Bonni J. GENZER, Plaintiff - Appellant, v. JAMES RIVER INSURANCE COMPANY, Defendant - Appellee., 934 F.3d 1156

PHILLIPS, Circuit Judge. In this appeal, Bonni J. Genzer, a rideshare driver for Uber Technologies, Inc., contends that James River Insurance Company, Uber's insurer, breached its contractual obligations by declining coverage for injuries she sustained in an accident on the return leg of a lengthy fare. Genzer also contends that, under Oklahoma law, the "mend the hold" doctrine limits James River to the grounds it gave for declining coverage before she sued. The district court granted

summary judgment in James River's favor, first ruling that Oklahoma has not adopted the mend-the-hold doctrine, and next holding that Genzer's claim falls outside the scope of the governing insurance policy. We agree on both issues. Thus, exercising jurisdiction under 28 U.S.C. § 1291, we affirm. BACKGROUND James River issued two Business Auto Coverage policies (the 100 Policy and 200 Policy) to Rasier LLC, Rasier-CA LLC, Rasier-DC LLC, and Rasier-PA LLC. The two policies were in force from March 1, 2017, to March 1, 2018.

The Rasier entities are affiliates of Uber, a company that coordinates ridesharing transportation services through smartphone applications. As explained below, the two policies covered different stages of Uber's ridesharing process. In rough terms, the 100 Policy applies when an Uber rideshare driver is fulfilling requests for transportation services.

As part of this, the driver must be occupying a "covered auto," which the policy's "Covered Auto Designation Symbol," or covered-auto endorsement, defines to include: Any passenger "auto" while being used by a "Rideshare Driver", in connection with the "UberPartner application" accessed using account credentials issued under a contract with a Named Insured, provided any of the following: a.

The "Rideshare Driver" has logged and recorded acceptance in the "UberPartner application" of a request to provide transportation services, and the "Rideshare Driver" is: 1) En route to the pick-up location of the requested transportation services including, but not limited to, picking-up of passenger(s); or 2) Traveling to the final destination of the requested transportation services including, but not limited to, dropping-off of passenger(s). b.

The "Rideshare Driver" has logged and recorded acceptance in the "UberPartner application" to provide transportation services and the "Rideshare Driver" is: 1) Located on a public airport premises during the course of the accepted transportation services including the picking-up and dropping-off of passenger(s); or 2) Located on a public airport premises immediately following the conclusion of the requested transportation services and while in the course of exiting the public airport premises. c.

The "Rideshare Driver" has logged into the "UberPartner application" and is "available to receive requests" for transportation services from TNC application users and "Rideshare Driver" is located on a public airport premises. J.A. at 90. The 200 Policy, by contrast, applies when Uber rideshare drivers are awaiting requests for transportation services.

Under that policy's covered-auto endorsement, a "covered auto" includes: Any passenger "auto" while being used by a "Rideshare Driver", in connection with the "UberPartner application" accessed using account credentials issued under a contract with a Named Insured provided the "Rideshare Driver": a. has logged into the "UberPartner Application"; and b. is "available to receive requests" for transportation services requested through the "UberPartner application"; and c. has not accepted a request through the "UberPartner application" and is not en route to or providing transportation services in response to a request accepted in the "UberPartner application"; and d. is not on a public airport premises.

Id. at 146. On April 17, 2017, Genzer accepted a fare through UberPartner, Uber's smartphone application for drivers, to transport a passenger about 139 miles from Will Rogers World Airport in Oklahoma City to Woodward, Oklahoma. After dropping off the passenger in Woodward, Genzer began heading back to the Oklahoma City area. On the return journey, Genzer was injured when an oncoming semi-trailer truck ejected a semicircular metal object that crashed through her windshield and hit her face. The truck's driver continued traveling and was never identified.

On May 3, 2017, Genzer submitted a claim for uninsured-motorist, medical, rental-car, and collision coverage. Genzer claimed that, at the time of the accident, her UberPartner application was set to the "Available" trip status and that she was "returning from taking a rider to" Woodward. See id. at 248-52. On May 8, 2017, Michael Pitts, a claims examiner for James River, informed Genzer's then-counsel that Genzer appeared to have been "offline at the time of the accident." Id. at 260. Pitts disclaimed coverage for Genzer's injuries on that basis.

On May 9, 2017, counsel responded that Genzer "was logged onto the Uber system on her return trip and looking for a fare for the return trip home when th[e] accident happened." Id. at 259. On May 10, 2017, Pitts replied that, whether "available or offline, there isn't coverage." Id. On May 23, 2017, Pitts sent Genzer a disclaimer-of-coverage letter on James River's behalf. The letter initially explained that the 200 Policy applies when a driver is "available" for ride requests, while the 100 Policy applies when a driver is "en route" to pick up a passenger or is "providing" transportation services. Id. at 203.

