invalidated if "undue influence" existed at the time of the testamentary act. In re Estate of Lint, 135 Wn.2d 518, 535, 957 P.2d 755 (1998) (citing Dean, 194 Wn. 661, 79 P.2d 331). "Undue influence" that is sufficient to void a will must be "something more than mere influence but, rather, influence `which, at the time of the testamentary act, controlled the volition of the testator, interfered with his free will, and prevented an exercise of his judgment and choice.'" Id. (quoting In re Estate of Bottger, 14 Wn.2d 676, 700, 129 P.2d 518 (1942)).

¶ 10 The applicable legal framework for determining whether a will is the result of undue influence was established in our seminal case Dean, 194 Wn. 661, 79 P.2d 331. For nearly eight decades, Dean has remained the governing case on undue influence, and it continues to be controlling precedent. The present case does not require us to disturb settled law. The trial court properly invalidated the will in contest for undue influence under the Dean framework.

A. Establishing the Presumption of Undue Influence

¶ 11 When challenging the validity of a will, the will contestant bears the burden of proving the will's illegality by "clear, cogent, and convincing" evidence.5 Dean, 194 Wash. at 669, 671, 79 P.2d 331. Circumstantial evidence may be used to establish suspicious facts that raise a presumption of undue influence. In re Estate of Martinson, 29 Wn.2d 912, 914-15, 190 P.2d 96 (1948). If the presumption is raised, the will proponent must produce evidence to rebut the presumption. Dean, 194 Wash. at 672, 79 P.2d 331. The absence of rebuttal evidence may be sufficient to set aside a will, but the contestant retains the ultimate burden of proof. Id.

¶ 12 The court in Dean identified certain suspicious facts and circumstances that could raise a presumption of undue influence:

Id. Whether the existence of the so-called Dean factors raises a presumption of undue influence is a highly fact-specific determination that requires careful scrutiny of the totality of the circumstances. Id.

¶ 13 The trial court properly held that the facts raised a presumption of undue influence based on the presence of all the Dean factors and other considerations. We reaffirm the Dean factors and find that the undisputed facts in this case substantially support the trial court's conclusion of undue influence.

1. Opportunity—existence of a fiduciary or confidential relationship

¶ 14 The first Dean factor establishes that a confidential or fiduciary relationship may give rise to a presumption of undue influence. The crux of these relationships is a level of trust that leads the testator to believe that the beneficiary is acting in his or her best interests, creating an opportunity for the beneficiary to exert undue influence. Kitsap Bank v. Denley, 177 Wn.App. 559, 571, 312 P.3d 711 (2013).

¶ 15 The trial court's findings of fact were sufficient to meet this Dean factor. A fiduciary relationship inheres in the role of attorney-in-fact, see In re Estates of Palmer, 145 Wn.App. 249, 263, 187 P.3d 758 (2008), and it is undisputed that Wells was Barnes' attorney-in-fact at the time the will in contest was signed, CP at 1185 (Conclusion of Law (CL) 12). Wells exercised her power of attorney by signing checks on behalf of Barnes. Id. at 1176 (FF 54). These facts are sufficient to find that a fiduciary relationship existed.

2. Causation—active participation in procurement of the will

¶ 16 The second Dean factor requires that the beneficiary's actions bring about or affect the testamentary instrument. In this case, although Wells was not present in the room when Barnes signed the will, she was Barnes' sole means of transportation and drove Barnes to the series of meetings that led to the execution of the new will.6 Id. at 1177 (FF 61), 1175 (FF 51), 1176-77 (FF 57, 60).

¶ 17 While the mere act of driving Barnes to the meeting with her attorney is not sufficient in and of itself to satisfy this Dean factor, see In re Estate of Malloy, 57 Wn.2d 565, 570, 358 P.2d 801 (1961), the new will was executed on the heels of what appeared to be Wells' systematic manipulation of Barnes. Wells alienated Barnes from her family by making numerous false statements that "fanned the flame" of Barnes' unfounded anger towards the Rovas. CP at 1181 (FF 73). Wells suggested that the Rovas had deliberately destroyed Barnes' address book—an irreplaceable item of great sentimental value to Barnes—when John Rova helped Wells unclutter Barnes' home, which had been declared unsafe due to Barnes' hoarding tendencies. Id. at 1175 (FF 50). She also accused John of trying to "throw Ms. Barnes under the bus" and stated that the Rovas wanted to put Barnes in a nursing home—untrue statements that "acted to further poison" Barnes' relationship with the Rovas. Id. at 1180 (FF 72). Wells also falsely told the rental property tenants that the Rovas were "greedy villains" who intended to evict them in order to sell the land, develop the property, and become millionaires. Id. at 1173-74 (FF 46). Wells further isolated Barnes by changing her long distance calling plan, making it difficult for family and friends to reach her by phone. Id. at 1179 (FF 69).