But, the letter advised, neither policy applies if the driver "has logged off the 'UberPartner application,' which is the case here." Id. After quoting both policies' covered-auto endorsements, the letter concluded that "[a]s Bonni Genzer was not logged into the 'UberPartner application' at the time of the accident her vehicle does not appear to qualify as a 'Covered auto,' under either the 100 or 200 Policies." Id. at 210. The letter identified no other reason for concluding that Genzer had not been driving a covered auto.

On July 14, 2017, Genzer filed suit in Blaine County, Oklahoma, asserting that James River's denial of uninsured-motorist coverage under the 100 Policy breached its contractual obligations. On August 10, 2017, James River removed the action to the United States District Court for the Western District of Oklahoma. On November 10, 2017, James River moved for

summary judgment, and Genzer cross-moved for partial

summary judgment that same day. On June 4, 2018, the district court granted James River's motion and denied Genzer's motion. ANALYSIS I. Standard of Review We review de novo a district court's grant of

summary judgment. Smothers v. Solvay Chems., Inc. , 740 F.3d 530, 538 (10th Cir. 2014).

Summary judgment is appropriate if "there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law." Fed. R. Civ. P. 56(a). We construe the factual record and all reasonable inferences therefrom in the light most favorable to the nonmovant. Young v. Dillon Cos. , 468 F.3d 1243, 1249 (10th Cir. 2006). The nonmovant, however, cannot defeat

summary judgment by relying on "ignorance of the facts, on speculation, or on suspicion." Conaway v. Smith , 853 F.2d 789, 794 (10th Cir. 1988). II. The Mend-the-Hold Doctrine Before reaching the coverage issue, we address a threshold matter of Oklahoma law. Genzer argues that James River's present rationale for not covering her uninsured-motorist claim differs from the rationale that it gave before litigation. She argues that, under Oklahoma law, the "mend the hold" doctrine bars James River from asserting during litigation a rationale for denying coverage that differs from its prelitigation rationale.

To resolve this argument, we must determine whether James River's pre- and during-litigation rationales indeed differ and, if so, whether Genzer's asserted version of the mend-the-hold doctrine operates to bind James River to its prelitigation rationale. A. James River's Shifting Rationale for Denying Coverage Uber insures its rideshare drivers only when they operate "covered autos" under the 100 and 200 Policies. The 200 Policy's covered-auto endorsement applies when a driver has logged in to the UberPartner application and is available for ride requests while located outside public-airport premises.

When a driver is available and at a public airport, subpart (

c) of the 100 Policy's covered-auto endorsement applies. Once a driver has logged and recorded acceptance of a ride request in the UberPartner application, subparts (

a) and (

b) apply. Subpart (

a) requires that a driver be either (1) en route to pick up the passenger or (2) traveling to the passenger's final destination. And subpart (

b) requires that a driver be located on public-airport premises (1) during the course of the accepted services or (2) immediately after completing the services. Only the 100 Policy provides uninsured-motorist coverage, which Genzer seeks under subpart (a)(2) in this litigation. Before filing her suit, though, Genzer appears to have sought coverage under the 200 Policy. Soon after the accident, Genzer filled out an Uber incident-report form, declaring that she had been "returning" from dropping off a passenger and operating in the "available" trip status when she was injured. J.A. at 249, 251.

Her counsel forwarded the report to James River and requested information on uninsured-motorist, medical, rental-car, and collision coverage. See id. at 248. Only the 200 Policy could apply on the facts as Genzer had recited them, because when injured she was in available status and located outside public-airport premises. So James River shared the 200 Policy with counsel. After some investigation, James River decided that Genzer had voluntarily logged off the UberPartner application and had not been available for ride requests when her accident happened.

It therefore e-mailed Genzer's counsel and disclaimed coverage under the 200 Policy on grounds that Genzer "was offline at the time of the accident." See id. at 260. It further advised that the 200 Policy doesn't include "first party medical" coverage-whether personal-injury protection or uninsured-motorist coverage -regardless of a driver's logged-in status. Id. In response, counsel insisted that Genzer had been "logged onto the Uber system ... and looking for a fare for the return trip home when this accident happened." Id. at 259.

James River replied that "available or offline, there isn't coverage for [personal-injury protection] or coverage for the vehicle's damages." Id. It then mailed Genzer a letter formally advising that, because she "was not logged into the 'UberPartner application' at the time of the accident[,] her vehicle does not appear to qualify as a 'Covered auto,' under either the 100 or 200 Policies." Id. at 210. Genzer sued. In her state-court petition, she demanded for the first time uninsured-motorist coverage under the 100 Policy.

Contrary to her original account that she had been "available" for ride requests when the accident occurred, Genzer asserted that she had been "providing rideshare transportation services." See id. at 5, ¶ 3. Genzer did not elaborate on her theory of coverage, but only subpart (

a) of the 100 Policy's covered-auto endorsement applies when a driver is "providing" services outside public-airport premises. So, after removing the action to federal court, James River disclaimed coverage under that provision. See id. at 43 (arguing that Genzer had not been traveling to pick up a passenger or to a passenger's destination). James River also asserted that the 200 Policy doesn't cover uninsured-motorist claims. See id. at 51, ¶ 15, 62-63. But it did not rely on Genzer's alleged offline status as grounds for denying coverage under either policy. Genzer then cross-moved for

summary judgment, conceding the 200 Policy's irrelevance but claiming coverage under subpart (a)(2) of the 100 Policy's covered-auto endorsement. See id. at 171 (arguing that the provision governs "[t]he scenario applicable here"). In Genzer's telling, the accident had occurred as she was traveling to the final destination of her passenger's "requested transportation services." Id. James River, in turn, reiterated that the accident had occurred both after Genzer had picked up and dropped off her passenger at the "final destination." Id. at 241.

James River also asserted a "factual dispute" over whether Genzer had been offline at the time of her injuries, but it called the dispute immaterial. Id. at 239. So James River has shifted its rationale for denying coverage. But its shift has tracked Genzer's shifting theory of coverage. Genzer initially represented that she had been "available" for ride requests but had not been transporting a passenger when the accident occurred. Only the 200 Policy could apply on the facts as Genzer had recited them, so James River disclaimed coverage under that policy on grounds that Genzer had been offline.

It then asserted the same rationale to disclaim coverage under both the 100 and 200 Policies. Meanwhile, it advised that the 200 Policy does not cover uninsured-motorist claims, even if a driver is online. Genzer changed her narrative when, upon filing suit, she invoked the 100 Policy and vaguely claimed that she in fact had been "providing" transportation services. Only subpart (

a) of the 100 Policy's covered-auto endorsement could apply to this new theory, so James River disclaimed coverage under that provision. And indeed, in a later filing, Genzer confirmed her reliance on subpart (a)(2). James River also disclaimed coverage under the 200 Policy on grounds that the policy doesn't cover uninsured-motorist claims. Genzer conceded that argument, thereby narrowing the dispute to coverage under the 100 Policy's subpart (a)(2). The mend-the-hold doctrine's applicability in these circumstances is unlikely in any jurisdiction.

Nevertheless, having determined that James River has shifted its rationale for denying coverage, we must decide whether Oklahoma has adopted the mend-the-hold doctrine in a form that would operate to bar that shift. We turn to that question now. B. The Mend-the-Hold Doctrine Doesn't Bind James River to Its Prelitigation Rationale for Denying Coverage The "mend the hold" doctrine is a form of estoppel that forbids a party to a contract from assuming different or inconsistent positions on the contract's meaning. See AM Int'l, Inc. v. Graphic Mgmt. Assocs. , 44 F.3d 572, 576 (7th Cir. 1995).

Many states that recognize the doctrine apply it to limit an alleged nonperforming party in a breach-of-contract action to defenses based on the party's prelitigation rationale for its nonperformance. See, e.g. , Hamlin v. Mut. Life Ins. Co. , 145 Vt. 264, 487 A.2d 159, 161-63 & n.2 (1984). The modern trend in these states limits the doctrine to situations in which a party has omitted defenses from its prelitigation rationale in bad faith or to another party's prejudice. See, e.g. , Hayden v. Mut. of Enumclaw Ins. Co. , 141 Wash.2d 55, 1 P.3d 1167, 1171 (2000) (en banc).

Other states decline to restrict parties during litigation to statements made outside the litigation context. In those states, the doctrine limits a party to defenses that it first asserts in the early stages of litigation, and then only when a during-litigation change in position would constitute bad faith or prejudice another party. See, e.g. , Trossman v. Philipsborn , 373 Ill.App.3d 1020, 312 Ill.Dec. 156, 869 N.E.2d 1147, 1166-67 (2007). Whether Oklahoma law would recognize such a doctrine, and in what form, is unclear.

As a federal court sitting in diversity, our task is to "attempt to predict how the highest court would interpret the issue." Belnap v. Iasis Healthcare , 844 F.3d 1272, 1295 (10th Cir. 2017) (brackets and citations omitted); see also Wankier v. Crown Equip. Corp. , 353 F.3d 862, 866 (10th Cir. 2003) ("Where no controlling state decision exists, the federal court must attempt to predict what the state's highest court would do."). In making this prediction, "we are generally reticent to expand state law without clear guidance from its highest court." Schrock v.