¶ 18 When viewed in the context of these actions, driving Barnes to the meeting in which she executed a new will can be reasonably seen as the last act in Wells' campaign to influence Barnes. These findings support the conclusion that the will would not have come into being but for Wells' activities and influence on Barnes.

3. Result—unusually or unnaturally large bequest

¶ 19 Under the third Dean factor, the effect of undue influence must manifest in the testamentary instrument in an "unnatural" or "unusual" way. See In re Estate of Peters, 43 Wn.2d 846, 864, 264 P.2d 1109 (1953) ("If fraud or undue influence had actually been exercised, we would expect to find some indication of this in the way in which the property was devised and bequeathed."). "Unusualness" or "unnaturalness" can be measured by comparison to the decedent's previous testamentary instruments, In re Estate of Chapman, 37 Wn.2d 682, 691, 225 P.2d 883 (1950), or bequests to other beneficiaries, Malloy, 57 Wash.2d at 570, 358 P.2d 801.

¶ 20 The trial court found that Barnes' new will was a "radical departure" from her prior wills. CP at 1187 (CL 20). Both of Barnes' prior wills included the Rovas: first as alternate beneficiaries, then as primary beneficiaries following the death of Barnes' husband and daughter. Id. at 1164 (FF 6, 8). The will in contest completely disinherited the Rovas in favor of Wells and her husband as the sole beneficiaries. Wells and her husband were never named as beneficiaries in Barnes' prior wills, Verbatim Report of Proceedings at 612, yet they received the entirety of Barnes' estate in the new will, leaving nothing to the prior beneficiaries. These facts are sufficient to support the conclusion that the Wells received an unusually and unnaturally large bequest.

4. Other considerations

¶ 21 In addition to the three main factors, Dean enumerates other considerations that could weigh in favor of finding undue influence. 194 Wash. at 672, 79 P.2d 331. These considerations speak to the testator's vulnerability to undue influence due to mental or physical infirmity and the nature of the relationship with the beneficiary.

¶ 22 The trial court properly concluded that all of the "`other considerations'" enumerated by the Dean court were present. CP at 1186 (CL 16). Barnes was elderly—nearly 95 when the will was executed—and "extremely vulnerable to undue influence due to physical limitations, [and] some degree of cognitive impairment." Id. (CL 15, 17). Barnes was dependent on Wells as her caregiver, and Wells' constant presence created ample opportunity to exert undue influence over Barnes. Id. at 1186-87 (CL 15, 19).

¶ 23 The trial court cited the unnaturalness of the will as a "critical factor" in its decision. Id. at 1187 (CL 20). A will is unnatural "when it is contrary to what the testator, from his known views, feelings, and intentions would have been expected to make." In re Estate of Miller, 10 Wn.2d 258, 267, 116 P.2d 526 (1941). The bequest to the Wells was "unnatural" in that they were not natural objects of Barnes' bounty: Wells was 51 years younger than Barnes, she and her husband were unrelated to Barnes, and Wells became consistently involved with Barnes only in the last few years of Barnes' life. CP at 1171 (FF 39, 40). In contrast, the Rovas are Barnes' closest living relatives and direct lineal descendants of the property's homesteaders. Id. at 1163-64 (FF 3, 5). They grew up near Barnes and spent a significant amount of time on the property. Id. at 1163 (FF 4). Up until the last few years of her life, the Rovas shared a close family relationship with Barnes, celebrating her 90th birthday together and including her in important family events, like the wedding of Karen Bow's daughter. Id. at 1165 (FF 12), 1172-73 (FF 44). Under these circumstances, the trial court stated that it "cannot conceive of Ms. Barnes disinheriting the [Rovas] and making this absolutely radical and unnatural change to her prior wills unless she was subjected to undue influence that the evidence suggests she was vulnerable to." Id. at 1187-88 (CL 20).

¶ 24 The trial court's conclusion that all the Dean factors and other considerations were present is sufficiently supported by its findings of fact.

B. Effect of the Presumption of Undue Influence

¶ 25 If the facts raise a presumption of undue influence, the burden of production shifts to the will proponent, who must then rebut the presumption with evidence sufficient to "balance the scales and restore the equilibrium of evidence touching the validity of the will." Dean, 194 Wash. at 672, 79 P.2d 331. However, the will contestant retains the ultimate burden of proving undue influence by "clear, cogent, and convincing" evidence. Id. at 671, 79 P.2d 331.