Wyeth, Inc. , 727 F.3d 1273, 1284 (10th Cir. 2013) (quoting Taylor v. Phelan , 9 F.3d 882, 887 (10th Cir. 1993) ). We review de novo a district court's

interpretation of state law. Wade v. EMCASCO Ins. Co. , 483 F.3d 657, 666 (10th Cir. 2007). Genzer contends that Oklahoma has adopted the prelitigation mend-the-hold doctrine. In support, she cites a single century-old Oklahoma Supreme Court case mentioning the doctrine: Morrison v. Atkinson , 16 Okla. 571, 85 P. 472 (1906). Yet Morrison did not adopt the mend-the-hold doctrine, much less in its prelitigation form.

Morrison mentioned the doctrine amid a general review of extra-jurisdictional estoppel jurisprudence, but it held only that "parties are restricted on appeal to the theory on which the case was tried in the court below." Id. at 472 (deriving this "true rule" from the estoppel cases). Morrison therefore announced the familiar rule that arguments not asserted at trial are waived on appeal. It did not, however, adopt a proscription against a party changing its prelitigation rationale for contract nonperformance.

Mend-the-hold language appears in a handful of Oklahoma Supreme Court cases from the decade after Morrison . But the court has never endorsed the doctrine as a constraint on an alleged nonperforming party in a breach-of-contract action changing its prelitigation defenses. Instead, the court's "uniform" conception of the doctrine has been that it bars positional shifts between trial and appeal. See Checote v. Hardridge , 31 Okla. 742, 123 P. 846, 849-50 (1911) (collecting cases); see also J.R. Watkins Med. Co. v. Coombes , 66 Okla. 126, 166 P. 1072, 1074 (1917) ; Render v.

Lillard , 61 Okla. 206, 160 P. 705, 712 (1916) ; Bailey v. King , 57 Okla. 528, 157 P. 763, 767 (1915), overruled on other grounds by Bryant v. Montgomery , 73 Okla. 104, 174 P. 1080 (1918) ; Wattenbarger v. Hall , 26 Okla. 815, 110 P. 911, 911 (1910). Nothing in these cases suggests that the court would extend the doctrine beyond the trial/appeal context. Lacking an Oklahoma case adopting the prelitigation version of the mend-the-hold doctrine, Genzer contends that the version is implicit in the duty of good faith that Oklahoma law imposes upon insurers in their contractual relations with insureds.

To support this argument, Genzer adduces cases involving insurers initially denying coverage in bad faith, then asserting legitimate defenses to coverage during litigation. See Haberman v. The Hartford Ins. Grp. , 443 F.3d 1257, 1270-71 (10th Cir. 2006) (applying Oklahoma law) ; Buzzard v. Farmers Ins. Co. , 824 P.2d 1105, 1108-10 (Okla. 1991). In limiting the insurers to their initial denials made in bad faith, these cases indeed resemble the prelitigation mend-the-hold doctrine-though they do not invoke it.

Any resemblance is irrelevant, though, because Genzer does not allege that James River initially denied coverage in bad faith. In fact, she seeks to limit James River to its prelitigation denial irrespective of its good-faith basis for that denial. Such an absolute bar to changing positions is plainly incongruous with a conception of the mend-the-hold doctrine rooted in the duty of good faith. Genzer ultimately resorts to extra-jurisdictional authority. Some of these cases apply the prelitigation version of the doctrine that, in Genzer's view, Oklahoma has adopted.

See, e.g. , Hamlin , 487 A.2d at 161-63 & n.2 (applying Vermont law). Most, however, apply the during-litigation version, which is inapposite to Genzer's claims. See, e.g. , Harbor Ins. Co. v. Cont'l Bank Corp. , 922 F.2d 357, 363 (7th Cir. 1990) (applying Illinois law) ; Emp'rs Ins. of Wausau v. Bodi-Wachs Aviation Ins. Agency , 846 F. Supp. 677, 685 (N.D. Ill. 1994) (same), aff'd , 39 F.3d 138 (7th Cir. 1994). Regardless, none are persuasive on Oklahoma law.

As a federal court sitting in diversity, we must apply Oklahoma law as it exists and not opine on the merits of a doctrine that Oklahoma courts haven't seen fit to adopt. We decline to extend the law in such a manner. In any event, even if Oklahoma had adopted the prelitigation mend-the-hold doctrine, we fail to see how it would apply on these facts. James River has not shifted from its prelitigation rationale for denying coverage in a vacuum.

It initially denied coverage on grounds that Genzer had been offline because Genzer represented that she had been logged in to the UberPartner application and "available" to receive ride requests when her accident occurred. When Genzer filed suit and claimed that she in fact had been "providing" transportation services, James River responded in kind, asserting that Genzer had already completed her passenger's requested services. So James River's denials have simply tracked Genzer's theories of coverage.

To hold James River to its prelitigation rationale, when Genzer has changed her theory of coverage after filing suit, "would be unreasonable to the point of absurdity." See Ryerson , 676 F.3d at 614. An insurer cannot reasonably be held to its original defenses when the insured's theory of coverage is a moving target. Additionally, though James River's rationale given for denying coverage during litigation differs from that given prelitigation, the rationales are not contradictory.

An Uber rideshare driver can both have logged off the UberPartner application and have already completed her passenger's requested transportation services. Both are disqualifying under the 100 Policy, and the existence of one does not preclude the existence of the other. Indeed, James River continues to argue that Genzer had been offline when her accident occurred, even as it argues that she had already completed the accepted services. Rather than abandon the former rationale, James River has simply shifted its focus to the latter one in response to Genzer's changing theory of coverage.

A party that asserts additional, consistent defenses to contract performance in response to the theory in the complaint has not unfairly mended its prelitigation hold. Even the cases on which Genzer relies largely rest on preventing parties from taking inconsistent positions. See, e.g. , Guinness PLC v. Ward , 955 F.2d 875, 899 (4th Cir. 1992) (applying "judicial estoppel or the doctrine of preclusion against inconsistent positions"). More important, perhaps, is that James River provided Genzer fair notice of its during-litigation rationale. Cf. Fry , 636 F.

App'x at 766 (noting that the mend-the-hold doctrine "seems to require only fair notice of the theory for denying coverage"). In its formal disclaimer-of-coverage letter, James River advised Genzer that "[t]he [ ]100 Policy applies after a Rideshare Driver has accepted a request for transportation services and is either en route to pick up a passenger or is providing those services." See J.A. at 203.

It then quoted the 100 Policy's covered-auto endorsement-including subpart (a), the provision that applies when a driver is either en route or is providing transportation services-and reserved its right to deny coverage under any of the endorsement's provisions. Id. at 207-08, 211. Together, the advisory statement, the endorsement, and the reservation of rights would have alerted Genzer that James River might later deny coverage on grounds that she had already finished dropping off her passenger.

With fair notice of James River's during-litigation defense to providing coverage, Genzer couldn't have suffered prejudice from James River asserting that defense. Nor does she even make such an argument. Instead, she urges application of the mend-the-hold doctrine in its most aggressive form, without regard for the actual effect of James River shifting from its prelitigation defense to coverage. Absent any prejudice to the opposing party, however, "invoking the doctrine of mend the hold to bar a valid defense is overkill." Ryerson , 676 F.3d at 614.

Accordingly, even if we assumed that Oklahoma would recognize the prelitigation doctrine, we see no sound reason to apply it to these facts. III. Breach of Contract Genzer argues that the provision of the 100 Policy's covered-auto endorsement under which she claims uninsured-motorist coverage in this litigation is ambiguous and must be construed in her favor. We see no ambiguity in the provision and conclude that Genzer's claim falls outside the scope of its plain terms. Oklahoma substantive law guides our

interpretation of the insurance contract in this diversity action. Yaffe Cos. v. Great Am. Ins. Co. , 499 F.3d 1182, 1185 (10th Cir. 2007). Under Oklahoma law, if a contract is unambiguous, we must "accept the language in its plain, ordinary and popular sense." Duensing v. State Farm Fire & Cas. Co. , 131 P.3d 127, 134 (Okla. 2005). Though ambiguities are construed against the insurer, "[i]nsurance contracts are ambiguous only if they are susceptible to two constructions." Max True Plastering Co. v. U.S. Fid. & Guar. Co. , 912 P.2d 861, 869 (Okla. 1996). We will not "indulge in forced or constrained

interpretations to create and then to construe ambiguities in insurance contracts." Broom v. Wilson Paving & Excavating, Inc. , 356 P.3d 617, 628 (Okla. 2015) (citation omitted). Genzer claims uninsured-motorist coverage under subpart (a)(2) of the 100 Policy's covered-auto endorsement. That subpart applies when a driver "has logged and recorded acceptance" of a "request to provide transportation services" and is "[t]raveling to the final destination of the requested transportation services including, but not limited to, dropping-off of passenger(s)." J.A. at 90.

Genzer asserts that this definition is susceptible to "multiple constructions as to the point that terminates coverage." Appellant's Open. Br. at 26. But the definition creates no such ambiguity. It plainly defines coverage as being coterminous with a passenger's "requested transportation services," which conclude when the passenger reaches his or her "final destination" and fully exits the vehicle with his or her belongings.