1. Wells failed to rebut the presumption of undue influence

¶ 26 The trial court properly found that the evidence presented by Wells was insufficient to overcome the presumption of undue influence in light of the totality of the evidence presented. CP at 1187 (CL 22). The Court of Appeals correctly stated that the scope of review is "limited to whether the unchallenged findings of fact support the conclusions of law." Barnes, 2015 WL 786791, at *2. However, the court reversed based on its own reweighing of the evidence in favor of an alternative theory for upholding the will—that "[t]he trial court's unchallenged findings of fact contain more than sufficient evidence that Barnes changed her will for a valid reason, unaffected by undue influence: that she had grown apart from, was suspicious of, and disliked the Rovas." Id. at *4.

¶ 27 This was error—the appellate court's role is to review findings supporting the conclusions the trial court did reach, not to look for evidence supporting an alternate conclusion the court could have reached. Wells does not challenge any of the trial court's findings or offer any evidence disputing the presence of the Dean factors, but selectively restates the trial court's findings to support her alternative theory for Barnes' will. While Wells' story may be persuasive in isolation, we must defer to the weight given to all the evidence by the trial court and its credibility assessment that the facts Wells points to do not balance the scales against the overwhelming evidence of undue influence.

2. The Rovas met their burden of proving undue influence by clear, cogent, and convincing evidence

¶ 28 Whether or not the presumption of undue influence is established or rebutted, the will contestant bears the ultimate burden of proving the will's illegality by "clear, cogent, and convincing" evidence. Dean, 194 Wash. at 671, 79 P.2d 331. We have long recognized that circumstantial evidence alone can be sufficient to support a finding of undue influence. In re Estate of Bush, 195 Wn. 416, 425, 81 P.2d 271 (1938) (quoting Olson v. Washington, 18 Cal.App.2d 85, 86-87, 63 P.2d 304 (1936)). However, a will contestant cannot rely solely on the weight of the presumption to invalidate a will, Dean, 194 Wash. at 673, 79 P.2d 331, and "mere suspicion of undue influence is not enough," In re Estate of Mitchell, 41 Wn.2d 326, 353, 249 P.2d 385 (1952). Rather, the contestant must establish undue influence by producing direct or circumstantial "positive evidence."7 Dean, 194 Wash. at 673, 79 P.2d 331.

¶ 29 Here, the trial court properly found that the evidence met the clear, cogent, and convincing standard in order to find undue influence. CP at 1187 (CL 21). The trial court did not delineate which evidence went to any particular proposition, but we have never held that evidence of the presumption could not also be considered as direct or circumstantial evidence of actual undue influence. As the taking of testimony unfolds at trial, the trial court must consider the evidence as a whole, regardless of which party offers it. The trial court's extensive findings of fact established an unrebutted presumption of undue influence based on the Dean factors, supported by further positive evidence of Wells' systematic influence over Barnes and active efforts to isolate and alienate Barnes from the Rovas.8 Taken together, the findings are easily sufficient to establish undue influence.

¶ 30 The Rovas met their burden and, under the appropriate standard of review, the trial court's conclusions are sustainable. We reverse the Court of Appeals and reinstate the trial court's judgment.

C. Attorney Fees

¶ 31 The Rovas request reasonable attorney fees pursuant to RCW 11.96A.150. Under RCW 11.96A.150(1), we have the discretionary authority to award reasonable attorney fees "to any party: (a) [f]rom any party to the proceedings; (b) from the assets of the estate or trust involved in the proceedings; or (c) from any nonprobate asset that is the subject of the proceedings."

¶ 32 The Rovas properly requested attorney fees in accordance with RAP 18.1(b). We grant the request for fees pursuant to RCW 11.96A.150(1) and order that the fees be paid by respondents Michelle and Dennis Wells.

CONCLUSION

¶ 33 The Court of Appeals exceeded the proper function of appellate review in these types of cases. We reaffirm the Dean factors and reiterate that the substantial evidence standard of review applies on appeal. Applying this precedent to the case before us, we reverse the Court of Appeals and reinstate the trial court's conclusion that Eva Johanna Rova Barnes' will is invalid due to undue influence exercised by Michelle Wells.

WE CONCUR: MADSEN, C.J., JOHNSON, OWENS, FAIRHURST, STEPHENS, GONZÁLEZ, and GORDON McCLOUD, JJ., and KEVIN M. KORSMO, J.P.T.

CHARLES K. WIGGINS, J., not participating.