Though it contemplates intervening stops en route to that destination-"including, but not limited to, dropping-off of passenger(s)"-its coverage plainly ceases at the last passenger's destination. Genzer's

interpretation is inconsistent with this plain meaning. In her view, subpart (a)(2) contemplates coverage for "the entire route" that she "had to take, as a practical matter, to accomplish her task," including her journey from her passenger's drop-off location back to her starting point or to a different location. Id. at 30. This

interpretation makes sense only if the "final destination" in subpart (a)(2) is the driver's terminus. But the provision describes the final destination in relation to the passenger's "requested transportation services." The passenger, of course, logs in to the Uber Passenger application, selects a destination, and requests transportation. A nearby driver then accepts and fulfills the passenger's request. But the driver plays no role in selecting the destination; indeed, the driver chooses where to travel only after fulfilling the passenger's request.

If such travel mattered, then subpart (a)(2) would cover travel to "the final destination of the driver" and not just "the final destination of the requested transportation services." It does not. Nor is it logical to construe the passenger's "requested transportation services" as somehow including the driver's destination. In the first instance, Uber passengers cannot "request" travel to the driver's destination; the Uber Passenger application lets them select only their own destination and any intervening stops while en route to that destination.

When they arrive and exit the vehicle, the application calculates and charges the fare, clearly concluding the "requested" services. Moreover, passengers care little where drivers travel after dropping them off; it is thus doubtful that they would deliberately "request" (or pay for) such travels. Indeed, passengers do not even know their driver's identity until after they request transportation services and a nearby driver accepts that request. An anonymous driver's future travel plans, then, do not factor into the passenger's request.

Genzer fixates on the phrase "including, but not limited to, dropping off of passengers," arguing that it means that "something beyond the mere dropping off of passengers was intended." Id. at 28. In her view, then, coverage doesn't necessarily terminate when the last passenger exits the vehicle. We agree that this quoted policy language comprehends occurrences other than passenger drop-offs-for example, stops along the way to visit a convenience store, to retrieve money from an ATM, or to drop off some (but not all) passengers.

The Uber Passenger application permits passengers to add such stops to their routes, and the policy logically accounts for that feature. But the prefatory language "traveling to the final destination of the requested transportation services" imposes a temporal limitation on such occurrences. Once the last passenger fully exits the vehicle at his or her final destination, the "requested" services are fulfilled and coverage terminates.

As a result, the last passenger's drop off must be the last occurrence that the "travel" in subpart (a)(2) "includes." Genzer obscures this construction, arguing that the phrase "final destination, including but not limited to dropping off of passenger(s)" isn't clear as to the point at which coverage ends. Id. at 26. But this unclarity arises because Genzer omits the preceding "traveling to the" language, rendering the "final destination" instead of the "travel" the term that "includes" occurrences other than the requesting passenger's terminus.

Elsewhere in her brief, Genzer proffers similar fragments of contractual language which, in isolation, appear ambiguous. See, e.g. , id. at 27 ("The term 'final destination of the requested transportation, including but not limited to dropping-off of passenger(s)' is ambiguous...."). Subpart (a)(2), read in its entirety, yields no ambiguity. Genzer further distorts the policy's language when she argues that, even after dropping off her passenger, she continued occupying a "covered auto" because she "had logged and recorded acceptance" of a ride request. Id. at 29 (emphasis added).

But subpart (a)(2) applies only if a driver "has logged and recorded acceptance" of a request. That is, Genzer must have presently been fulfilling a ride request when the accident occurred. At that point, though, she had already dropped off her passenger at the requested destination, and the acceptance had lapsed. Genzer's insistence that her previous acceptance remained operative for coverage purposes would mean that, after accepting a single ride request, a driver would continue occupying a "covered auto" throughout her travels, even when offline and driving for strictly personal reasons.

Such a result would be contrary to the 100 Policy's structure, which provides coverage only for discrete stages of the ridesharing process. Additionally, had the 100 Policy's drafters intended for coverage to continue beyond the requesting passenger's final destination, they could have included such language as they did in subpart (b)(2). That provision covers situations in which a driver is located on public-airport premises "immediately following the conclusion of the requested transportation services and while in the course of exiting" the premises.

Subpart (b)(2)'s clear provision for coverage after the "conclusion of the requested transportation services" implies the deliberate omission of such coverage from subpart (a)(2). Ultimately, Genzer appeals to our obligation to construe ambiguous policy provisions against insurers and in favor of the type of coverage that insureds expect. She contends that she expected coverage until she reached her final destination and insists that, from her "perspective," she hadn't yet reached that destination when the accident occurred. See Appellant's Reply Br. at 3-4.

We indeed evaluate ambiguous insurance policies to "determine whether an insured could reasonably have expected coverage." Yaffe , 499 F.3d at 1186 (citation omitted). But there is no ambiguity to construe in Genzer's favor. We conclude that subpart (a)(2) of the covered-auto endorsement covered Genzer from when she accepted her passenger's request for transportation from Will Rogers World Airport in Oklahoma City until she dropped off the passenger at the requested final destination in Woodward.

The 100 Policy thus provides no coverage for the injuries that Genzer sustained in the accident during her return journey from Woodward. Though we sympathize with Genzer's misfortune and injuries, this outcome is dictated by the covered-auto endorsement's plain terms. CONCLUSION For the above reasons, we affirm the district court. The "100 Policy" refers to James River's Business Auto Coverage Policy Number CA436100OK-02, while the "200 Policy" refers to Policy Number CA436200OK-02.

Genzer stresses that she did not know the destination to which the passenger had requested transportation until after she accepted the fare. We rely in part on the district court's recitation of undisputed material facts, which the parties do not contest. The district court noted that it is unclear whether Genzer intended to return to central Oklahoma City, Will Rogers airport, or her home. The court found this fact immaterial, however, and assumed for purposes of resolving the

summary-judgment motions that Genzer "intended to return to the Oklahoma City area." J.A. at 278. We agree on the fact's immateriality. The 100 Policy defines "an 'uninsured motor vehicle' to include 'a hit-and-run vehicle [when] neither the driver nor the owner can be identified.' " J.A. at 279. The parties and the district court agreed that the truck's driver qualified as an uninsured motorist under this definition.

The "available" status means what it says-that the rideshare driver is logged into the UberPartner application and is available to receive requests for transportation services but has yet to accept a passenger's request for transportation services. See J.A. at 90, 257. Genzer initially claimed that she was returning from taking a rider to Watonga, Oklahoma. It appears, however, that Genzer took the rider to Woodward and that the accident occurred in Watonga. As explained in detail, infra , Pitts disclaimed coverage under the 200 Policy.

At the district court, the parties disagreed on whether Genzer had voluntarily logged off the UberPartner application or was in "available" mode at the time of her injuries. Yet they ultimately agreed on the dispute's immateriality. We use the term "prelitigation" to refer to the period preceding the complaint's filing. The 200 Policy affords no uninsured-motorist coverage for rideshare drivers in Oklahoma because the named insured has waived such coverage. The incident-report form included three checkboxes for the driver's "Trip status": "Available," "En Route," and "Transporting Rider." See J.A. at 249.

The first status roughly corresponds with the 200 Policy, the second and third with the 100 Policy. Genzer checked only the "Available" box. The parties suggest that Genzer originally submitted a claim under the 100 Policy and not the 200 Policy. But Genzer submitted a claim under neither policy. She evidently lacked access to the policies, so her counsel submitted a generic claim and requested a copy of the governing policy to review. Only the 200 Policy could apply to the facts in Genzer's incident report, because it covers available-status claims when the driver is located outside public-airport premises.

That Genzer requested uninsured-motorist coverage on facts implicating the 200 Policy doesn't render her claim a 100 Policy claim. The e-mail transmitting the policy is in the record, but the attachment is not. It is clear that James River shared the 200 Policy, though, because the e-mail referred to the "policy for available status." J.A. at 260. James River has repeatedly described the 200 Policy as the policy that governs available-status claims.

See, e.g. , id. at 203 (advising that the 200 Policy applies when a driver is "available" for ride requests); Appellee's Br. at 15 n.32 (observing that the 200 Policy is "the policy that applies" when a driver is "available"). After all, the 200 Policy exclusively governs available-status claims, while the 100 Policy primarily governs claims that involve a driver traveling to pick up or drop off a passenger. Clearly, the "policy for available status" means the policy that exclusively governs available-status claims. The 200 Policy governs both Oklahoma and Kansas.

The policy includes personal-injury-protection and uninsured-motorist endorsements for Kansas but not for Oklahoma. Hence, the policy does not include first-party medical coverage for accidents in Oklahoma, even if the requirements for driving a "covered auto" under the policy are otherwise met. The 100 Policy, though, does cover first-party medical claims in Oklahoma. Like the 200 Policy, it includes a personal-injury-protection endorsement for Kansas but not for Oklahoma. But it includes uninsured-motorist endorsements for both Kansas and Oklahoma.

The blanket assertion that Oklahoma's coverage doesn't include "first party medical," then, confirms that James River was discussing the 200 Policy. James River interprets this language to mean that "no [uninsured-motorist] coverage was available" under the 100 Policy, irrespective of Genzer's logged-in status. Appellee's Br. at 18. Yet personal-injury protection and uninsured-motorist coverage are distinct types of first-party medical coverage. And coverage for vehicle damages is not medical coverage at all. Hence, James River's statement cannot be construed as a disclaimer of uninsured-motorist coverage.

Further, James River was advising on coverage under the 200 Policy, not the 100 Policy. Up to that point, the parties had discussed only the 200 Policy, and nothing suggests that they had refocused on the 100 Policy by the time James River made this statement. But even accepting James River's

interpretation, it doesn't explain why James River disclaimed coverage. The phrase "available or offline, there isn't coverage" states only the conclusion that there isn't coverage, not why there isn't coverage. It certainly does not contemplate James River's current rationale for denying coverage, i.e. , that Genzer had already dropped off her passenger before the accident. The rationale that there isn't coverage whether Genzer was "available or offline" doesn't admit of such a specific meaning. The parties disagree at length over the import of this letter.

Genzer insists that, to the extent the letter disclaimed coverage under the 100 Policy's covered-auto endorsement, it did so under subpart (c). James River counters that the letter disclaimed coverage under subpart (a). But the letter didn't expressly narrow its disclaimer to either subpart (

a) or (c). Instead, it quoted the full covered-auto endorsement and then asserted a blanket denial "[u]nder this definition," based on Genzer's allegedly not having been "logged into the 'UberPartner application.' " See J.A. at 207-08. The letter cited no other reason why Genzer had not been driving a "covered auto." James River technically disclaimed coverage under subparts (

a) through (

c) of the 100 Policy's covered-auto endorsement. See J.A. at 52-53, ¶¶ 22-27; id. at 58-61. But it noted that only subpart (

a) could apply because Genzer had not been located on public-airport premises when her accident occurred. See id. at 43 & n.2. It thus focused its denial on that provision. The parties' arguments rely on Oklahoma law, so we assume that Oklahoma law applies. See Trans-W. Petroleum, Inc. v. U.S. Gypsum Co. , 584 F.3d 988, 993 (10th Cir. 2009) ("The parties' arguments rely on Utah law; therefore, we will assume that Utah law applies."). Federal district courts in Oklahoma-including in this case-have expressed skepticism that Oklahoma has adopted the mend-the-hold doctrine. See Fry v. Am. Home Assur. Co. , No.

CIV 14-131-RAW, 2015 WL 519706, at *2 (E.D. Okla. Feb. 9, 2015) ("It does not appear Oklahoma has adopted this doctrine...."); J.A. at 286 ("[I]t is unclear that the 'mend the hold' doctrine applies in Oklahoma.") (citations omitted). Likewise, a panel of this circuit wasn't convinced in 2016 of the doctrine's applicability. See Fry v. Am. Home Assur. Co. , 636 F. App'x 764, 766 (10th Cir. 2016) ("[A]ssuming without deciding that Oklahoma would recognize a doctrine along these lines...."). Having scoured the relevant case law, we conclude that Oklahoma hasn't adopted the doctrine.

Genzer asks us to certify this question to the Oklahoma Supreme Court. But Genzer did not ask the district court to certify the question, and we see no compelling reason to do so now. See Lehman Bros. v. Schein , 416 U.S. 386, 391, 94 S.Ct. 1741, 40 L.Ed.2d 215 (1974) (noting that certification is committed to the "sound discretion of the federal court"). Genzer fails to differentiate the doctrine's pre- and during-litigation versions, but her argument presupposes Oklahoma's adoption of the prelitigation version.

In her opening brief, Genzer contends that these cases apply the mend-the-hold doctrine, though she evidently acknowledges in her reply brief that their bad-faith theory is distinct from the mend-the-hold doctrine. Compare Appellant's Open. Br. at 10 (referring to Buzzard and Haberman 's "application of this doctrine"), with Appellant's Reply Br. at 12 (using the disjunctive "or" to distinguish "a bad faith theory or a mend-the-hold theory").

Genzer also proffers insurance-law treatises discussing the general rule that an insurer waives any grounds for denying coverage that it fails to assert in its initial coverage denial. See 13A Couch on Ins. § 195:55 ; 16C J. Appleman & J. Appleman, Insurance Law and Practice § 9260, p. 393-95 (1981). These treatises, however, do not reflect every state's adoption of the prelitigation mend-the-hold doctrine. Indeed, many states expressly reject the general rule that the treatises discuss. See, e.g. , Ryerson Inc. v. Fed. Ins.

Co. , 676 F.3d 610, 614 (7th Cir. 2012) (observing that Illinois law "does not confine [a defendant] to the defense (or defenses) that he announced before the suit"). Accordingly, while these sources provide a useful analytical starting point for understanding insurance-law principles, we must examine Oklahoma authority for Oklahoma law. Of course, the analysis is different for during-litigation positional shifts.

A party that asserts one defense to contract performance in response to the complaint, then when that defense fails asserts a different defense-even a consistent one-might be attempting unfairly to take a better hold. See Harbor Ins. , 922 F.2d at 363. But here, James River did not assert a new defense when its first defense "failed." Rather, it denied coverage before litigation based on Genzer's factual account, then asserted different grounds for denial in response to the complaint.

Subpart (a)(1), which applies when a driver is "[e]n route to the pick-up location of the requested transportation services," J.A. at 90, is inapplicable.